The British journal of psychology

^mw^^mmmmmmmmmi

Digitized by the Internet Archive

in 2007 with funding from

IVIicrosoft Corporation

http://www.archive.org/details/britishjournalof06brituoft

THE

BRITISH JOURNAL

OF

PSYCHOLOGY

CAMBRIDGE UNIVERSITY PRESS

Honlion: FETTER LANE, E.G.

C F. CLAY, Manager

«Rjinbur£rt): loo, PRINCES STREET

Uontjon: H. K. LEWIS, 136, GOWER STREET, LONDON, W.C.

lonDon: WILLIAM WESLEY AND SON, 28 ESSEX STREET, STRAND

tenltn: A. ASHER AND CO.

UtipM: F. A. BROCKHAUS

CI)tcnflo: THE UNIVERSITY OF CHICAGO PRESS

Bomfcap. anU Calcutta: MACMILLAN AND CO., Ltd.

Toronto: J. M. DENT AND SONS, Lto.

rokgo: THE MARUZEN-KABUSHIKI-KAISHA

A/i rights reserved

THE

BRITISH JOURNAL

OF

PSYCHOLOGY

EDITED BY

CHARLES S. MYERS

WITH THE COLLABORATION OF

W. BROWN W. H. R. RIVERS

C. BURT A. F. SHAND

G. DAWES HICKS C. S. SHERRINGTON

A. KIRSCHMANN W. G. SMITH

W. McDOUGALL C. SPEARMAN

T. H. PEAR JAMES WARD

CARVETH READ H. J. WATT
G. UDNY YULE

Volume VI 1913— 14

Cambridge

at the University Press

1914

\

(JCambrttige :

PRINTED BY JOHN CLAY, M.A.
AT THE UNIVERSITY PRESS

r

CONTENTS OF VOL. VI.

Part 1. June, 1913.

PAGE

The nature and development of attention. By G. Dawes Hicks . 1

The psychology of visual motion. By Henry J. Watt . . 26

The comparative method in psychology. By Carveth Read . 44

Some observations on local fatigue in illusions of reversible
perspective. By J. C. Flugel. (One Diagram) .... 60

Binocular and uniocular discrimination of brightness. By
Shepherd Dawson. (Six Figures) 78

The quantitative investigation of higher mental processes. By
Stanley Wyatt. (Four Figures) 109

Publications recently received . . . . . . .134

Proceedings of the British Psychological Society . . .136

Part 2. October, 1913.

Are the intensity differences of sensation quantitative? I. B)'^
Charles S. Myers 137

Are the intensity differences of sensation quantitative ? II. By
G. Dawes Hicks 155

Are the intensity differences of sensation quantitative 1 III.
By Henry J. Watt 175

Are the intensity differences of sensation quantitative? IV.
By William Brown 184

The aesthetic appreciation of musical intervals among school
children and adults. By C. W. Valentine 190

Note on the probable error of Urban's formula for the method
of just perceptible differences. By Godfrey H. Thomson . .217

The effects of ' observational errors ' and other factors upon
correlation coefficients in psychology. By William Brown. (One
Diagram) ............ 223

The main principles of sensory integration. By Henry J. Watt . 239

Publications recently received . . . . . . .261

vi Contents of Vol. VI.

Parts 3 attd 4. February, 1914.

PAGE

Freud's theory of the unconscious. By William Brown. (One
Diagram) ............ 265

The analysis of some personal dreams, with reference to Freud's
theory of dream interpretation. By T. H. Pear. (Seven Figures) . 281

The conditions of belief in immature minds. By Carveth Read . 304

An experimental investigation of perception. By Frank Smith . 321

The colour perception and colour preferences of an infant during
its fourth and eighth months. By C. W. Valentine. (One Diagram) 3G3

The testimony of normal and mentally defective ^children. By
T. H. Pear and Stanley Wyatt 387

The conditions which arouse mental images in thought. By
Charles Fox. (One Diagram) . . . . . . . .420

On changes in the spatial threshold during' a sitting. By
Godfrey H. Thomson. (One Diagram) ...... 432

Review ........... 449

Publications recently received ....... 453

Proceedings of the British Psychological Society . . . 455

-Til

LIST OF AUTHORS

PAGE

Brown, William. Are the intensity differences of sensation quanti-
tative? IV 184

Brown, William. The effects of 'observational errors' and other

factors upon correlation coefficients in psychology . . . 223
Brown, William. Freud's theory of the unconscious . . . 265

Dawson, Shepherd. Binocular and uniocular discrimination of

brightness ........... 78

Flugel, J. C. Some observations on local fatigue in illusions of

reversible perspective ......... 60

Fox, Charles. The conditions which arouse mental images in thought 420
Hicks, G. Dawes. The nature and development of attention. . 1

Hicks, G. Dawes. Are the intensity differences of sensation

quantitative? II. . . . . . . . . .155

Myers, Charles S. Are the intensity differences of sensation

quantitative? I. . . . . . . . . .137

Pear, T. H. The analysis of some personal dreams, with reference to

Freud's theory of dream interpretation . . . . .281

Pear, T. H., and Wyatt, Stanley. The testimony of normal and

. mentally defective children ....... 387

Read, Carveth. The comparative method in psychology . . 44

Read, Carveth. The conditions of belief in immature minds . . 304
Smith, Frank. An experimental investigation of perception . .321
Thomson, Godfrey H. Note on the probable error of Urban's formula

for the method of just perceptible differences . . . .217

Thomson, Godfrey H. On changes in the spatial threshold during

a sitting ........... 432

Valentine, C. W. The aesthetic appreciation of musical intervals

among school children and adults . . . . . .190

Valentine, C. W. The colour perception and colour preferences of an

infant during its fourth and eighth months ..... 363

Watt, Henry J. The psychology of visual motion .... 26

Watt, Henry J. Are the intensity differences of sensation quanti-
tative? Ill •. 175

Watt, Henry J. The main principles of sensory integration . . 239
Wyatt, Stanley. The quantitative investigation of higher mental

processes ........... 109

Wyatt, Stanley, and Pear, T. H. The testimony of normal and.

mentally defective children ....... 387

v'/u

PROCEEDINGS OF THE BRITISH PSYCHOLOGICAL

SOCIETY

PAGE

Meetings on March 8, May 3, 1913 . . . . . . .136

Meetings on June 7, June 8, November 8, 1913, January 24, 1914 . 455

Volume VI JUNE, 1913 Part 1

THE BRITISH

JOURNAL OF PSYCHOLOGY

THE NATURE AND DEVELOPMENT OF
ATTENTION.

By G. DAWES HICKS.

I. 1. The two opposed classes oj theories of attention.

2. Tlie error underlyiiig both of them.

3. In both there is wanting sufficient recognition of the early

stages of mental development.

II. 1. Attention and apprehension.

2. Distinction of the act or process of appi'ehending and the

content apprehended.

3. The rudimentary or primitive act of apprehension.

4. Apprehension is always a process of discriminating and com-

paring,

III. 1. The com,mon or essential characteristics of attention.

2. The rudimentary or primitive act of attention.

3. The nature of clearness and distinctness.

4. The importance of movement-factors in primitive attention.

5. Secondary attention. The nature of 'interest.'

6. Voluntary or deliberate attention. The 'sense of effort.'

7. Conclusion.

"Attention," said Ebbinghaus, "is a veritable embarrassment of
psychology. In some even comprehensive treatments of the mental life
it has still up to the most recent time been as completely neglected, as
for the most part it had been neglected in the English association-
psychology. In others it appears to be inserted in the whole system in
the strangest, now and again it may be said in the most helpless,
fashion. That under these circumstances great divergence prevails
both in the presentation of its features and in the more detailed

J, of Psych, Yi 1

2 The Nature and Development of Attention

accounts of its nature is intelligible. As compared either with the
ordinary use of the term or with the inner relatedness of the facts
appropriately grouped together as facts of attention, the majority of
these accounts are too narrow ; they are too inclined to take into con-
sideration but one side of the matter or something which is only to be met
with under special conditions ^" These remarks may fittingly serve as
a preface to the following attempt to traverse once more a field that
may seem to have been already sufficiently explored, and from which
attempt, it may be thought, little or nothing of fi-esh importance is
likely to accrue. In spite of much to the contrary that has been
written of late, I venture, however, to plead that the subject calls for
the mode of treatment I have here in view. An account, such as that
presented by Titchener^, of the experimental work that has been done
on attention exhibits only too plainly the barrenness in psychological
significance of what has thereby so far been achieved, and illustrates
also how intimately the value of what is yielded by experiment depends
upon the theoretical standpoint from which specific problems are framed
and their solution sought. There is no possibility of entering upon
an investigation of the conditions and laws of attention without some
conception of the nature of the process itself, and if that conception be
not reached through a careful analysis of the facts of the mental life,
the alternative will be the acceptance of the crude uncritical generalisa-
tions of popular reflexion with which, in that case, every interpretation
of the experimental results will be saturated.

I.

§ 1. A variety of ways of handling the facts of attention has become
familiar in the history of psychology. In the long run, however, they
will all be found to fall under one or the other of two broadly contrasted
modes of explanation which Lotze characterised in a well-known passage ^
On the one hand, attention has been depicted after the fashion of
a varying power of illumination, a sort of waxing and waning light,
which may be directed by the mind upon the several presentations
it receives, and in accordance with the variations of which, other things
remaining the same, will be the clearness, the precision, the complete-
ness in apprehension, of what is attended to. Lotze himself remarks
upon the little help there is to be got fi-om a view of this kind. By

1 Grundzilge der Psychologic, i. 2»« Aufl. 1905, 611,

2 The Psychology of Feeling and Attention, 171 sqq.
' Metaphysik ; Buch iii., Psychologic, § 273.

G. Dawes Hicks 3

concentrating attention upon presentations we are said to increase their
intensity, their clearness, their distinctness. But a mere gazing at
anything, even though it be heightened to infinity, would in itself be
powerless to bring about this result. A mere putting forth of energy,
supposing an intelligible meaning could be found for such an expression,
would be utterly fruitless in accounting for the effects produced, unless
it were shown to be a means of bringing the content in question into
comparison with and relation to the elements surrounding it. On the
other hand, attention has been held to be no more than a name for
describing the varying degrees of intensity which the several presenta-
tions and re-presentations entering into the mental life themselves
possess, — a name, that is to say, for indicating a property or attribute
of the presentations or re-presentations themselves, in virtue of which
they secure special notice on the part of the apprehending subject, not
a property accruing to them in and through the act of attending. The
presentations are regarded as rising, so to speak, into consciousness,
through their own strength ; they are supposed to act and react upon
one another, to blend or fuse with one another, to inhibit one another,
to revive one another, and so forth, and the growth in clearness, distinct-
ness, liveliness, of the predominant presentation is accordingly traced
back to mutual influences of the kind mentioned. In short, the energy
or activity which on the former view was ascribed to the process of
attending, is transferred on this view to the contents attended to, and,
in accordance with the dictum that "psychology is concerned with
nothing beyond presentation and its laws," the only psychological
explanation that can be offered of attention will be in terms of the
movement, the reciprocal relations, of presentations or ideas.

§ 2. These two modes of contemplating the phenomena in question are
usually considered to be essentially antithetical, and so, no doubt, in many
respects they are. But antithetical positions frequently participate in
a common error, and I think we have now before us an instance in point.
The error which, as it seems to me, lies at the root of both methods of
procedure, and which is responsible for most of the perplexities to which
they give rise, consists in treating presentations, contents of apprehen-
sion, as objects. I can prepare the way for what is to follow by dwelling
for a moment upon this initial matter. One of the disadvantages
appertaining to the use of the term 'presentation' is the difficulty,
well-nigh the impossibility, of freeing that term firom the incidental
significance which has come to attach to it on the strength of our
relatively mature experience. In mature experience we have certainly

1—2

4 The Nature mid Bevelojjment of Attetition

come to look ordinarily upon the act of apprehending as distinct and
separate on the one hand from the real thing and on the other hand
from what we call the presentation of the real thing. We habitually
distinguish, for example, the act, or process, of hearing both from (say)
the vibrating string of a violin and from the sound heard. I have no
quarrel with this distinction which, so far as it goes, is legitimate enough.
But the distinction may have, and I should say obviously has, a very
different significance in the two cases just indicated. From a psycho-
logical point of view, at any rate, we are entitled to say that the act of
apprehending is one existing fact, and that the vibrating violin string
is another and a separately existing fact. But are we likewise entitled
to say that the act of apprehending is one existing fact and that the
presentation, the content of the act of apprehending, is another and
a separately existing fact ? For reasons that will presently appear, I
reply to that question without hesitation in the negative. Meanwhile,
however, it will be well to draw out certain implications of such nega-
tive answer. According to a sufficiently prevalent mode of looking at
the matter, what takes place when a sound is perceived would be
described somewhat as follows. The vibrating body, by transmitting
its movements to the air, affects the auditory end organs, a certain
change in the condition of the auditory nerves and the cerebral centres
with which they are connected ensues, and in consequence there arises
in the mind an auditory presentation, which by a mental act directed
upon it is apprehended or attended to. The auditory presentation, that
is to say, will exist as a reaction or modification of the mind, whether it
be apprehended, whether it be attended to, or not. The presentation
will be one mental fact, the act of apprehending or of attending will be
another mental fact. Now, the negative answer just given to the question
proposed involves the rejection of this interpretation. I leave out of
account here, as irrelevant to the present issue, the problem around
which much discussion has recently centred, whether namely a secondary
quality, such as a sound, is rightly described as mental, or as physical,
or as neither mental nor physical. I confine myself to the specific set
of occurrences involved in the perception of a sound, and roughly the
analysis I propose to substitute for the foregoing is this. On the
occasion of the vibrations of air, which emanate from the vibrating
violin string, affecting the auditory nerves and thereby the appropriate
cerebral centres, there arises, whether as effect or correlative or in
virtue of whatsoever relation between body and mind be postulated, not
forthwith a presentation, not forthwith a sound, but a state or process

G. Dawes Hicks 5

of consciousness, an act of apprehension, which, in this case, is the act of
distinguishing or recognising a sound. The auditory presentation, that
is to say, is the content of the act of apprehension and has no separate
existence, as an independent reaction of the mind, prior to the act of
apprehension, and upon which the act of apprehending is, so to speak,
subsequently directed. There is no having a presentation apart from
the process of apprehending itself The distinction, important though
it be, between the act of apprehending and the content apprehended,
does not, therefore, warrant the assumption of two separate mental
facts or occurrences. Such, however, is precisely the assumption which,
either tacitly or avowedly, is made when presentations are treated as
objects. Presentations, so conceived, have unavoidably conferred upon
them a quasi-substantive mode of existence. They come to be regarded
as independently existing entities which form themselves into series,
group themselves, operate upon one another, and which in fact discharge
the specific functions we are in the habit of ascribing to what we call
things. No psychologist would, I take it, care to speak of presentations
as literally ' things,' and yet surely when presentations are dealt with
by aid of the notions just referred to, it would be extremely difficult to
say in what respect they are supposed to differ from ' things.'

§ 3. The objection I am pressing seems, then, to me to be fatal
to both the classes of theories signalised by Lotze. I urge further that
each of these modes of treatment discloses, on nearer scrutiny, the bias
that inevitably results from a too exclusive limitation of the problem to
conscious experience of a highly developed type.

The notion of attention as a focussing of psychical energy upon the
different aspects, elements and objects in the one field of consciousness
is, I suppose, a notion which is naturally suggested by ordinary reflexion
upon what appears to be obviously a feature of the mature mental life,
— the facility, namely, which we possess of attending in various degrees
to the same fact, or to what we call the same fact, and of giving a
similar degree of attention to facts extremely varied. From this merely
popular and supei-ficial introspection of the higher forms of attention to
the assumption of a unique and specific activity is, however, a far cry.
If no better grounds can be furnished than those derived from a merely
popular way of describing experiences which in themselves no one
doubts, the assumption of a unique and specific activity would amount
to no more than a resuscitation of the antiquated doctrine of faculties.
A similar line of consideration should constrain us to the admission of a
specific activity of perceiving, of imagining, of remembering, of thinking,

6 The Nature and Development of Attention

and so forth. And notwithstanding the disposition, on the part of
some recent writers, to reinstate the old faculty psychology, no one has
attempted to set aside the reasons that led Herbart, for example, to call
for its unreserved rejection. Briefly those reasons were these. The
doctrine has no explanatory value. It is a resort to the idea of an
occult power from whose assumed mode of operation the observed
changes that are accounted its effects in the contents apprehended are
deduced. A simplicity, a singleness, is implied in the cause, the specific
mode of psychical energy, which is altogether out of keeping with the
extreme complexity of the results it is said to produce. A single
agency of the kind referred to is wholly irreconcilable with the intricate
manner in which mental processes are connected with one another, and
is in violent contradiction with the ascertained facts of mental growth
and development. Let it suffice to dwell here on the last point. Even
the most cursory inspection must convince us that the states of mind
called states of attending, although evincing certain common funda-
mental features, exhibit differences in character clearly dependent on
the different stages of mental evolution at which they have made their
appearance. When, for example, in a well-known passage, Hamilton
distinguishes three grades or kinds of attention, (a) a mere vital and
irresistible act, {h) an act determined by desire, which, though in-
voluntary, may be resisted by our will, and (c) an act determined by
a deliberate volition, he may be said to be mapping out three prominent
phases in the historical development of mind. The conditions which
determine attention in these three forms will turn out to be to a large
extent of dissimilar origin. In the case of the first, they will be mainly
of the mechanical order, psychological analysis will be confronted with its
limit, and the explanation, hypothetical it may be, will tend to pass into
the region of physiology ; in the second, they will be prevailingly of the
character that it is customary to include under the comprehensive term
' interest ' ; and finally, in the third, they will be those involved in the
reflective pursuit of a deliberately conceived end or purpose. Even,
then, so much recognition of difference in the way in which attention is
manifested is really inconsistent with the doctrine of a specific faculty.
Such other grounds as are advanced for the view are generally drawn
from those unquestionable experiences of the mature consciousness in
which what is called the ' effort to attend ' is involved. But, in the first
place, the ' effort to attend,' in the form in which we are familiar with
it, is evidently an experience possible only for a mind of a highly
developed type. It involves clearly anticipatory representations,

G. Dawes Hicks 7

accompanying desires, the consequent repression of ideas which otherwise
might form part of the sphere of apprehension at the moment, and
generally certain of those motor experiences which make their appear-
ance in any exercise of effort on our part, and which themselves can
never find expression in any single species of sense-presentation. There
is usually involved, further, the admittedly complex factor, consciousness
of self, and it can scarcely be questioned that experiences in which the
representation of self is an ingredient are of far too composite a
nature to entitle us to draw any conclusion from them to the primitive
stages of the activity of attending. In the second place, it has to be
borne in mind that what may appear to us introspectively, in our
mature experience, to be a simple and unanalysable process need not
by any means necessarily be a simple and unanalysable process as a
psychological fact. Mental processes of the most intricate kind re-
peatedly come to appear simple the more readily and habitually they
are performed. If, then, attention is to be regarded as a special activity,
one certainly requires better evidence for it than such as is furnished
by the antithesis between what is roughly described (by Wundt, for
example) as the attitude of 'passive receptivity' and the attitude of
active exertion or effort. Important though this difference may be, and
no one questions its fundamental importance as gradually making its
appearance in the development of mind, yet, unless it can be shewn
to be ultimate, it affords no presumption whatever in favour of the
hypothesis we are considering.

On the other hand, the notion of attention as capable of explanation
in terms of the "movement, the reciprocal influence upon one another, of
presentations, may be thought to arise naturally from ordinary reflexion
upon the more rudimentary or non- voluntary stages of the process. And
this, no doubt, to a large extent is true. As contrasted with others,
certain presentations seem to have the power of thrusting themselves
into prominence, of becoming ' dominant,' to use Mr Bradley's phrase, in
consciousness. " A flash of lightning by night, the report of a firearm,
the sudden prick of a knife, or a violent internal pain, all these for the
moment so occupy our notice that everything else becomes feeble or is
banished." Here, again, however, I believe we may be easily misled by
trusting too implicitly what inspection of our mature experience seems
obviously to yield. The facility which we possess of sharply distinguish-
ing presentations from one another, of holding them definitely apart, of
regarding them as though they were so many separate, detached, indi-
vidual objects, may be an acquired facility, a facility attained by means

8 The Nature and Development of Attention

of the process of attention itself. If isolation, independence, definiteness
of outline, be not originally given, if, as features in what is apprehended,
they come about gradually in the history of the mental life, and largely
through the very process we are seeking to account for, — and this,
surely, is the position to which psychological analysis unequivocally
leads, — then clearly these must not be presupposed as data on the basis
of which a theory of attention can be framed. It must not be assumed,
that is to say, either that presentations are in truth so many discrete
and independent entities, or that in their original character they are
objects of the mind's contemplation. For, according to the general
principle just indicated, presentations will only appear to be the one or
the other in consequence of there having been brought to bear upon
them reflective considerations, the legitimacy or illegitimacy of which
as psychological categories it will be needful to determine.

II.

§ 1. If, then, we reject on the one hand the view according to
which attention is a special activity, and on the other hand the explana-
tion of its characteristics by reference to the interaction of presentations,
it remains to make the attempt to shew in what way the undoubted
peculiarities of the facts of attention can be accounted for by tracing
the way in which the process of attending is connected with the simpler
and more primordial processes of mind. Professor Ward's extended use
of the term attention " so as to include what we ordinarily call inatten-
tion" has, at least, the advantage that the essential continuity of
attentive consciousness with inattentive consciousness or consciousness
simply is thereby emphasized. The difference between these two forms
of consciousness is, it is maintained, a difference not of kind but of
degree. But true and significant though I conceive this contention to
be, it would, I think, be a mistake to deprive psychological terminology
of the only single expression it has for denoting a body of facts which
do exhibit, under a variety of differences, certain common characteristics
as contrasted with the states of consciousness which we ordinarily call
those of inattention. That underlying this specific distinction there is
a more fundamental and generic identity is indeed precisely the thesis
I am concerned to defend, but I feel the need of a term by which com-
pendiously to describe the set of specific features which in that case call

G. Dawes Hicks 9

for explanation^. For reasons that will immediately appear, the term
' apprehension ' seems to me a suitable designation for the generically
identical process which, following Professor Ward, I believe to be at
work in both the contrasted phases of conscious experience, and the
problem I have now to face is that of trying to shew how an intelligible
account can be given of attention as a differentiated mode of this one
fundamental process.

§ 2. So long as analysis is confined to what is met with at com-
paratively advanced stages of mental development the problem thus
formulated hardly admits of successful solution. In dealing with the
mature conscious life, there is always the temptation to accept as simple
and ultimate what in truth may be an exceedingly complicated product,
and it is just in this particular that introspective observation is almost
certain to lead us astray. It is well-nigh impossible to disentangle the
intricate phenomena that come before us in the higher forms of attention,
unless we endeavour to work our way back from these to the earlier and
less evolved phases of experience, out of which they may be conceived
to have originally emerged. Pursuing that course, many considerations
seem unmistakably to indicate that what is vaguely called the power
of concentrating attention is a derivative function, that its ingredients
are by no means invariably the same, that it is largely of the nature
of a habit, and that there is no more of ultimate activity or faculty
involved in it than such as belongs to the fundamental process of
apprehending, of which it is a complicated and highly developed
modification.

One of the considerations referred to is that to which I have already
alluded. In mature experience no distinction we can draw seems more
obvious and self-evident than that between the act of apprehending and
the content apprehended. We unhesitatingly distinguish hearing from
the sound heard, seeing from the colour seen, and so forth. And it is
undoubtedly on this distinction that the view of attention as a specific
unique activity is in the long run based. For here likewise nothing
seems more manifestly in accordance with fact than that the act of
attending and the content which results therefrom are two distinct
and separate entities. Yet the slightest reflexion, as Hume would say,
is sufficient to dispel the idea that this distinction, for the mature

^ I doubt the advisability of employing such a mode of expression as • degrees of
distribution of attention ' to indicate the specific features in question (cf. Ward in Mind,
1887, XII. 58), because it almost inevitably suggests the thought of attention as being
comparable to a ' moveable light.'

10 The Nature atid Development of Attention

consciousness seemingly so indubitable, could conceivably have entered,
even in the crudest form, into the experience of the primitive mind.
It is a distinction the recognition of which presupposes a superstructure
of reflexion such as is utterly beyond the scope of the early mental life,
and can only be possible for a mind that has attained, in some measure,
at least, to the stage of self-consciousness. We distinguish the content
from the act of apprehending, but what in our mature experience gives
stability and definiteness to the content as thus distinguished is the
presence of a number of thoughts or concepts which connect the content
in question with the objective order of real fact. So too we distinguish
the act of apprehending from the content apprehended, and what gives
stability and definiteness to the distinction is a number of thoughts
or concepts which connect the act of apprehending with the train of
experiences constituting what we call the self. It is quite true that
even a mind of the simplest type must contain within itself the
indispensable condition of the distinction which subsequently makes its
appearance in the course of the successive stages of mental evolution.
It is quite true that psychical existence has precisely the peculiarity of
exhibiting in all its states a unique double-sidedness of aspect; its
states are, that is to say, from the very first, states in and through which
the subject is aware, and aware of something. And I anticipate the
objection that the content may as a matter of fact be separate from
and independent of the act of apprehending even though it be not
recognised as thus separate and independent by the experiencing
subject. But the point is that the start is bound to be made from
an act of apprehending which is such only as the awareness of a
content. To maintain that the two are separate even though they are
not recognised as separate does not help in the least to account for the
gradual emergence of the distinction as a recognised distinction. It is
still from the concrete fact, an awareness of a feature or features which
as defining the act of awareness we call the content of it, that any such
account must proceed, and even though there were an actual severance
between act and content, such actual severance would avail us not at
all in tracing the way in which the distinction in question comes to be
drawn by the conscious subject. Just because, therefore, the distinction
cannot be present as a recognised distinction in primitive experience,
the fact of its recognition in mature experience does not warrant the
conclusion we are too hastily inclined to rest upon it, that the act of
apprehension is one existing entity, and the content of that act, or the
presentation, another. In other words, the assumption of a primitive

G. Dawes Hicks 11

severance in fact between act and content seems to me to be " an
illegitimate transference," as Adamson puts it, 'to the supposed original
condition of the inner life of a distinction which has definite meaning
only in its later form^" that is, in the form in which it points to the
contrast between the act of apprehending and a real thing appre-
hended ^

§ 3. The rudimentary act of apprehension, from which conscious
life may be supposed to start, would not, so far as I can see, be fitly
described as a state either of cognition or of feeling or of conation.
After all, these terms are generalities, and serve mainly to indicate
certain broad differences which are recognised in mature experience
and which have become fixed in our ordinary nomenclature. They
name with sufficient accuracy three lines along which the mental life
develops, each of which reaches, by degrees, a certain measure of inde-
pendence of the others. But the meaning assigned to these terms is
determined wholly by our knowledge of the processes in the developed
stage, and it is prima facie extremely improbable that an equal degree
of independence could be claimed for the three attitudes of mind as
evinced in less developed stages. Cognition, in the current acceptation
of the term, certainly carries with it, as part of its significance, recog-
nition of the antithesis between subject and object, an antithesis which,
I have tried to shew grounds for thinking, cannot be included as from
the outset among the facts of which there is consciousness. So, too,
feeling, as originally experienced could not involve the highly character-
istic feature which it afterwards acquires, — definite reference, namely,
of any particular state of feeling to the experiencing subject, a feature
which colours the connotation of the term as it is ordinarily employed.
Once more, the characteristics usually ascribed to conation, as essentially
involving the representation of an ' is to be ' in contrast with what
actually is, are obviously, I should say, not characteristics that belong
to the early experience of impulse and bodily movement. So far, then,
as the rudimentary states of mind are concerned, all we are entitled, I

^ Lectures on the Principles of Psychology in Vol. ii of the Development of Modem
Philosophy, 182.

2 The assumption really means, of course, that the relation between the act of appre-
hension and the external thing is simply being reduplicated in the relation between the
act of apprehension and its content. I agree with Professor Alexander (this Journal, iv.
244) that, as thus conceived, what is called content blurs the relation of act and that which
is apprehended. The content becomes, then, that which is apprehended, and, in order to
render suc/t apprehension intelligible, we should still have to regard it as also possessing
the aspect which I am denoting by ' content.'

12 The Nature and Development of Attention

think, to say is that they would contain within themselves the roots
from which these three diverging stems take their origin. One may
venture to lay down the general proposition that the nearer we approach
the beginning of consciousness, the less apparent will be the distinction
of cognition from feeling and striving. Increasing indistinctness of
presentation need not merge presentation into feeling, but in the early
life of mind any prominent difference between a sense-feeling and a
sense-presentation would seem to be precluded. The elementary com-
ponents of mind may, then, not inappropriately be described as in
essence acts of apprehension, the contents of which may be of varied
kinds, either those which come to be distinguished as sense-qualities,
or those which may come to be experienced as sense-feelings, or those
which become impulses or strivings and find a terminus in bodily move-
ment. Without, therefore, assigning to cognition, in the ordinary
significance of the term, a supreme position, it may, I think, be con-
tended that the functions which find their most complete realisation in
cognition are the common conditions of all phases of inner experience.
Feeling and striving may have equally with cognition their direct
antecedents in the primitive processes of mind, but it would be an
error to ignore their character as dependent upon acts of apprehension,
and to credit them with a substantive existence such as in no wise,
at any stage of development, really belongs to them. It is only as
based upon acts of apprehension that feeling and striving can be said
to be mental facts at all.

§ 4. Can we characterise further the acts of apprehension which
are thus conceived to constitute the very essence of mind ? I think we
can. With the fine discernment that often gives to his psychological
work a far-reaching suggestiveness, Aristotle insists, in more than one
section of the De Anima, upon regarding at(r0T}(rt<i not as mere passive
receptivity, but as a Bvvafii<: KpcTiKrj and each specific mode of ai<Tdrj<n<i
as a fieaov KpiriKov. In other words, Aristotle refuses to allow that
sense-presentations and their images can be rendered explicable with-
out some element of that discriminative activity which is usually
supposed to be the distinctive feature of the higher mental processes
alone. Sheltering myself here under an authority which no modern
psychologist need be ashamed to own, I venture to urge that dis-
crimination or comparison is involved in having even the crudest, even
the most indefinite, sense-content to which the name presentation
can be assigned. A very obvious objection to which a contention of
this sort is exposed may be at once forestalled. It is quite true that

G. Dawes Hicks 13

the terms 'discrimination' and 'comparison,' as ordinarily employed,
indicate a reflective operation of an intricate kind with the capacity for
which it would be ridiculous to endow a primitive mind. After dis-
carding, on account of their implications, various conceptions by the
aid of which the first beginnings of consciousness have been repre-
sented, it would certainly be a gross inconsequence, on my part, to
picture those first beginnings as in reality processes of thinking. No such
absurdity is intended. It is not meant that in order to be aware of a
sense-quality, ^, it is necessary that a conscious subject should perform
the mental operations expressible in the propositions " A is different from
B," or " A resembles (7." Far fi-om it. That would involve the use of
the abstract ideas of likeness and difference, — ideas which manifestly
are only possible for a self-conscious mind. But, in our own experience,
difierences and resemblances are recognised long before we have any
conception of difference or resemblance. Or to put it otherwise, A may
be distinguished from B, long before the precise features in which they
differ are appreciated. This more rudimentary process of differentiating
and distinguishing may itself be looked upon as exhibiting any number
of varying grades or stages, and I see no difficulty whatever in supposing
that in its crudest form discrimination lies at the root of the earliest
manifestations of consciousness. I, at any rate, can find no means of
realising what a state of mind could be which does not involve at least
some faint, confused, indefinite recognition of a content possessing a dis-
tinguishable character. If, then, every act of apprehension be described
as an act of discriminating, assimilating and comparing, no violence is
really done to the consideration of the relatively extreme simplicity and
narrowness of range that admittedly must be held to be characteristic
of the mental lives that are lowest in the scale of evolution. Whilst,
on this view, it is true that the general character of apprehension,
whether it be the first dim obscure strivings of sense or the highly
elaborate exercise of abstract reasoning, remains the same, yet the
specific modes of its exemplification are infinitely various, each stage of
its advance being dependent in the last resort upon increase of facility
in distinguishing and relating the elements presented in the world of
concrete fact in the midst of which a conscious life is lived. Ap-
prehension, so conceived, may certainly be said to be an ' activity ' ;
every phase of conscious life, as thus interpreted, is a state or condition
in which the subject is ' active.' But ' activity ' does not here signify
a mysterious unlocking of force or putting forth of energy; it points
rather to a certain aspect of what has been called discriminating and

14 The Nature and Development of Attention

comparing, — to its aspect, namely, as a process, or occurrence, in time.
I imagine, however, that, as applied to mind, the general term * activity '
is chiefly of service in negatively excluding misconceptions, such as that
of passive receptivity, than in positively throwing light upon the nature
of consciousness.

III.

§ 1. What explanation, on the basis of the foregoing account of the
structure of the primitive mental lifeS can now be ofifered of the process
of attention in the sense in which we have agreed to understand that
term ? The view which has been taken is clearly incompatible with the
notion of an abrupt introduction of any unique power or faculty into
the constitution of mind. It will not be as the effect of the exercise of
any special activity, but as a consequence of the conditions under which
the processes of the mental life develop that we shall seek to render
explicable the results which by general agreement are called results of
attending.

The common or essential characteristics, evinced in all processes of
attention, call first for consideration. They may be said to be these :
(a) a certain selection or limitation within the field of what is appre-
hended of some feature or features and the relative neglect or disregard
of the rest ; (6) the increased clearness and distinctness of the content
attended to, its greater vivacity and prominence, its more precise and
adequate appreciation ; and (c) the experiences in the form of feeling-
tone that sometimes precede and sometimes accompany the attitude of
attending.

§ 2. I start by dwelling on the last of these characteristics. It can
hardly be doubted — few psychologists, I think, ever have doubted —
that the earliest changes in the contents of apprehension which entitle
such contents to be described as facts of attention must be initiated by
differences in the intensity of feeling-tone accompanying some particular
content apprehended. A content possessing a higher degree than
others of painful or pleasurable feeling-tone would become naturally
differentiated from the rest; there would be a natural tendency for
other contents of a less pleasurable or less painful character to be
excluded from the sphere of consciousness. It has, indeed, been argued

' I have worked oat in detail what I have only cursorily outlined above in various
contributions to the Aristotelian Society; see Proceedings, N.S. vi. 271; viii. 160; and
X. 232.

G. Dawes Hicks 15

that to lay stress on this consideration is to overlook what is alleged to
be the fact that in our own experience pleasure or pain never comes until
after the object has been attended to, and that therefore such pleasure
or pain cannot give rise to the act of attending. We can, it is con-
tended, be neither pleased nor displeased by what is not yet in
consciousness \ But the argument misses the whole point of the
consideration, which, of course, is that a content may be ' in conscious-
ness ' before it is attended to, and unless that be admitted, it is difficult
to see how any psychological explanation of the circumstances we are
now concerned with is possible at all. I take it, then, that experience
in the form of feeling-tone can and often does precede the act of
attending in the strict sense of the term. At first, the variation in
the intensity of feeling-tone could produce no more than the purely
mechanical effect of detaining for a longer time than would otherwise
have been the case the content in question before the mind. By the
mere fact of its being thus singled out, so to speak, from among the
other transient contents of apprehension, its presence in consciousness
would be prolonged ; using, if one may, a spatial figure, it would tend
to spread itself over the field of consciousness at the expense of the other
contents, and within certain limits increase of duration would con-
tribute to producing the results for which we are trying to account.
For retention for a longer time in consciousness would in its turn give
rise to a two-fold effect. It would afford, on the one hand, a greater
opportunity for the establishment of associations between the content
in question and other contents apprehended by the conscious subject.
It would afford, on the other hand, a greater opportunity for easy and
rapid reproductions, in the form of re-presentations, or images, of the
said content.

As regards the former effect, one appeals to the familiar fact that
a certain duration of the sensory stimulus is necessary in order that an
act of apprehension should take place at all. I do not suppose that the
relation between duration of stimulus and completeness of apprehension
is by any means a simple one. In our own experience, for example,
duration of even a comparatively intense stimulus beyond a certain
period is unfavourable to apprehension. But there is no reason to
doubt that within normal limits, a longer duration of stimulus does
afford opportunity for a more complete comparison of the content with
other contents and of its more accurate discrimination from them. In
like rnanner, an increase in the intensity of the stimulus will have as

» W. B. Pillsbury, Attention, 56 and 287.

16 The Nature arid Development of Attention

its concomitant, besides an increase in the intensity of the content
apprehended to which I shall refer later on, a larger measure of feeling-
tone, and this, in accordance with what has already been said, would
also make for the results j ust indicated.

As regards the latter effect, every reproduction of the apprehended
content must obviously afford additional opportunity for increased
precision of discrimination. Each specific appearance of a revived
content would enable it to be marked off with greater definiteness of
outline from what was simultaneously present at the moment, and
every increase of definiteness of outline would render possible an
increased precision when a renewed content of like kind was appre-
hended. For one thing, the general setting of the latter would be no
longer identical with what it was on the occasion of the previous
appearances. And this new setting would serve as a means for insti-
tuting firesh comparisons, so that repetition would necessarily bring
about, although by no means in uniform fashion, increased clearness
and distinctness of what was revived or reproduced.

§ 3. There is, so far as I can see, no way of accounting for the
clearness and distinctness that constitute the principal result of
attending except by thus connecting the whole process of attending
with the elementary function of apprehension — with the act, namely, of
discriminating, of gradually recognising the marks whereby one content
is contrasted with or assimilated to another. As I have urged above,
clearness and distinctness would not be produced either by a mere
increase of mental energy directed upon a limited portion of the field
of consciousness, or by a mutual interplay of presentations, conflict-
ing with and reinforcing one another. The difficulty that confronts
the latter hypothesis is well illustrated in Titchener's treatment of the
matter. Despairing, apparently, of finding any means of solving the
problem in terms of the reciprocal action and reaction of presentations,
Titchener boldly cuts the knot by declaring that clearness is an inde-
pendent attribute of sensation, which, within certain limits, may vary
independently of the other concurrent attributes, such as quality,
intensity and duration*. It is one thing, however, to cut a knot and
another to untie it. On the one hand, Titchener is unable to doubt
that " some sensible qualities are, intrinsically, clearer than others," he
inclines to the view that "differences of clearness are, like intensive
differences, ultimate and distinctive," and he considers clearness to be
" an attribute of sensation, conditioned upon nervous predisposition,

1 Op. cit. 171 sqq.

G. Dawes Hicks 17

just exactly as quality is an attribute of sensation, conditioned upon
nervous differentiation." On the other hand, he admits that one and
the same sense-content may vary in clearness, according to the degree
of ' concentration ' of attention, he places novelty, rarity, unaccustomed-
ness among the conditions of clearness, and he recognises the great
importance of the associative relationship " between the sensation and
the whole circle of ideas dominant at the moment " as determining the
clearness of the former. I know not how these two portions of the
theory are to be wrought together into a coherent view. If clearness
be an ultimate and intrinsic property of sense-contents it will hardly do
to speak of clearness as purely relative in character, dependent on a
number of conditions extraneous to the content of which it is a property.
I confess I am unable to attach any intelligible meaning to the former
of these contentions. It is evidently based upon the assumption that
isolated sense-contents are already given, prior to apprehension, endowed
with all the characteristics which are described as attributes, clearness
among the rest. The rejection of that assumption leads, it seems to
me, directly to the view of clearness upon which I have been proceeding.
If conscious experience does not start with a multiplicity of sense-
presentations, each definitely marked off and separated from the rest,
but with a vague, ill-differentiated whole, out of which by successive
acts of discriminating there gradually emerge definite presentations,
then it must follow that the clearness and distinctness of such presen-
tations will depend upon the number of marks or features which the
conscious subject is able to discern in each, and it will only be in and
through the process of discriminating that the presentation will have
clearness, or indeed any individuality of character at all. A sense-
content, in short, increases in clearness just in proportion to the number
and kind of distinctions that contrast with or resemblance to other
presentations has enabled the discriminating subject to constitute.

§4. Another factor calls for notice, a factor, moreover, that will
be seen likewise to make for the result we have had in view. I refer to
the intimate connexion of elementary sense-experiences with bodily
movements. Every stimulation of the sensory mechanism has mechani-
cally connected with it a certain stimulation of the motor mechanism ;
the junction between the two is so close and uniform that hardly any
change in the one fails to find a response in the other. It is with the
elementary experiences, hard to discover psychologically, which precede
and accompany the execution of movement that we have here specially
to do. The physiological conditions of sense-apprehension themselves

J, of Psych. VI 2

18 The Nature and Development of Attention

evoke to a greater or less extent certain efferent discharges causing
muscular contractions as a natural fact. With recurring sense-contents
more or less resembling one another, there will come, therefore, to be
associated, though in various degrees, the experiences that have accom-
panied the movements thus initiated. And in accordance w^ith the
character in part of the sense-contents and in part of the movements
associated with them will be the prominence of the former in conscious-
ness. The sense-contents accompanied with a large quantity of the
experience of movement will naturally retain a more prominent place,
and if they incite at the same time feelings of pleasure or more specially
of pain, and if they give rise to motor experiences that are related to
the continuance of pleasure or the removal of pain, they will secure
relatively the largest place in consciousness. Thus I conceive that in a
general way we can understand how it is that with the characteristics
of attention, the retention for a longer period in consciousness of an
apprehended content and the increase of clearness and distinctness of
that content which ensue through its being compared with and related
to other contents, there should naturally come to be associated those
experiences of strain or tension which manifest themselves so pro-
nouncedly when will, in the strict sense of the word, has been developed,
and the conscious subject can control series of movements or even trains
of thought.

§ 5. So far I have been dealing with what can appropriately enough
be called primary attention. I pass now to secondary attention, if one
may adopt that term for the process as it is exhibited at the stage when
a definite consciousness of a world of external objects has been attained,
but which is still more or less now voluntary in character. I think it
will be found that the principle upon which I have been proceeding is
adequate here also to render a satisfactory explanation of the facts. A
reference to some familiar experiences will aid us in carrying on our
inquiry.

In the first place, I point to the well-established fact that certain
limits are imposed by the range of experience already possessed by
a conscious subject upon the kind of objects which can be for him
possible objects of attention. In other words, the same fact presented
to differently equipped minds will not be attended to in the same way.
The objects of sense-perception, to which for the present we can confine
our inquiry, exhibit a great diversity of qualities, and some of these
qualities seem inevitably to force themselves into notice, to attract or
engross, as it is said, attention. Often this feature of attractiveness or

G. Dawes Hicks 19

fascination on the part of an object has been regarded as the dis-
tinctive feature of so-called non-voluntary or automatic attention in
contradistinction to voluntary or deliberate attention. But when once
allowance is made for the circumstance just alluded to, the characteristic
ceases to be one th^t can be fixed upon as differentiating, in any
decided manner, the two modes of attention in question. The features
or aspects which * attract ' or ' engross ' the attention of one conscious
mind will not, and we may safely say cannot, ' attract ' or ' engross '
the attention of another mind which we will describe, in popular
language, as "differently constituted." Whilst halting on a certain
occasion in front of one of the most beautiful views obtainable of
Grasmere and its hills, a lover of the Lake-country was accosted by
a Manchester tripper, and had to find an answer to the query, " Is there
anything to see in this place ? " The proverb, the eye only sees what it
brings with it the power of seeing, needs no better illustration. The
limitation to which the proverb refers depends psychologically upon the
possibility of recognising in the new material presented features capable
of being assimilated with the stock of ideas, the particular trains of
experience, already possessed, so to speak, by the individual subject.
For the most part we attend, to use the orthodox mode of expression,
only to that which ' interests ' us. Interest, however, is not, as one
might be tempted to imagine from much that has been written about
it, an unique and unanalysable relation between consciousness and its
object. The term is but a convenient symbol for indicating the complex
set of conditions to which I have been referring. " It is quite certain,"
says Mr Bradley, that interest "consists to a large extent in pleasure
and pain\" and I have no desire to call in question that assertion. But
I think it chiefly of importance to emphasize the consideration that the
statement, " what does not interest will not be attended to," amounts to
saying that whatsoever is not brought in some way into connexion with
the previous experience or past life of the subject will not ' attract ' or
' engross.' It is often maintained that there are certain presentations
of so vivid, so impressive, so striking a character that they cannot fail
to elicit recognition, that they force their way mechanically, as it were,
into consciousness in spite of any subjective tendency to oppose them.
This, however, is, by no means, unreservedly true, or, indeed, true at all,
as thus expressed. One need not appeal to exceptional instances, such
as that of Hegel writing the last pages of the Phdnomenologie within
hearing of the "thundery of Jena." Ample confirmation could be

1 Mind, XI. 310,

2—2

20 The Nature and Development of Attention

furnished from every-day life of the fact that so far as strength or
intensity can, in this reference, be measured or spoken of in an absolute
sense, the strongest, the most vivid, presentation may not succeed in
arousing attention, still less in becoming clear by virtue of its intensity.
To be attended to, even such presentations must in some way be
assimilated with contents that form, or have formed, part of the subject's
experience; they can only be attended to when the conditions of the
inner life do not prevent it. And these reservations rob the contention
of any significance it might be supposed to possess as militating against
the truth of the principle which I am emphasizing.

In the second place, another circumstance may be referred to that
will likewise be of service. The range of attention, the kind of relations
which can be simultaneously apprehended, will vary with the successive
stages in the development of mind. And not only the range of atten-
tion. There will be variation also, although not capable of exact
numerical determination, in the complexity of the facts that can, at any
moment, be attended to^. In the earlier stages of mental growth,
attention can only be given to relatively simple facts. What can be
held together in one and the same act of attention is, in these earlier
stages, comparatively meagre in character, formed into one whole by
relatively few and for the most part external links of connexion. In
a mind of a maturer type, facts of far greater complexity will be
simultaneously attended to, the field of simultaneous attention will be
of far greater extent. So that here, again, and from another side, we
are led to our former result that capability of being connected with the
trains of ideas and feelings, or, more generally, with the experiences, of
the conscious subject is an essential requisite of anything being attended
to. Meanwhile we are considering the case not of attention which
arises deliberately and through resolve, but of attention that occurs on
account of the 'interesting' or ' attractive ' qualities of what is presented
in perception. A novel or striking occurrence, let us say, impels us to
notice it. At first sight, that may seem to conflict with the principle
we have been laying down. But the opposition is apparent only. The
absolutely unknown does not interest us. That which is new will not

1 It is worth remarking, perhaps, upon the totally misleading problem which has been
set, in this connexion, for experimental solution, — that, namely, of determining how
many objects can be attended to at once. What are we, in such a reference, to understand
by one object? Every object is complex, comprising a multiplicity or manifold of features.
It is this complexity that is the vital matter so far as the range of attention is concerned ;
the mere number of what are arbitrarily taken to be single objects is a very subordinate
consideration.

G. Dawes Hicks 21

be entirely new ; to be recognised at all it will have to be connected in
some way with what is already part of our experience. Such links of
connexion may be small or large in number, and the number will
depend to a considerable extent upon the stage of our mental develop-
ment. If the number be small, the object will be comparatively
obscure. A young child's astonishment at the sight of a locomotive
will not, for example, secure clearness and distinctness of apprehension.
The indispensable condition of clearness and distinctness, namely, rapid
and easy assimilation with what past experience has prepared, would, in
that case, be wanting.

To recapitulate, — along the path we have been following, we
seem to be on the track of a perfectly general law of mind. An
object may be apprehended with varying degi'ees of definiteness,
distinctness, and clearness, and every increase of definiteness, distinct-
ness, and clearness is equivalent to recognition of an additional dis-
tinguishable mark or characteristic in the object in question. The
possibility of recognising such distinguishing marks is conditioned by
the variety of ways in which the object can be brought into connexion
with elements with which it can be compared, contrasted, and related.
It is, in mature experience, a sufficiently familiar circumstance that the
character of what is observed depends largely upon the sum of acquired
ideas which the observer can bring to bear upon it, and in this
circumstance is to be found the explanation of the maiii features of
what we have called secondary attention. An object will become for
us definite and precise, in so far as we have the means of instituting
a comparison between what is immediately given and what we are
already aware of as the accumulated product of previous experience.

§ 6. If the principle which has so far guided us be sufficient to
account for the phenomena of attention hitherto considered, there is
certainly a strong presumption that the special features that call for
explanation in the higher form of attention, known as voluntary or
deliberate, will exhibit themselves as following naturally from the
varying circumstances that make their appearance in the coui-se of
mental growth and development. And I believe that, in truth, this is
the conclusion which a careful examination of the facts does enable us
to reach. The activity of discriminating, comparing, and relating works,
as we have seen, through means of the material freshly offered being
brought into connexion with material that has been already supplied to,
and discriminated by, the apprehending subject. Obviously, then, the
direction taken by this activity in the successive stages of its history

22 The Nature arid Development of Attention

will be largely influenced by any important distinctions that may
gradually disclose themselves among the material which we have
described the subject as possessing. One such important distinction —
it is not too much to say, the most important distinction — which thus
comes gradually to recognition is that indicated by the terms self and
not-self. By degrees in the development of intelligence there is effected
a definitely recognised separation between the trains of thoughts,
sentiments, feelings and sense-presentations which are more or less
constant and habitual, and which thus come to be regarded as consti-
tuting the prevailing centre or background of individual personality,
and the relatively transient presentations and apprehended contents
wliich come and go, and which the subject learns to contrast with and
to distinguish from the totality of the former. The contents of our
knowledge or experience, or rather certain of them, tend more and
more to wear the aspect of an inward possession, and to become the
instrument, as it were, by which we apprehend the world of objective
fact. So soon as this distinction has attained any prominence in
consciousness, it must of necessity influence in a very decided manner
the direction, as we may put it metaphorically, of attention. For it
will then become possible for the subject to differentiate between the
cases where attention comes about through a presented object being
connected with the contents of representations or ideas that are not
specially included in the consciousness of self, and the cases where the
activity of comparing and relating is carried on through means of those
ideas and feelings which are included.

The consciousness of self has by no means the same content in all
phases of its development. At first that content would be comparatively
poor and meagre in its features, and would be mainly dependent upon
what later it never entirely loses, the mass of vague sense-experiences
and feelings that arise from the vital processes of the body. Now the
sense-experiences which arise as the concomitants of bodily movements
are pre-eminently of the motor type — those presentations of tension or
strain (Spannungsempfindungen) that are the ways in which muscular
contraction is apprehended. Comparatively early in the evolution of
animal forms, bodily movements begin to lose their originally chaotic
and random character, and tend to fall into regular ordei-s and groups.
And the presentations of tension or strain are not, therefore, appre-
hended separately or in isolation ; they form parts of a continuous series,
which is experienced more or less as a whole, and, what is of no less
importance for our present purpose, is revived or reproduced as a whole.

G. Dawes Hicks 23

By way of illustration, reference may be made to our own experiences
when engaged in overcoming resistance, say in lifting a heavy weight.
Presentations of tension or strain come to us from all parts of the body,
not merely from the muscles that are directly concerned. But these
presentations are experienced as one complex mass, and it is easy to see
how this complex mass comes to be looked upon as having a single
and independent existence of its own, which is taken to be the cause
of the movement instead of being, as it actually is, its concomitant and
consequent.

The truth is that the conditions which are involved in the rising
into consciousness of the distinction between self and the objects of the
external world coincide for the most part with the conditions under
which there comes gradually to be formed what may properly be
described as the individual will of the apprehending subject. The
conditions which are concerned in the formation of the will, in this
sense, are, one may safely assert, very largely bodily movements, or
rather the experiences which are the concomitants of bodily movements.
Certainly the conscious subject's control over the movements of the
body is an acquired control, and cannot be supposed to be, in any sense,
an original acquisition of conscious life. If we assume the presence, at
the outset, of rudimentary impulses as the primitive germs of conation,
the temptation must be resisted of conceiving of the movements that
follow from such impulses as in any way foreshadowed or prefigured in
the impulses themselves. Looked at from the standpoint of an indi-
vidual conscious subject, the connexion between specific presentations
or re-presentations other than motor and specific bodily movements is
an arbitrary connexion ; the connexion was not, I take it, primordially
revealed to consciousness through some innate mode of presentiment.
Only through experience, by the aid of association, can the conscious
subject have come to connect the primitive impulses with definite
objects, and, as a consequence, the impulses have come to assume the
more complicated form of desires. A desire, in the ordinary sense of
the term, involves the representation in idea of what is desired, and the
possibility of forming such a representation obviously depends upon
there having been experienced a long series of prior presentations and
the satisfactions which they afforded. Or, to put the matter otherwise,
there is implied, in the state of desiring, some recognition of the
distinction between the real and the merely imagined or ideal, — recog-
nition, in other words, in however obscure a fashion, of a relatively
independent order of facts over against which the conscious subject
stands opposed. Before anything like a normal correlation came to be

24 The Nature and Development of Attention

established between desires and the means of realising them, there
must have preceded endless experiences of executed movements which
were not thus conditioned. Only through experience could a conscious
subject obtain the data which enable the discrimination to be made
between what simply occurs and what occurs in consequence of a repre-
sentation on its part. Only through experience could there be formed
in consciousness the representation of that which will yield satisfaction
to the subject. Only through experience could such a representation
become so welded together in consciousness with specific bodily move-
ments as to secure its realisation. Both the control of movement and
the realisation of an idea or representation depend, therefore, upon the
establishment of empirical connexions between certain phases of the
inner life and certain modes of the bodily organism.

Regular groups or series of movements come by degrees to be
associated with specific re-presentations or ideas, so that on the occur-
rence in consciousness of the latter the former tend inevitably to be
re-instated. The actual execution of the movement or the actual
realisation of the idea, if the latter involves change in the external
world, is wholly an affair of the bodily mechanism, and the details of
that mechanism are completely hidden from the conscious subject.
Though the action is inspired by conscious purpose, consciousness
obtains no information about, nor does it direct, the intricate adjust-
ments through which the action is carried into effect. We exercise
voluntary control over our movements by dwelling on the object
immediately before us in conjunction with the idea of what we desire
to accomplish, and the bodily movement supervenes in accordance with
natural law. In short, the realisation of what is represented in idea, as
also the voluntary control of bodily movements, is itself, as a psycho-
logical fact, the result of attention ; such realisation and control are
only possible through the elements of a present situation being dis-
criminated fi'om and contrasted with the elements of a contemplated
end or purpose.

Now, the very circumstance that attention, in the form denoted
above by the term secondary attention, thus lies at the root of any
control we can exercise over bodily movements points of itself to the
source of those experiences of effort or of activity which are frequently
so prominent in voluntary or deliberate attention. Careful analysis
has shewn that in the experience of effort, even of the effort described
as ' intellectual effort,' muscular factors of diverse and varied kinds are
invariably brought into play, and that kinaesthetic presentations
and re-presentations invariably form part of what is then experi-

G. Dawes Hicks 25

enced'. Furthermore, it is not difficult to understand how the feeling or
sense of effort should come to wear the aspect of self- activity. For motor
presentations as being dependent upon subjective impulses and as being
likewise extremely uniform in character, as contrasted with the visual
and other presentations to which they lead, naturally come to be con-
nected with the consciousness of self in the closest and most definite
way. But to conceive of such effort as self-activity in the sense of
being itself the cause of the process of voluntary attention is, it seems
to me, an error similar to that of supposing that the sensations of strain
or tension experienced in lifting a heavy weight are themselves the
causes of the bodily movement.

§ 7. That the self-conscious subject does exercise control over his
bodily movements and over his trains of thought, it is far from my
purpose to call in question. The mental activity, however, involved in
such control is, I should maintain, the activity which is involved in
apprehension, in various degrees, throughout, and which cannot legiti-
mately be identified with ' effort ' as an apprehended fact. Mental
activity, surely, may be one thing, and the activity apprehended in
' strain ' or ' effort ' another, and one need not be thought to be denying
the first because one fails to find it specially manifested in the second.
According to the view I have been taking, every state of mind, whether
the consciousness of effort be an ingredient in it or not, is a state of
mental activity ; that activity may be involved no less decidedly in a
sense of ease than in a sense of strain. In listening to one of Wagner's
operas or in following an abstruse argument, the mind may be in-
tensely active and yet the consciousness of effort may be at a minimum.
On the other hand, when effort is declared to be the " distinct con-
sciousness of opposition between what we call self and muscular
resistance " (Baldwin), one seeks in vain to understand how the muscles
can offer opposition to the self. Further, as Dewey points out,
muscular resistance, whatever else it may imply, must involve as a fact
of consciousness motor presentations, and if the self is actually exerting
force that too must find expression in motor presentations, so that the
factors it is sought to dispense with would have, after all, to be admitted.
In contradistinction to a theory of this sort, I have been trying to shew
grounds for thinking that the distinction between non-voluntary and
voluntary attention indicates in truth differences in the stages at which
the process is viewed, and is explicable when account be taken of the
conditions under which the mental life develops.

^ See, for example, the experiments referred to by Dewey in his article on " The
Psychology of Effort," Philos. Rev. vi. 43.

THE PSYCHOLOGY OF VISUAL MOTION.

By henry J. WATT.

/. Criticism of Wohlgemuth's physiological theory.
II. TJie introspective nature and affinities of the after-effect of
seen movement.

III. The correlation between the introspective features of the

after-effect and those of the previous objective movement.

IV. Wertheimer's criticism of certain psychological theories.
V. The present theoretical outlook.

Two elaborate studies of visual motion have recently been published
by A. Wohlgemuth^ and by Max Wertheimer^. Both of these im-
portant papers add much to our knowledge of the facts, and excel in
clearness and precision of work. They are also alike in rejecting
all the psychological theories that have been advanced in their several
fields of research and in formulating a physiological theory in explana-
tion of the facts. Neither writer, however, makes any contribution to
psychological theory. The possibility of such a thing is hardly even
suggested ; it is presumably annulled by the mere offer of a physio-
logical theory. But the matter is not debated.

This situation seems to me so anomalous as to be worthy of special
notice, the more so as the facts of the case hardly warrant the attitude
adopted by these writers. In this paper I propose to deal briefly with
the theory and outlook of these two works, which may be considered
typical of a certain trend of opinion prevalent at the present time. In
view of their general importance, however, and for the sake of brevity,
I shall assume for the most part that the reader is already familiar
with them and need only be reminded of their contents as each point
arises.

^ "On the After-effect of Seen Movement," this Journal, Monograph Supplement,
No. 1.

2 " Experimentelle Studien iiber das Sehen von Bewegungen" (Habilitationsschrift,
Leipzig, 1912), Ztschr. f. Psychol, lxi. 161 fif.

Henry J. Watt 27

I. Criticism of Wohlgemuth's Physiological Theory.

The after-effect of seen movement, which is the object of
Wohlgemuth's investigation, is familiar in various natural situations.
If fixedly we gaze at a streaming waterfall or look down upon a
rushing river for half a minute or so and then turn to look at the
ground, the latter will seem to be streaming in a peculiar manner in
the direction opposite to that in which the water flowed (relatively to
our field of vision). For experimental purposes a simple form of this
process is devised. A sheet of paper bearing alternately black and
white lines of some little breadth is fixed upon a drum, which is
rotated so that the lines move across the field of vision more or less
slowly in a direction perpendicular to their length. Under suitable
circumstances the lines will appear to move backwards when the motion
of the drum is stopped. As the eye has been fixed and steady all the
time, this peculiar after-motion cannot be due to any motion of the eye
after the stopping of the drum, but must be taken as the after-effect
of the preceding motion. Some theorists have therefore supposed it
appeared because we were deluded by the previous objective movement
into being accustomed to motion and therefore into expecting motion
for a longer time than it was really there and thus into seeing what we
expected. But this can easily be disproved by the application of
incognitive methods, which prevent us from knowing from time to time
what really happens. If the same after-effect follows whether the
observer knows what is really happening or not, it cannot be the result
of an illusion of judgment.

If the mind does not work at all to produce this after-effect, then
apparently the only task for theory is to extend the accepted notions
regarding the general physiology of neural processes so as to cover the
facts ; or to imagine a neural mechanism which will shew why motion
is sometimes perceived where nothing really moves and why it then
runs in a certain direction, opposite to that of the preceding movement.
For his theory Wohlgemuth assumes that retinal elements al and a 2,
61 and 62, are each connected with a "subcortical centre of movement,"
consisting of summation cells Al and A2, Bl and £2, and also in
pairs with a Schaltzelle 81 for the a's and 82 for the 6's. J. 1, A2, and
81 are also connected with one another, as are Bl, B2, and ^2.
Impulses are sent by ^ 1 and A2 to the cortex, but this system of
centres of movement is independent of other centres, e.g. those for
brightness, colour, local sign, etc.

28 The Psychology of Visual Motion

(a) Wohlgemuth assumes (pp. 99 ff.) that in the hypothetical centre
of movement, owing to the part played by the Schaltzelle SI, a state of
facilitation lasts in Al, so long as the objective movement stimulates
the eye, but that as soon as this movement is stopped, the state of
facilitation in ^ 1 is replaced by a state of fatigue in A 1. By this
means, during the objective movement, ^41 is more excited than A2,
while during the after-stage A2 is more excited than Al. The
psychical counterparts of these relations of intensity are, for the
former, movement having the direction A 1 — A 2, for the latter,
movement in the opposite direction.

Now this assumption posits the unfailing occurrence of so special a
case that it seems to me to vitiate the whole theory. We should rather
expect many possible relations between facilitation and fatigue :
facilitation frequently still increasing with psychical counterpart of
similarly directed movement and after-effect, occasional balance with
no visible movement or after-effect, frequent fatigue after longer
stimulation with a reversal of both seen movement and after-effect,
and thereafter periodic return to a state of balance. But, as we
read on page 85, " no after-image of the sectors moving in the
same direction as the objective movement could at any time be
detected."

(b) The theory offered virtually begs the question. For, in order
to suppose that the physiological basis of the experience of pure
motion exemplified by the after-effect is a difference of excitation
amongst the cells Al, A2, etc., and that the physiological basis of the
direction of the felt motion is the spatial distribution of this difference
of excitation amongst the cells Al, A2, etc., it must assume that the
cells Al, A2, etc., already function as the physiological basis of
different localisations, and that real directions within the complex
of cells Al, A2, etc. (with or without actual physiological connexions
between these cells) form the basis of felt directions ; or it must assume
that the cells Al, A2, etc., individually and as a complex, are con-
nected and correlated with those other centres that are the physiological
basis of localisations and directions. In either case the theory takes
the physiological basis of localisation and direction for granted and
only offers a theory of motion, treating it as a sort of intensive state,
which refers and is attached to these localisations and directions, and
endows the " sukzessive Aufspringen eines gleichartigen Eindruckes an
verschiedenen Orten " — which we might perceive merely as such, were
we beings devoid of the peculiar experience of motion — with this

Henry J. Watt 29

unitary quality of continuity, namely " ein Hindurchgehen durch die
zwischenliegenden Raume\"

But can motion really be treated in this way ? Is it not rather the
case that motion has a direction of its own, which may coincide with,
or be opposed to, some other direction of which we are conscious apart
from any motion ? Is not also the velocity of a motion a characteristic
of its own ? Are not the motion, the direction, and the velocity, of
motion — whether it correspond to a real motion or be pure motion in
the sense of the after-effect — the essential aspects of this experience,
its vividness being necessary in some degree, but as such relatively
unimportant ? The vividness of the experience may be to some extent
interchangeable with its velocity, in so far as an increase in velocity is
accompanied by an increase in vividness ; but surely it would be
contrary to experience to allow this vividness to usurp the place of the
velocity of the motion itself, not to speak of its direction. And if
motion presents a continuity that is not given in, or derivable from, the
data of our space and time Anschauungen, should we not expect to find
an explanation of this continuity included in the physiological theory
of motion ? But it is evident that this continuity is taken for granted
in the theory as stated.

It seems then that the theory in question offers an explanation
really only of the vividness of the experience ; and if against this must
be written the arbitrary assumption which I have stated under (a), the
balance leaves nothing to the credit of the theory. We must discover
first of all what is the neural basis of pure motion, its direction, and
its velocity ; it will hardly be very difficult thereafter to find a basis for
its vividness.

On page 19 Wohlgemuth says that Borschke and Hescheles admit
that, as seen, the movement of two sets of straight rods at right angles
to one another " can only be regarded as squares, moving in an oblique
direction." This movement can, of course, be described as one pleases;
but if it is felt as the movement of squares in an oblique direction,
that must be due to psychical, or shall we say, central, reasons; for
it is essentially the Gestalt of the square which determines the
apparent movement. If a point on one of the rods were marked out
by colour or shape, we should at once in so far be free from this
apparent oblique motion of the squares^ In connexion with this the

1 Cf . Ebbinghaus, as quoted by Wohlgemuth, op. cit., p. 108.

2 Cf. Pleikart Stumpf, "Ueber die Abhangigkeit der visuellen Bewegungsempfindung
und ihres negativen Nachbildes von den Beizvorgangeq auf der Netzhaut." Ztschr,/.

30 The Psychology of Visual Motion

forced explanation given of the results of experiment 28 on page 107
should be consulted.

(c) The theory constitutes, as it stands, a lapse from the presumable
parallelism of mind and body; it fails to shew that the relations of
mind and body, whatever they may be, follow any general scheme
or plan ; in fact, it suggests that they vary arbitrarily from one
experience to another. For all would agree, I think, that the neural
basis of the an-angement of the simplest sensory experiences in
respect of their adherent localisations is, proximately or ultimately,
the arrangement of neural units of some kind. Of course, we should
not expect to be aware of the experiences correlated with these neural
units, nor of their localisations, apart from some degree of excitation in
these neural units. But neither should we expect to find that the
essential aspect of their stimulation, with which alone experience is
correlated, is the difference of excitation in them. For even if differ-
ence of degree of excitation were a necessary feature of the neural
basis of the experience of motion, and of its direction and velocity,
these experiences must first and foremost be correlated with the
arrangements and interconnexions of the neural units and only
secondarily with their difference of excitation. Difference of excitation
would, then, be only a means of bringing different localisations with
different clearness and insistency to the mind.

Thus we might revert to the simple theory of common sense and
expect motion to be based upon the successive stimulation of neural
units correlated with different positions. And it is to be noted that we
have as yet no evidence that bears against this view or shews that the
effect of motion is producible from simultaneously stimulated neural
units, be they stimulated equally or differently. The facts of the after-
effect of seen movement do not, of course, afford this evidence. They
offer no other evidence than do the ordinary facts of motion. It is only
in the eyes of such a theory as Wohlgemuth's that the stimulation of
the neural units subserving motion is simultaneous and different.
[When the stimulation of the elements of a neural complex in
different degrees is said to be simultaneous, that means, of course, for
Wohlgemuth as for others, simultaneous and continuous over a short
stretch of time.] But he extends this explanation not only to the

Psychol. Lix. 324 : " Im Vorbild stimmte die gesehene Bewegung nur nicht dann mit der
berechneten Richtung iiberein, wenn irgendwelche Anhaltspunkte andere Auffassungen
begunstigten," etc. Compare the effect of using broken lines and spirals, where the seen
movement always corresponds to the objective movement.

Henry J. Watt 31

after-effect of motion, but to ordinary visual motion. His hypothesis
thus stands in sharp contradiction to the facts upon which it ultimately
rests, that is, both to the facts of experience and to the facts known
regarding the elements of the peripheral stimulation and their relative
qualities, intensities, positions, and times. The positions of these
elements differ in different times, so that in the several neural units
stimulated by them, at least in those proximate to the stimulation,
there must necessarily be successive differences, be they differences of
intensity or of quality or of both together. Is it not, then, most
reasonable to suppose that whenever motion is given, these successive
differences occur throughout all the elements of its neural basis, be they
proximate or remote ?

Therefore it seems that Wohlgemuth's physiological theory of
motion fails to shew that the relations of mind and body follow any
general scheme or plan ; or if it does so implicitly, it places a false
emphasis on the part played by the intensive differences of neural
processes in the correlation of mind and body.

II. The Introspective Nature and Affinities of the
After-effect of Seen Movement.

The first task of psychology seems to me to be a thorough study of
all distinguishable varieties of experience and their arrangement on the
basis of their resemblance to one another, whether the resemblance be
that of appearance {e.g. of attributes) or of functional properties and
variations. We must form a periodic table of experiences, as it were,
and we must take that table as the basis and object of explanation of
every theory which is to be called psychological.

From Wohlgemuth's valuable historical and experimental researches
it appears that the after-effect of seen movement has the following
characteristics or properties^ : —

01 The after-effect is an apparent movement, in a direction

opposite to that of the previous objective movement.

E 16 Its velocity is comparable with that of an objective

movement.

E15 Its velocity acts as a velocity. It adds itself to an objective

movement.

1 C refers to conclusions by agreement between Wohlgemuth and his predecessors,
E refers to Wohlgemuth's own experiments (the numbers are those of his text), cf. pp. 110 fit.

32 The Psychology of Visual Motion

j^33 Certain observers mistake it for a real objective movement

and are unwilling to believe the contrary (p. 87).
EIQ — 13 It varies in vividness (cf. pp. 46 ff.).
03 It is definitely localised.

C4 It has a definite position in time.

These are its positive features. Negatively it appears that : —
E 32 It is not like "a shadow passing across the stationary surface."

J^33 As compared with an objective movement it has a hollow

ghost-like appearance. Or it may have all degrees from
reality to evanescence and ghostliness. As Wohlgemuth
observes it, it is an experience sui generis. For him it
never approaches the appearance of real objective move-
ment. It lacks the solidity and reality that is given
by change of position in space (cf. pp. 87 f,).
We must, therefore, conclude that, no matter how unusual the
isolation of pure motion in the after-effect may be, nor how "unreal"
it looks, it does greatly resemble its prototype of objective movement,
fusing with the latter both phenomenally and functionally.

III. The Correlation between the Introspective Features of
THE After-effect and those of the previous Objective

Movement.

The next question is whether the introspective kinship thus
established is confirmed or contradicted by the evidence regarding
the correlation between the introspective features of the after-effect
and those of the previous objective movement upon which it is
dependent. With which feature or features of simple sensation is the
after-effect objectively connected ? We may pass in review the chief
attributes of sensation : (a) quality, (6) intensity, (c) order (local sign),
{d) position in time, (e) extensity, and (/) duration. Of these, however,
only the first four really come into question.

(a) Quality. The relevant facts are these : —
E 17 — 18 " The after-effect is independent of the quality of the

light." The latter may be varied without variation of

the former. Cf. £'19 below.
^29 "Fatigue produced by alternating movements of opposite

sign is independent of the colour of the light producing

it, i.e. the fatigue is maintained in light of different

colour."

Henry J. Watt 33

El, (711, and E2 — 4 may also be cited, which shew the manner in
which bhe clearness and vigour of the contents of the visual field
reinforce the vividness of the after-effect. The after-effect is also
noticeable in the dim field of subjective vision (eyes closed).

The conclusion, then, must be that the after-effect cannot well be
produced apart from quality of some kind, but it is independent of
the variation of the quality, as such, of the light. It is presumably
produced by a factor which accompanies quality and which becomes,
to some extent, more insistent as quality becomes more insistent.
The explanation which Wohlgemuth offers of Ell — 18 that "each
new colour is a new stimulus" (p. 106) hardly seems consistent with
E2d. Wohlgemuth's theory may explain the latter, but it can hardly
explain the fornier. Wohlgemuth himself seems to feel this difficulty
(cf. pp. 107 and 109).

(6) Intensity.

EIQ "In the case of different colours difference of brightness is

not essential for the production of the after-effect."
E14! "If a moving series of alternating dark and light stripes

excite the retina, a slightly better after-effect seems to
be obtained if the stripes be of equal width; but if the
alternate dark and light stripes be not of equal width
it seems not to matter which stripes are increased and
which decreased in width."
E2 — 4 "The after-effect is more marked in a brightly illuminated

objective field... than in a darker field."
Eo — 6 "If during the passage of images over the retina, a
stimulus of a given intensity alternates with one of
less intensity, the after-effect of movement produced
is more vivid than if such stimulus alternates with a
(more or less complete) cessation of stimulus."
01 "The uniform passage of light stimuli over the retina in

any given direction... produces the after-effect."
The decisive case is ^ 19, which shews that a variation of intensity
is not an essential condition. E14! is only compatible with Wohl-
gemuth's theory if the special assumption discussed under la is
admitted. The other results, along with those referring to the
difference between the light- and the dark-adapted eyes, are con-
comitant variations, which may depend not only upon the variations
of intensity, but upon that of one of the other attributes. The
J. of Psych. VI 8

34 The Psychology of Visual Motion

explanation of E o—Q which Wohlgemuth gives (p. 104) seems
strained: "When a black stripe succeeds a white one the synapses,
which had been fatigued, immediately regain their former state."
But the main theory supposed these synapses to be in a state of
facilitation. If they are fatigued, A 1 should be more fatigued than
A 2, having been excited more strongly longer, and the movement
should have turned apparently to the direction opposite to that which,
it shewed at first.

The conclusion then must be that the after-effect cannot well be
produced apart from intensity of some kind, but that it is independent
of the intensity, as such, of the light. It is presumably produced by
a factor which accompanies intensity and becomes to some extent
more insistent as intensity becomes greater. If the after-effect is to
be got, moreover, either the quality or the intensity must be varied.
Both of these may, but need not, be varied at once. The after-effect,
therefore, cannot well be dependent upon either of these attributes,
but it may be dependent upon a factor which changes with differences
in either or both of these^

(c) Uniocular order (local sign). ' T '

C 1 Quoted above.

C2 "This after-effect is more marked if the eyes... remain

fixed on a stationary point."
(78 "The after-effect is producible by any rate of the stimu-

lating movement."

1 It is necessary to refer at this point to the preliminary notice of experimental results
issued by Pleikart Stumpf, in which he says: "Es zeigte sich namlich zunachst die
auffallende Tatsache, dass bei sukzessiver Helligkeitsanderung einer Farbe des einen
Farbenpaares sich eine Stelle finden lasst, bei der der Bewegungseindruck in den meisten
Fallen vollkommen verschwindet, oder in einigen besonderen Fallen doch ein Minimum
an Deutlichkeit erreicht. Zu jeder Farbe lasst sich auf diese Weise ein bestimmtes Grau
finden, das mit ihr, so miissen wir wohl annehmen, einen unwirksamen Erregungsiiber-
gang bildet, so dass kein Bewegungsempfindungsprozess mehr zustande kommen kann "
{op. cit., 328 f.). The grey is that which gives the lowest fusional frequency with the
colour concerned. If Stumpf s observations are correct, their inconsistency with those of
Wohlgemuth may be the result of the difference of method adopted. Stumpf's method is
essentially stroboscopic and his bands of colour are very narrow — two millimetres. No
account is given, however, of the means of obtaining the necessary variation of brightness
in the grey bands, which to give the result stated must have been most laborious. Until
full details are given, Stumpf's result must be held in suspense. In view (1) of the
restriction of a "minimum in some cases" which he indicates, (2) of the absence of any
reference by Stumpf to differences of velocity, and (3) of the cumulative effect of certain
differences in Wohlgemuth's results, Stumpf's case must be supposed to be an exceptional
one. At all events, the theoretical procedure upon which alone I wish to insist here,
must be applied to all relevant and stable experimental results.

Henry J. Watt 35

(710 "Pseudo-movements, e.g. stroboscopic movements, produce

an after-effect exactly as an actual movement does."
G 7 — 9 " The after-etfect increases in one or several ways, within
limits, with the number of stimuli simultaneously
affecting a given area ol the retina, and or with the
frequency with which the stimuli pass given retinal
elements."
E 10 — 13 " The after-effect at first increases very rapidly with the
objective velocity, but soon reaches a maximum and then
gradually diminishes with further increase of speed."
^14 Quoted above.

E2\, 1 — 4 In the periphery of the field of vision the after-effect is at
first more vigorous, but diminishes and disappears very
rapidly.
E'il, 5 "Any after-effect in a not-stimulated area is of opposite
direction to that of the stimulated area." [Not weaker
or less rapid.]
Eb — 6, 1 Here Wohlgemuth says that "distinctness of contours is
not the essential factor in the production of the after-
effect." But it is evident from page 37 that "distinct-
ness of contours" is only an alternative reading for
"difference of brightness."
E 28 " After fatigue has been produced by a long series of

movements alternating in sign (so that the after-effect
is greatly reduced), the after-effect of movements at
right angles to the direction of the previous ones is only
very slightly affected, if at all."
EIQ "When several objective movements of different directions

stimulate the same retinal area simultaneously or
successively, an after-effect is produced which is the
resultant of the after-effects of the various movements."
C\ obviously admits the influence of order and 02 provides a
better basis for its regular introduction. Contour is the chief form of
accentuation of visual position, so that the greater the number and
frequency of the moving contours the greater the variation of orders
{El — 9). The impression of motion comes into full effectiveness more
or less suddenly after a certain rate of motion has been obtained, but it
becomes less clear with the higher velocities {E%, ElO — 13). E14!
calls for the operation of a factor which is independent of the division
of the period between the light and dark portions. It is, on the other

3—2

36 The Psychology of Visual Motion

hand, a well-knowu fact that orders and distances are clearer when
they are regular and symmetrical. E 5 — 6, 2, which shews that a grey
stripe is more effective in alternation with a white one than a black
one is, may be supposed to involve a greater clearness of orders. For
when black and white are juxtaposed, they must intensify each other
by contrast, and so make irregularities of brightness of their surfaces
less noticeable than they would be if the black were replaced by grey.
That is to say, grey favours the distinction of positions, or, in other
words, it allows of the existence of many orders, besides that given in
the contours. With ^26 we may compare what was said above about
the apparent movement of squares in an oblique direction, when two
sets of parallel rods move at right angles to one another. If the after-
effect is correlated with the neural basis of orders, directions, and
motions, there is no reason why fatigue for one direction should affect
the receptivity towards another direction at right angles to the first
(^28). Wohlgemuth's explanation of this result, on the contrary,
must be said to be highly strained {vide p. 107). As regards £"21, 1 — 4,
it is a commonly accepted fact that motion is more insistent in the
periphery of the field of vision, but that positions there are not so
highly differentiated as in the centre of the field. We might, then,
expect a more insistent after-effect of briefer duration, rapidly dis-
appearing. It is difficult to see what relation CIO, especially as
described by Wertheimer, has to the varying intensity of pairs of
movement centres. But their relation to differences of order and of
time is obvious.

The conclusion must, therefore, be that the after-effect is correlated
with, and directly or indirectly dependent upon, the order-differences
of sensation given by the objective movement which excites the after-
effect. There is no fact which suggests that the after-effect is inde-
pendent of this attribute of simple visual sensation. E 1 and E 2 — 4
only imply that the presence of clear qualities and high intensities
involves clearer sensational orders than does a darker or obscurer
field.

Of (79 Wohlgemuth says: "This result is probably merely a
question of fusion of two retinal fields like results Nos. [C] 6 and 7 "
(p. 103). These binocular cases do surely belong to quite a different
class of integrative processes to be studied separately from uniocular
cases.

{d) The only other attribute which could come into question at all
is that of position in time which represents rate of succession of stimuli.

Henry J. Watt 37

It is undoubtedly involved in the production of motion and its after-
effect, both in Wohlgemuth's theory and in any other. It is definitely
involved in G I, 08,0 10, E 7—9, E 10—13, and E 14. Neither motion
nor its after-effect is to be correlated with simultaneous sensations.

The following results do not apparently favour or disfavour any
particular theory of the psychological or neural basis of motion and its
after-effect : 0 5, E 22, E 25, E 27, E 30.

The preceding investigation thus bears out the suggestions given by
the psychical affinities of the after-effect of motion. This not only
resembles motion, but it is related by direct psychical correlation with
the experience of motion evoked by preceding objective movement, and
with the conditions which favour or indicate a greater clearness of the
orders of the sensations aroused by the objective stimulus. I offer no
physiological theory alternative to that of Wohlgemuth. Nor do I
mean to suggest that the after-effect of seen motion is linked to the
preceding objective movement by any bonds of psychical causation.
But I would maintain that the introspective nature of the after-effect
is such that it resembles motion and order, while the correlations
which experiment has established between the objective motion and the
after-effect are such as to lead one to believe that the physiological
basis of the after-effect is identical with that of motion and that both
are connected with, and dependent upon, the physiological basis, not
of intensity, but of order. A purely psychological statement of the
resemblances and correlations between experiences must precede, not
only every psychological theory regarding their connexion, but also,
and a fortiori every physiological theory of their basis. It cannot be
a safe proceeding to construct physiological theory by inference from
psychological facts while the task of systematization of the psychological
facts is neglected, whether a psychological theory of these facts is given
or not.

IV. Wertheimek's Criticisms of certain
Psychological Theories.

In Wertheimer's experiments, as in Wohlgemuth's, motion is
seen when there is no real motion at all, but only the successive
appearance, at times separated by varying intervals, of (usually) two
brief stationary visual stimuli, a and b, separated by a short space, or
at right angles to one another, like the two parts of the letter L. The

38 The Psychology of Visual Motion

motion seen may be indistinguishable from the seen motion of a single
real object, e.g. a short line turning through a right angle, or it may be
double, as if first one small line made a movement through say 30°
downwards from the vertical and then another small line through say
30° into the horizontal position; or it may even sometimes be so
evanescent as to appear, apart fronfi the motion of either small line,
as a sort of pure, abstracted motion in a definite part of the field
between the two lines, a mere ' going over ' or torsion. All this, more-
over, withstands the test of incognitive methods, just as does the
examination of pictures shewn by the cinematograph. One may know
the theory of the cinematograph or not, it makes no difiference. So
here again, the mind does not play a part in the production of the
motion, not even by associating the parts omitted by the cinematograph
with the parts shewn by it. For the motion will be seen even for
objects that have never been actually seen by the spectator, e.g. an
aeroplane, just as well as for the most familiar objects. Besides how
could reproduction of the lost stages make the pictures move when
they are not in motion at all, so long as they can be apprehended by
vision as pictures? Thus we seem again to be driven to the physiology
of the central nervous system for an explanation.

No objection need be raised against Wertheimer's physiological
theory from the psychological side. Evidently it is only a theory of
this kind which, as Wertheimer shews, can explain the facts relating to
the production of motion by the stroboscope, the cinematograph, or
other similar devices. I wish only to call attention to his criticism
of those theories which attempt to regard motion as a form-quality
(Gestaltqualitdt) or as a complex quality (Gomplexqualitdt) or the like,
and which attempt to construct a psychological theory of motion from
this leading idea. In Wertheimer's view these theories are put out of
court by the fact that they demand that the motion which arises when
the stimulations a and b are given in the manner described, shall apply
to, and embrace, phenomenally both a and b. But, as Wertheimer has
shewn experimentally, there are such things as " partial movements," —
a moving over one space and b moving over another space, the
two movements being separated from one another by a small space;
there is also such a thing as singular movement, when only a or b
moves; and, best of all, the seen motion may not apply to, or embrace,
a or 6 at all : these may be completely at rest and there may be in the
space between them the phenomenon of pure motion or torsion, an
experience much like Wohlgemuth's after-effect. Thus a theory which

o.l "^T- Henry J. Watt i "^"^ 39

suggests that motion is founded upon at least two contents, in this
case a and b, may be dismissed without further comment. Besides
such a theory would have to explain all the other facts gathered by
Wertheimer, which, needless to say, it could never do^ Wertheimer,
finally, offers a physiological theory of the facts, his theory of "physio-
logical short-circuit." = •:..• t.'ir. ;

i It may very well be that this or that theory of the type criticized
lias, in its ignorance of the facts, attempted to explain what was
known of the facts of stroboscopic movement by using a and b as
"(oundmg contents" (fundierende Inhalte). But a critic may be ex-
pected to see the virtue, as well as the vice, of a theory. Like the
eastern monarch who was invited to witness a horse race, and replied:
"I already know that one horse can run faster than another," may we
not also say: "we know already that there can be two disconnected
movements, or that one thing may move and another be at rest, or
that a motion may take place in the space between two things without
affecting either?" Surely if Wertheimer offers a physiological theory of
his facts, he thereby discredits his criticism of the form-quality type
of theory! A felt motion may have any manner of cause you please,
so long as the felt motion is supposed to correspond to its subservient,
central neural basis. In Wertheimer's experiments a and b are mere
stimuli, not founding contents.

'-- ft -r ' y The Present Theoretical Outlook.

I indicated in the opening lines of this paper that neither Wohl-
gemuth nor Wertheimer explicitly discusses the general attitude he
adopts towards psychological and physiological explanations. They do
not say why a psychological theory need not be offered for certain
facts, nor why a physiological theory of these facts is admissible.
Probably the reader is supposed to be sufficiently disciplined in these
matters already. But if the relations between the component parts
of a complex attitude remain obscure, there is grave danger that one
of these parts may be over-emphasized and overworked, so that con-
fusion results. It will therefore be well to discuss this attitude, to
clarify the relations of its parts, and to find which should dominate the
others if the best and most harmonious results are to be obtained.
The whole situaiion may be seen analytically by means of a sui'vey from
two opposite points of inquiry.

1 Cf. Wertheimer, pp. 242 f.

40 The Psychology of Visual Motion

(a) What reasons can be given for the absence of psychological theory?
There are three which may be imputed to these authors. Either
(1) they feel convinced that there is nothing for such a theory to explain;
or (2) they see in experience no beisis upon which pure psychological
theory might be built up ; or (3) they are convinced that the facts of
experience are mere discrete differences which can be explained only
by physiological theory, based upon the special relations between
experiences and the various features of the stimulative processes which
evoke them. Acceptance of the third situation obviously excludes
occupation of the first two. For if there jb nothing to explain, there
is no need for a theory of any kind ; and if experiences are not con-
nected in some way, but are mere discrete differences or qualities, the
physiological entities (mechanisms, etc.) deduced therefrom will also be
discrete and unconnected and therefore useless. And that is what
we find; for just in so far as Wohlgemuth and Wertheimer identify
pure motion or the after-effect with ordinary motion, they construct their
physiological theories to accommodate both ; and in so far as Wohl-
gemuth distinguishes motion from successive and continuous change of
position^ he must be held to give a purely illusory theory of motion or
he assumes the existence of what he calls a " subcortical centre of move-
ment^" and that, after all, is nothing but a ready made, specially created
machine, which cannot have evolved out of the fundamental neural pro-
cesses. But surely both the body and the mind must evolve ; and if so each
must evolve out of its own fundamental processes by the inner necessity
and illumination that is given by progressively increasing effectiveness.
To treat experience as a heterogeneous collection of elementary varie-
ties, more or less similar, but essentially independent, therefore renders
every scientific endeavour based upon the study of experience nugatory.
Experience, like the starry sky which guides the sailor, is not merely
one of the happy accidents of creation, merely "just so," and no more.
It was a world of life before the sceptics tried to take it as an occa-
sional, natural chart to the dark oceans of neural physiology. And it
will be all the better a chart when it has again taken its place in
knowledge as an ordered, inwardly coherent world.

If then we neglect the systematization and theoretical study of
experience, we upset the natural hierarchy of the component parts in
the complex task of the psychologist or psycophysiologist and so
achieve confusion. Neglect to systematize experience leads to neglect
to systematize the physiological mechanisms we imagine by inference.
* Wohlgemuth, p. 88. 2 qp, ^it. p. 99.

Henry J. Watt 41

And without systematization there can be no theory of the evolution
either of the brain or of experience.

(b) Adopting an opposite point of inquiry, we may now ask: What
insight justifies the confidence with which a physiological theory of
'certain facts is offered and admitted'} When double contacts give
single touch, anyone apparently may understand that that is explicable
only by a theory which assumes the existence of a single point of
maximum central excitation and explains the way in which that arises
out of the given double peripheral excitations. When an after-effect
of negative sign arises from preceding objective movement, or when
a movement arises from one or more resting stimulations, anyone may
likewise understand that no laws of mind lie hidden here. The
assumption of an indubitable parallelism of mind and body seems to be
the only justification of these views. But this assumption, as we have
seen, is abandoned by Wohlgemuth in his special physiological theory
of motion and its after-effect. Why should we, then, retain it at all?
Why not maintain that, when single touch results from double contacts,
both the central and the peripheral excitations are double, and that
single touch is due to the fact that for the two excitations the soul has
rendered only one experience? Alternatively we might assume that
the two excitations really did arouse two sensations, but that these two
fused for some reason into one. Such assumptions have indeed been
made, not perhaps for double contacts, but for those binocular stimulations
which result in single fused vision. And no charge of absurdity or of
obvious error could be brought against them. But these two cases
of single touch and single vision from double stimulations are essenti-
ally parallel in nature. For the former only physiological explanations
are generally admitted ; for the latter physiological reasons have also
been given, but they have been held to be utterly inadequate and
psychological interpretations have been favoured instead ^ If the physio-
logical or the psychological line of explanation is preferable in special
cases, there must surely be clear ground for the preference.

This ground seems to me to be a tacit recognition of the possibility
and validity of pure psychological theory. In dealing with single touch
or any other similar sensations, we recognise that there is nothing
psychologically simpler and more primitive than elementary sensation
itself to which we might appeal for an explanation of its characteristics.
Consequently, if one class of sensations shews features which another
does not possess, we feel justified in assuming that the anomaly must
' Cf. W. McDougall, Body and Mind, London, 1911, chap. xxi.

4^ The Psychology of Visual Motion

be due to th6 peculiar nature of the stimulus or of the receptor of that
sense, i.e. it must be due to physical or physiological causes. All
sensations, then, must be of one psychological class and of one psychical
type and must behave, apart from extraneous causes, in the same way.
This assumption is quite admissible as a working hypothesis, since ucf
positive arguments can be brought against it, no matter how difficult
it may be to establish it. On the other hand, the admission that, apart
from discrete differences in quality and in the extent of range of
variation of any attribute, sensations may be of different types, is
scientifically self-destructive. For the departure from type means the
failure of generalisation and therefore the absence of explanation.
There can be no true science of psychology at all, unless the simplest
sensations conform essentially to one type. Hence the common appeal
from the psychology of the sensations to physiological theory implies
both the admission of the assumption of types and the recognition of
a fragment of pure psychology.

Similarly it is justifiable to offer a physiological theory of the after-
effect of seen movement and of stroboscopic movement; for there seems
to be no obvious psychical reason why the after-effect should be of
a direction opposite to that of the preceding objective movement. If
previous writers have offered psychological theories, a closer examination
of the facts shews that the processes they appealed to are not involved
in these experiences^ Nor is there any apparent psychical reason why
the presentation of a successive series of stimuli differing in position
should arouse the experience of a continuous movement over a distance
or of many small neighbouring movements, etc. Besides, these peculiar
effects are so like the experiences evoked by objective movements that
we may at once assume that the physiological basis of the latter is
identical with that of the former. Psychological theory has, then,
only to classify and systematize the varieties of movement experience
and to set them into relations of resemblance to the already classified
simple sensations. The result of this task defines the problem for the
physiological theory of motion, which has not only to imagine a neural
basis of motion, but has also to shew how it is connected with the
neural basis of the simple sensations, besides indicating, by reference
to the incidental features of the physical processes taking place in these
neural structures, how the anomalies of the correlation of external or
preceding stimuli and consequent experiences (reversed after-effect,
movement from stationary stimuli) are to be accounted for.
» Cf. Wohlgemuth, pp. 90 f. ; Wertheimer, pp. 240 f.

Henry J. Watt 43

Tliis attitude towards the problems of simple sensations and the
simplest other sensory experiences is confirmed by a consideration of those
cases in which two systems of sense-organs, eyes or ears, work together
to make certain experiences possible. A careful survey of the problems
is here made inevitable, because the facts suggest the view that the
unity of binocular vision has no unitary neural counterparts We seem
compelled to allow that we get unitary vision not only from double
peripheral, but also from double central excitations. That the method
of approaching these cases must also give first place to positive psycho-
logical classification and theory I have attempted to shew elsewhere ^

1 Cf. McDougall, loc. cit.

2 .« rpjjg delation of Mind and Body," this Journal, 1912, v. 299 ff.

{Manuscript received 6 March 1913.)

\

THE COMPARATIVE METHOD IN PSYCHOLOGYi.

By CARVETH READ.

§ 1. Bise of Comparative Science.

§ 2. The Comparative Method assumes continuity of descent hy
heredity or tradition.

§ 3. Its explanatory force consists in shelving what this line of
descent probably was.

§ 4. Breadth of ilie field of evidence ; as depending on the psycho-
logical assumptions involved.

§ 5. The causes of whatever modifications of any faculty may
have occurred in the course of its descent must also be
assigned.

§ 6. Comparative Psychology requires the construction of Animal
Psychology.

§ 7. The difficulties of Animal Psychology.

§ 1. Comparative Psychology is merely Psychology treated by the
Comparative Method ; or the application of the Comparative Method to
the study of mental phenomena, or to the interpretation of behaviour ;
so far as the results of Zoology and Ethnology require for their com-
plete understanding an appreciation of the mental processes of other
men and animals ; and so far as such appreciation enables us to under-
stand each type of mind by comparison with others, and thereby the
better to understand our own.

The comprehensive idea of comparative science is modern ; we are
now familiar with such terms as Comparative Anatomy and Com-
parative Philology. But sporadic attempts to throw a light upon some
kind of fact by comparing different examples of it are old enough.
Some of the sophists, after comparing the varying laws and customs
of nations, concluded that there was no such thing as natural justice.
Every empirical induction, of course, depends upon a comparison of

* Founded on a lecture delivered at University College, Loudon, October 9, 1912.

Carveth Read 45

cases. All attempts at dividing, defining, classifying, involve the
making of comparisons : the great ancient example of this is Aristotle's
classification of animals. In Psychology, Aristotle distinguished four
grades in the activities of souls : growth and nutrition (common to
plants and animals), perception, memory and reason ; each with its cast
of impulse or desire. In modern philosophy we find Descartes and
Locke confidently comparing the minds of men and brutes. But all this
is little else than classification : classification depends upon the making
of comparisons; but the mere making of comparisons is not the
Comparative Method.

The comprehensive idea of comparative science is not older (I
believe) than the latter part of the eighteenth century, and was first
applied to Philology, when an acquaintance with Sanscrit began
to spread amongst European scholars. Its similarity to European
languages was perceived ; and the idea arose of classifying languages
according to their agreement or difference in vocabulary and in principles
of word-formation and syntax, and of investigating especially the Indo-
Germanic languages, as exhibiting derivation and differentiation from a
common ground according to laws of change. Next came Comparative
Anatomy, and awakened the greatest interest and wonder by shewing
what homologies, or resemblances of structure, were to be discovered
in abundance throughout the animal kingdom.

Yet Comparative Anatomy, though profoundly interesting, failed
for some time to enter upon the true Comparative Method. It is very
interesting to learn how like the wings of a bird are to the forelegs
of a lizard and to the arms of a monkey ; that the pineal gland, the
seat of Descartes' soul, is the same organ that in Sphenedon and
some other lizards gives rise to a vestigial eye that opens on the
roof of the skull ; that the bones of a gorilla differ from a man's only
in their proportions : but what then ? Hundreds of such comparisons
and identifications leave us in a state of intoxicated wonder ; and
wonder, according to Aristotle, is the beginning of science ; but,
certainly, it is not the end. The accumulation of myriads of such
facts constitutes the statement of a problem, but gives no solution, nor
any method of finding one. It was not till the last century had
passed its fifth decade that the key to this problem was supplied by
the theory of evolution and of the genetic relationship of all organisms.
And this is why Comparative Philology struck into the true method
earlier ; because the unity of the human race was already believed in ;
and so it was easy to understand that a primitive people, speaking one

46 The Com2)arative Method in Psychology

language, might, after separating (under the Tower of Babel), and
wandering far and wide, come to speak different dialects, and at last
become mutually unintelligible. Hence the idea of tracing original
relationship, by comparing extant or recorded tongues, and finding
what they have in common ; and, further, of tracing the history of each
tongue as far as possible by documents, and (where these fail) by means
of the corresponding words and forms of other tongues, and by the laws
of their modification : this idea easily arose and became extraordinarily
fruitful.

§ 2. The Comparative Method, then, always requires as its basis
the assumption of continuity of descent, or of tradition, in the pheno-
mena it deals with, and is applicable wherever such' conditions are
found. If continuity of descent is found in the development of the
mind, the method is applicable to Psychology ; and I purpose to draw
the outlines of this method, and to illustrate its stages from Psychology
and the Sciences with which Psychology is most closely implicated.
Most of my illustrations concerning instinct may seem to belong to
Zoology, rather than to Psychology ; but that is unavoidable at present,
since there has hitherto been so little successful work done upon the
subjective side of instinct that purely psychological materials are
wanting. In my opinion, moreover, there is no such thing as a science
of pure Psychology ; and since, if attempted, the data must be entirely
introspective, it could only hold of the human mind:

The Comparative Method is not merely a drawing of comparisons,
but of explanation by means of comparisons ; and the explanation
consists in reconstructing the antecedents of the phenomenon we are
investigating and giving its- history. Many people may have wondered
at the English word " am," so conspicuous in our speech and so isolated
in its appearance, and wonder may have convinced us of our ignorance
(as Aristotle says it does) ; but most of us get no further. Philologists,
however, on comparing other Indo-Germanic languages, have found
that our substantive verb has three roots, hhu (to grow), was (to dwell),
and as (to breathe) ; and that am, like the Latin sum, the Greek
eifd, and the Sanskrit asmi, may be regarded as an abbreviation of a
hypothetical Aryan a^uma, where ma is the sign of the first person ;
so that the pleonasm " I am " has come to mean " I exist " from having
originally meant " I breathe " with the implication of existence. It is
(by the way) instructive to the Psychologist to learn that our sub-
stantive verb has been formed from three roots, none of which primarily
expressed the abstract idea of existence which they all include : our

Carveth Read 47

forefathers had the idea in an obscure and nascent form, before they
had an appropriate word for it ; and we still use all three signs,
having forgotten their more concrete significance. Biologists and
Psychologists proceed in the same way as Philologists, dealing with
genera and species, instead of with languages and dialects, and trusting
to find by a comparison of species, under the idea of common descent,
the antecedents of structures and functions that occur in any of them
whose history cannot be directly known. ''■■

Thus Darwin, in explaining the comb-building instinct of our
hive-bees, which, taken by itself, is so marvellous, refers first to the
humble-bee, which uses its old cocoons to store honey in, and adds to
them some short tubes and irregularly rounded cells of wax. But the
gap between the humble-bee's work and the finished architecture of the
hive-bee is prodigious. However, he knows of a certain Mexican bee
{Melipona domestica), whose comb (as well as her own bodily structure)
lies about midway between those of the foregoing species ; and he
argues that certain easily- conceivable alterations in the habits of the
Mexican bee, all favoured by the principle of economy or utility, would
result in such structures as those from which we get our honey. Com-
parative Psychology must adopt a similar course.

Nothing, for example, is more conspicuous than the superiority of
the human over the animal mind. Hence it has been felt to be a
matter of first importance to shew that the higher mammalia, and
especially the anthropoid apes, supply a sort of mean of intelligence and
character between man and the lower mammalia : that the ape is as
much above the monotrerae as man « is above the ape. And in
Darwin's Descent of Man, chapters iii. and iv. consist chiefly of
facts to prove that the most characteristic human faculties are fore-
shadowed in the higher mammalia and especially in the Primates ; so
that it is credible that the immediate ancestors of man, could we
restore them, would present to us all the intermediate stages. So
far as we may infer mental faculties from a skull, the remains of
Pithecanthropus, found in Java, verify this position: the skull's capa-
city lying about midway between that of a chimpanzee and that of
an Australian.

§ 3. The Comparative Method explains by pointing out a possible
course of causation in the production of some phenomenon by tradition
or inheritance ; it does not show that such or such steps were actually
taken, but only that (judging by parallel cases) their having been taken
is more or less probable. If we could produce a record of the actual

48 The Comparative Method in Psychology

steps, there would be no call for the Comparative Method: the recovery
of actual combs of the hive-bee, shewiog all its variations from the
beginning, would make it quite needless to refer to the workmanship
of allied species. The use of the Method is to construct or confirm an
elaborate hypothesis concerning a series of antecedents that have been
lost. Even in reconstructing the famous genealogy of the horse from
an ancestor with five toes, the evidence (I understand) depends upon
a number of specimens which make it highly probable that a certain
series of changes in his structure took place, but do not supply the
unbroken series of intermediate forms. Hence if one hears of only one
or two cases of such reconstruction of genealogies, the evidence may
seem feeble ; but this feeling passes off, as more and more cases are all
explained by the same method. When we are investigating some
ancient instinct, or the rise of intelligence, we certainly shall not find
the earlier stages of it preserved in the geological record ; but we may
find some of its correlative structures and some of its products, and
may learn much concerning the climatic and biological environment
in which it was exercised. Fossil evidence of the existence of gall-
wasps and their galls is found throughout the tertiary strata: we are,
therefore, sure that the correlative instincts of wasp and larva were then
exercised. The history of an instinct, then, though the very steps of
its descent to extant species be irrecoverable, may conceivabfy be
reconstructed as well as that of the horse's anatomy.

Geology (I may observe by the way) seems remote from Psychology,
until we see that, in the department of Palaeontology, it is very
helpful and even necessary. It has been suggested (for example) that
the relatively small brain of Dinosaurs, compared with mammals of
equal bulk, may explain the extinction of their order: inasmuch as
the small brain implies low intelligence, and incapacity to modify
behaviour in the presence of changing conditions of life. If so, and
if no other cause of their extinction can be shewn, we may turn the
argument about, and rely upon the superior adaptability of intelligence
compared with instinct, and its consequent biological utility, as one of
the reasons (or the chief of them) why intelligence gives a predominance
to those organisations that trust to it.

The course of causation that is pointed out by the Comparative
Method comprises : (1) a probable line of descent, along which a
certain organ, or instinct, or institution (in social affairs) has been
handed down by inheritance or tradition, and has been gradually
modified; and (2) the causes of each modification so far as they can

Carvbth Read 49

be assigned. An hypothesis thus constituted is, of course, subject to
the usual logical conditions : it must be comprehensive, consistent with
all the known facts, and better in these ways than any rival hypothesis.

As to the line of descent, Darwin says : " In searching for the
gradations through which an organ in any species has been perfected,
we ought to look exclusively to its lineal progenitors ; but this is
scarcely ever possible, and we are forced to look to other species and
genera of the same group, that is to collateral descendants of the same
parent form, in order to see what gradations are possible, and for the
chance of some gradations having been transmitted in an unaltered or
little altered condition. But the state of the same organ in distinct
classes may incidentally throw light on the steps by which it has been
perfected " {Origin of Species, c. vi.). Accordingly, in discussing the
comb of the hive-bee, he confines himself (as we have seen) to collateral
descendants of the same parent form ; but when he explains the possi-
bility that the vertebrate eye has been produced by natural selection,
he refers to the facetted eyes of insects and crustaceans to illustrate
the range of variation, and to primitive forms in which there is no lens,
and even to aggregates of pigment-cells without any nerves. Now, for
Psychology, the history of the eye indicates the history of the power
of vision, so far as it depends upon the peripheral organ. Similarly, if
the antecedents of the human mind are sought only in the Primates, the
inquiry is confined to "collateral descendants of the same parent form ";
but we seek to throw light upon our subject matter from "the state
of the same organ in distinct classes," or to classes whose connexion
with ourselves is more and more remote, when we turn for parallels
to dogs and cats, or reptiles, or fishes, or even go to the ant and consider
her ways.

§ 4. The field of evidence, then, from which parallel cases may be
drawn seems to become of indefinite extent. No one has used the
Comparative Method more powerfully than Dr Frazer in elucidating
the origins of social beliefs and institutions ; and to find evidence for
his speculations he ransacks the whole storehouse of human records
contemporary and historical. In The Magic Art (vol. Ii., cc. xiv. to xvii.),
discussing the worship of Vesta at Rome, he derives from various
sources some significant hints concerning similar practices among the
Latin tribes and the Kelts and the Greeks, who stand to the Romans
in the relation of varieties or proximate species, like the humble-bee
and M. domestica to the hive-bee ; but for the closest parallel to the
institution at Rome he turns to the Hereros or Damaras of South-

J. of Psych. VI 4

50 The Comparative Method in Psijchology

Western Africa. One may ask how a case so remote can throw any
light upon the argument. No doubt a parallel from Keltic or Greek
religion would have been more acceptable : and that similar institutions
at some time obtained amongst those peoples is highly probable ; but
no record remains of them, and therefore none can be adduced. Con-
cerning the Hereros recent and credible testimony is forthcoming ; and
it is justifiable to resort to it for the following reasons. It can be
shewn that the modes of inference and the other mental conditions
determining belief, and the beliefs resulting from such conditions,
amongst races at about the same level of culture, are very similar, and
are more alike the more primitive they are. Their social conditions
as to family and tribal organisation have much in common. The main
external conditions of their life are also similar : dependence on the
sun and weather, on the fertility of flocks and herds and fields (if they
own or cultivate any), on the kindling of fires or the preserving of a
perpetual fire. And causes being similar, so are the eflfects. Hence
arise similar doctrines concerning the divine nature of fire, and insti-
tutions for kindling, maintaining and worshipping it. Plain signs,
if not explicit evidence, of the existence at some time of such beliefs
and practices can be found in Asia and America, as well as in Europe
and Africa; and it is therefore not unreasonable to confirm an hypothesis
concerning the customs of the legendary age of Rome by describing
the customs of the Hereros. Indeed it has this advantage, that the
more remote, racially and geographically, the peoples are one from
another, the less likely are their resemblances in belief and ritual to be
traceable in any way to imitation ^

This argument of Dr Frazer's depends entirely upon the psycho-
logical position, that the modes of inference and other conditions

1 On the other hand chronological remoteness of peoples (in the above case, about
2500 years) leaves more opportunity for possible transmission of influence by imitation,
or even for tribal migration. No one will suppose that the Hereros have imitated,
however indirectlyi or are actually descended from, the early Romans; but Professor
Sollas suggests that the Bushmans with their remarkable pictorial art are the same people
who painted the walls of caves at Altamira and elsewhere on either side of the Pyrenees
(say) 20,000 years ago. There is time for the migration to have taken place, and perhaps
some indication of the route pursued may be given by painted rocks in various parts of
Africa from Morocco to the Cape. Moreover there is some evidence in their stone-carvings
that the artists of the Pyrenees were, like the Bushmans, steatopygous {Ancient Hunters,
etc. , c. 9). Herodotus doubting the claim of the Sigynnae, dwelling beyond the Danube,
to be descended from the Medes, adds tbat, nevertheless, given a long time, nothing is
impossible (B. v.). The art of the Bushmans differs in some characters from that of their
supposed ancestors (and no wonder !) ; but if this hypothesis should be confirmed, we have
here a remarkable proof of the persistence of racial traits.

Carveth Read

51

determining belief and (through belief) action, are everywhere the
same amongst men, so far as men are unscientific, and that therefore
under similar conditions we shall find similar practices. The extensi-
bility of this principle sets a limit to the Comparative Method. The
cruder conditions of belief give way, in some circumstances (as under
stress of trade and industry), and in some minds (least subject to the,
'crowd'), to more exacting tests of truth, which culminate for us in
scientific method. Few of us, indeed, apply this method outside of our
special studies ; but so far as any people do so, it is impossible for the
practices connected with Magic and Animism to establish themselves ;
since such practices depend upon social beliefs concerning the con-
nexions of phenomena accepted without analysis; which may, therefore,
be groundless and absurd ; whereas scientific investigation seeks the
necessary relations of things.

Similarly for the whole Animal Kingdom there are certain indis-
pensable assumptions which mark the limits of Comparative Psychology.
The first requisite is the presence in all animals, even the simplest, of
some degree of consciousness ; the second is the connexion between
sensation and reaction, and between reaction and feeling. At a higher
level the associability of sensations must be assumed ; higher still
the associability of perceptions, and (again) of ideas. To determine
the marks by which to judge whether a species lives at one or
another level is a leading problem of our study. Not that the levels
are sharply demarcated ; but to assume in any case a principle of
subjective interpretation which the given species does not share in,
must lead to disappointing results. Where, in interpreting the
behaviour of a given species, a principle requiring a certain stage of
development is assumed, parallel cases can only be looked for at, or
above, that stage. And if there is no fundamental similarity of mental
processes throughout organic nature, it is impossible to reason freely
from one case to another, or to trace by parallel reasoning any line
of descent for the mental faculties. But it is the laws of mental
functions that must be assumed everywhere to prevail: many particular
functions may be dispensed with. There are secondary characters of
mind, such as colour-sensibility, or vision as a whole, or hearing, or
' image,' as to which it is important to find out experimentally whether
they are present or absent; but their absence does not disable our
judgment of an animal's mind, such as it is.

And a similar observation may be made in Ethnopsychology ; we
may distinguish between fundamental and secondary characters of

4—2

52 The Comparative Method in Psychology

culture, depending on fundamental and secondary traits of mind.
Fundamental, for example, are magical and religious rites; secondary
are art-products : the former depend upon universal modes of inference
and conditions of belief; the latter may be due to several motives,
varying in strength from tribe to tribe, and one or more of them
possibly absent here or there — magic, religion, commemoration of
events, communication with others, industrial value, or artistic delight
in the work. Any art-product — a drawing, decoration or dance — being
given to find its motive or motives, we are at first perplexed by the
possibility of a " plurality of causes "; but this difficulty generally gives
way upon examination of the details. On comparing as many
examples as possible that are known to have chiefly some definite
motive, marks may be found of one motive, others of another, and so
on ; so that, looking for such marks in the given object, we may
assign it to one of the motives, or to some of them. It is true that
an expert may form a respectable judgment upon such matters
without being able to explain why ; but if he could explain, it would
be by bringing forward the distinctive marks of each class, which lie
submerged and unanalysed in the total impressions of his experience.
If the motive of any work of art cannot be assigned either scientifically
or empirically, we are so far left without a clue to the psychology of the
artist and his people.

§ 5. As to the causes of modifications that take place in the line of
descent, they are of two kinds : (1) General advantages, reducible for
the most part to ' utility,' such as protective resemblance, or economy —
say economy of wax, which (Darwin thinks) must have favoured the
survival and spread of bees that built the best comb ; or adaptability to
new conditions, which must promote the success of animals of superior
intelligence. (2) Particular conditions, such as the Arctic glaciation, which
(according to Wallace) may have originated the migratory instincts of
many birds : or increase of predatory enemies, which would give survival
value to alertness, quickness of perception, promptitude of action (in flight
or defence) in herbivores, according to the disposition of each species.

If particular conditions are assigned for any change of structure,
instinct, faculty or custom, there should be evidence, direct or indirect,
(1) that the alleged conditions — say glaciation — really occurred at the
time, and to the extent, required for producing and confirming the
modification ; and (2) that they might be expected to produce it, either
for general physical reasons a priori, or because in parallel cases similar
results have happened (that is, by the Comparative Method). The

Carveth Read 53

glaciation of the whole Arctic region and a great area of the temperate
zone, in the Pleistocene period, has been very generally admitted ; but
the extent and continuity of it is disputed by Dr Scharff {Origin and
Distnhution of Animal Life in America) ; and the adequacy of such an
event to produce migration is disputed by Mr Dixon {Migration of
Birds), who urges that such a climatic change would bring about not
the migration but the extinction of species. I mention these objections
to Wallace's hypothesis by way of illustration, offering, of course, no
opinion as to their validity.

In Psychology the principle of utility is frequently appealed to in
order to account for successful modifications in conduct both in animals
and in men. It explains (for example) the appearance of the parental
instincts in a few fishes and reptiles and in most birds ; since the death-
rate is reduced, fewer offspring become necessary to maintain the
numbers of the species, and the development of individuality is favoured.
The utility of numbers for mutual defence, or of giving the alarm when
danger threatens, explains the gregariousness of most Ungulates ; the
utility of numbers for attack explains the gregariousness of wolves and
probably of our own remote ancestors. Sociality having been esta-
blished, it becomes a particular condition of most other traits that dis-
tinguish social animals ; and the further development of society by
internal differentiation of occupation and rank supplies at every stage
particular conditions of nearly every peculiar belief, custom, and
character. The establishment of the Kingship, for example, seems
to be such a condition of the growth of a belief in gods who are kings,
of the rites of their worship and of the feelings with which they are
regarded.

The general utility of intelligence for the sake of adaptability is
plain. As to the particular conditions of the development of intelli-
gence in the Primates, including ourselves. Professor Elliott Smith lays
great stress upon the habit of treC'-life, characteristic not only of the
lower Primates but also of their ancestors. The fact that they had this
habit will not be disputed, nor the persistence of it during a sufficient
period to have important consequences, since the remains of lemurs are
found in the Eocene strata. The habit of living in trees diminished
the predominance of the olfactory sense, and favoured the activity of the
visual and auditory distance-senses, as well as the tactile sense and the
kinaesthesis, and also the development of those areas of the brain that
subserve attention and association^ Perception of an enemy or of food,
^ Address to the Anthrop. Sect, of the Brit. Ass., 1912.

M The Comparative Method in Psychology

by the distance-senses, sight or hearing, implies an interval of time,
longer or shorter, before contact can result in injury from an enemy or
in the obtaining of food : an interval that may be filled with many
mental processes favourable to escape or possession, and therefore giving
advantage to those individuals in whom such processes occur and in
whom they are most effective'. And perhaps I may be allowed to refer
to the hypothesis set forth in Natural and Social Morals (c. vii. 2), that
it was the adoption by man of a carnivorous diet, and of the habit of
hunting in pack the larger kinds of game, that supplied the main condi-
tion for the early developments of human traits both physical and
mental in our anthropoid forefathers. That vegetarian animals may
adopt a carnivorous, and carnivorous a vegetarian diet, or a mixed one,
is shewn by many examples : not only domestic dogs and cats, but also
wild pigs, squirrels, bears, etc., and some monkeys are partially carni-
vorous, to the extent of eating birds' eggs and even young birds. The
possibility of extensive changes of diet is implied in the hypothesis that
all the higher mammalia are derived from an insectivorous stock. Again,
the earliest men we know of were hunters, and their remains are found
with those of large game. The advantage of hunting in pack is shown
-by the dogs and wolves ; and it is possible that some of these joined
our own pack at the beginning, and have shared our fortunes ever since.
To the pack language was useful, and that gave the basis of nearly all
our intellectual superiority.

Between the derivation of instincts and customs by heredity or
tradition and their modification by general influences or particular con-
ditions, there is no incompatibility ; on the contrary, the latter pre-
supposes the former. Dr W. H. R. Rivers, in his Address to the
Anthropological Section of the British Association (1911), on the Ethno-
logical Analysis of Culture-, tells us that, having formerly been too
much addicted to the explaining of customs and institutions everywhere
by independent evolution, he became convinced by his experience in
Oceania of the necessity of attributing many important social pheno-
mena to racial mixture and the blending of customs. The proofs that
he gives of this must convince everybody else, even if we had not many
examples to the point in the history of our own country. But he does
not mean (I believe) to disparage the idea of the independent evolution
of customs. Customs must exist before they can be blended. And as

1 Cf. Sherrington, The Integrative Action of the Nervous System, c. ix.
« Cf. Dr Rivers's paper, "The Sociological Significance of Myth," in Folk Lore,
Sept. 1912.

Carveth Read 55

for the Comparative Method, it is as applicable to the modification of
customs or institutions, of instincts or intelligence, by migration, con-
quest or imitation, by changes of climate, habitat or food, as it is to
descent by tradition or heredity. Where, for example, it may be im-
possible to find direct evidence of a conquest, such as one supposes to
have changed the customs of a people, indirect evidence may perhaps
be adduced that similar changes have resulted from conquest in other
countries. If we look for similar cases in animal life, we shall hardly
find that any species has altered its behaviour by imitation^ of another
species ; but migration and conquest (exposure to new enemies) must
often have extensive consequences ; such as the reawakening of fear in
the birds of a desert island, after man has come amongst them and dis-
turbed their "ancient haunts of peace." And indirect evidence of a
change of behaviour may be all that is available. There is no direct
evidence (I believe) that the swallows that now build under our eaves
formerly built elsewhere; but we cannot suppose that they first began
to build nests when they recently found houses convenient to shelter
them ; and we know that an allied species, the house-martin, now builds
in caverns and in cliflPs as well as in houses, and that many a species of
bird, spread over considerable areas, builds under different climatic or
defensive conditions, different kinds, of nest differently located. So
that the modification of instinct in the house-swallow presents no
difficulty.

In these cases, from the observation of certain modes of behaviour,
we infer a change and the causes of it. Conversely, if there are grounds
for believing that a certain change of conditions has had certain conse-
quences, we may look for similar consequences wherever a similar change
has taken place : if, for example, the development of vision and hearing,
as distance-senses, is held, with good reason, to have led to the improve-
ment of intelligent behaviour, it may be surmised that the earlier
development of the olfactory sense contributed, in some measure, to the
same sort of progress.

§ 6. If we are to apply the Comparative Method to the explanation
of the human mind, we must allow ourselves great latitude of compari-
son. The human mind cannot, of course, be explained merely from
itself. If we refer the behaviour of an adult Psychologist (whose mind
is necessarily our starting point) to its beginnings in childhood, it soon
appears that the child's mind is quite as much in need of explanation.

"^ I mean, of course, ♦ by conscious imitation ' — to be distingnished from biological
mimicry.

56 The Comparative Method m Psychology

After a few months it exhibits numerous faculties, keen and adroit, out
of all proportion to its experience; so that Romanes could rank the
child of two and a half years on a level with the adult gorilla : though
the comparison is deceptive ; for in human affairs, in which he is at
home, the child is plainly in advance of the gorilla ; whilst in the gorilla-
world, which we do not appreciate, he would be far behind. As the
child developes year by year, we see that his powers are not merely an
acquisition of his own life, whether by experience or by education in the
widest sense, but the realisation of an inheritance ; and this is clearly
indicated by the disproportionate size of his brain, and the rate of its
growth in infancy, relatively to the size of his body and to the rate of
his brain's growth in later life. If we turn to less cultivated races than
our own, or to what are called ' savages,' we find in all of them extensive
knowledge (according to their needs), intelligence and dexterity, tradi-
tions and institutions, which make it quite hopeless to seek our own
beginnings in them. Of the anthropoids, in their native state, or
monkeys, or lemurs, we know too little, but we know enough to see the
absurdity of looking for a beginning there. And the same thing is true
of all vertebrates and of all the invertebrate metazoa. It is plain that
organisation always implies antecedents, that orderly growth and de-
velopment always imply heredity. Hence there is no alternative to the
task of attempting to construct a comprehensive Animal Psychology.
This is not itself Comparative Psychology, except in so far as it is con-
structed by the Comparative Method ; but the study of animals by
experiment and observation upon each species must supply a large part
of the necessary data for the Comparative Method. Indeed, a good
many Botanists will tell us that a thorough explanation of behaviour
cannot be obtained without including Plant Psychology ; since plants
have their own organs of perception, and, when these organs are stimu-
lated, messages pass by protoplasmic channels to other parts of the
organism, and are responded to by appropriate movements, comparable
to those which in animals we regard as signs of life and mind.

§ 7, The task of working out Animal Psychology will need very
many years of labour and the co-operation of very many students. It
is far more difficult than the earlier investigators were aware of; so
that much of their work not only needs revision, but must often be
treated as a warning against certain fallacies. It is more difficult than
the Ethnologist's task ; for we are sure that human institutions begin
with man, and therefore we may hope to reconstruct the history of
culture, both in its facts and in its motives, from our observations, and

Carveth Read 57

from the records and remains of man in his various tribes and races ;
but we are equally sure that the subject of Psychology, sensibility and
reaction, has no beginning but with life itself. On the other hand, a
considerable part of the work, especially descriptive work on the senses
and nervous systems of animals, and on their instincts, has already been
done for us by Zoologists. Animal Psychology is a way of regarding
a certain area of the zoological field, namely, the behaviour of animals,
considered as susceptible of subjective interpretation ; and I hope that
few Zoologists would admit that it is separable from their own study.
Zoology, Anthropology and Psychology are conveniently distinguished
for the sake of special work ; but, like the Siamese twins, they cannot
be separated without sacrificing the lives of all of them.

The chief difficulty of Animal Psychology, however, is not its com-
prehensiveness and vastness ; to overcome that would be merely a
question of time. There is, besides, the intrinsic difficulty of finding
a subjective interpretation of the facts. In dealing with men of other
races, or of other levels of culture, we may be confident that the general
ground of their mental constitution, their senses, impulses, emotions,
and the laws of the formation of habit and belief, are very much like
our own (at the pre-ratiocinative stage) ; so that, beginning with our
own experience and proceeding cautiously and methodically, we may
hope to understand theirs. With anthropoids, too, and monkeys we
feel upon pretty sure ground ; and even dogs, though belonging to
another branch of the tree of life, yet, as social animals, lie open to
sympathetic interpretation, and we are upon most points of behaviour
upon terms of mutual understanding with them. But, as soon as we
leave the human race, we lose the power of verifying inferences con-
cerning the mental experience of others by obtaining from them direct
introspective descriptions or replies to questions ; and below the level of
the higher mammalia our difficulties rapidly increase. We may say,
slightly altering the words of Spinoza, that " the minds of animals differ
from ours as their bodies do" {Eihica, in. 57); and, accordingly, the
more their bodies differ from ours, the harder it is to understand their
minds. Moreover the likeness or difference of bodies is no matter for a
superficial judgment ; it is concerned with intimate structures, recondite
and perhaps still impenetrable. How much could we infer of the
difference in character between the cat and the dog from anything we
know of their bodies ? Our children shew a delight in climbing trees,
presumably by remote inheritance (which must have some unknown
physical basis), although the characteristic changes that have taken

58 The Comparative Method in Psychology

;place in the human body all go to disqualify them for the Primate's
ancient habitat. In some animals, again, there seem to be organs of
sense in which we have no share, such as the lateral line in fishes ;
organs serving similar purposes to our own, but so differently constituted
that the experience obtained through them must be very different, such
as the multiple eyes of insects and crustaceans ; and other organs whose
significance we cannot even guess. Besides, since we do not perceive
merely with the sense-organs, but with the brain, it is impossible to
assume that animals with brains very different from ours perceive, as
we do, even what their sense-organs, taken by themselves, seem able to
perceive. For example, the eye of a glow-worm is capable of forming
a complex image, which Exner saw and photographed ; but it does not
follow that the glow-worm's brain is capable of the considerable syn-
thesis which the perception of such an image requires ; to say nothing
of its interpretation by ideas. There are the still more puzzling instincts
in which many animals seem so wise concerning matters about which
it is impossible they should know anything either by experience or by
instruction.

With all such problems, to take the easiest way, and interpret them
by the nearest analogy in our own experience, is the error of ' anthropo-
morphism.' It was common amongst the earlier investigators of the
minds of animals ; and the danger of it (I believe) G. H. Lewes first
clearly exposed in Problems of Life and Mind (Third Series : The Study
of Psychology, c. viii.). But freedom from anthropomorphism can never
be more than a matter of degree. We must either give up the attempt
to understand the subjective side of an animal's behaviour; or else, with
whatever precautions, find some analogy .in our own consciousness.
Some ' analogy,' however, means some functional equivalence, not neces-
sarily the same sense-modality, quality of emotion, etc. Does a grass-
Jiopper 'hear,' for example ? It does hear; but it may not hear tones
as we do.

Other discouragements of our study, besides those that beset most
empirical sciences, arise from the expense of time and trouble that must
often be put up with in obtaining the simplest fact. If we work by
experiment in the laboratoiy, the animals sometimes (as one of my
colleagues observed) seem to take no interest in the proceedings. If we
betake ourselves to observation in the field, hour after hour may pass
with no apparent result, except the roasting of our own necks in the
sunshine, of which even such enthusiasts as Fabre and the Peckhams
complain. All naturalists are agreed upon the need of endurance : it is

Caeveth Read 59

generally called 'patience' — a most misleading term. Patience is
a passive quality, characteristic of those who put up with evils which
they regard as irremediable, or which they have no courage to oppose :
the weather, or the government. But the watching and waiting of a
naturalist is the perseveration of an instinct, like the anchorage of a
cat at a mouse's hole : it is the eager desire to know, overcoming all
allurements, discouragements and distractions.

If these drawbacks to the study of Animal and Comparative Psycho-
logy deter anyone from pursuing it, we may parody a saying of Kant's :
" it is not necessary that every man should be an Animal Psychologist " ;
but whoever neglects it will never get to the bottom of human nature.
Not that to understand mankind is thie sole motive to study animals :
they well deserve study for their own sake : but it seems to me to add
(at least) to the value of such investigation that, through the Compara-
tive Method, it will enable us to know ourselves.

{Manuscript received, 19 February 1913.)

SOME OBSERVATIONS ON LOCAL FATIGUE IN
ILLUSIONS OF EEVERSIBLE PERSPECTIVE.

By J. C. FLiJGEL.

(^From the Psychological Laboratory, University College,
University oj London.)

Local fatigue as manifested in the * WvndmilV_or ^ Revolving Xlross'
Illusion. McDougalVs observations confirmed. Local J^atigue not
manifested by all subjects. Differences of local fatigability in the
same subject. Experiments showing the extrem,ely specific nature of
this local fatigue. Local fatigue unaffected by other simultaneous
mental processes. The rdle of local fatigue in illusions of reversible
perspective. The general interest of local fatigue. Summary,

In a paper in the last number of this Journal^ a considerable amount
of experimental evidence was brought forward, indicating that the
I direction of the attention is a factor of great importance in illusions
of reversible perspective. In the course of a long investigation with
several trained observers, a very close correspondence was shown to
exist between attention to any part of a reversible figure and the
forward appearance of that part ; while reversals of perspective were
found to coincide with observable movements of the attention from one
part of the figure to another. It was also pointed out that these results
gave rise to a number of interesting problems, which had not been
dealt with in that paper.

Among the problems immediately concerned with the direction of

the attention itself, was the question : — Why is it that the attention

jyj. cannot be steadily maintained upon any one part of the figure, so as to

'' maintain the same perspective indefinitely, if the observer so desires ?

On turning to a consideration of this question at the conclusion of the

^ J. C. Flugel, "The Influence of Attention in Illusions of Eeversible Perspective."
This Journal, 1913, v. 357.

J. C. Flugel 61

previous research, it seemed to the present writer that the answer was
veix^ossiblyioJbe found in the phenomenQn-Of,.locaL fatigue brought
to_ light by the researches of McDougalH. According to this author,
the inability to prevent reversals of perspective is due to the rapid rise
of fatigue in the higher nervous paths subserving the perspectives,
w^hich produces in turn a series of rapid changes in the relative resist-
ances of the two systems; the whole process being only a particular
case of that general instability which is found throughout the higher
nervous arcs.

This explanation would seem to be readily applicable to our present
problem, the only change required being that the fatigued systems
should correspond not only to the perception of the perspective but also
to the clearness of the sensory presentation of a particular part of the
figure (since it is here not only the change of perspective but also the
accompanying movement of attention from^bne part of the figure to
another, which has to be explained). This change seemed the easier,
in so far as McDougall had already demonstrated the existence of local
fatigue, even on the purely sensory level, in some of his observations on
binocular rivalry ^ In view^of these considerations, it was decided to
repeat McDougall's observations on local fatigue in illusions of reversible
perspective with some of the subjects of our own previous experiments.
It was hoped that, by so doing, we might obtain direct evidence as to
the influence of local fatigue vUpon the direction of the attention, and
that we might at the same time be able to throw some light upon
various other p^'oblems arising out of our previous results, such as the
cause of the m^rijed individual differences in the power of controlling
the attention and the relation of involuntary changes in the direction
of the attention to the "fluctuations of attention" in general. The
present paper is devoted to a brief description of the experiments thus
initiated^

McDougall's demonstration of local fatigue rests upon the following
considerations. If the reversals of perspective are due to the rapid rise

1 "The Physiological Factors of the Attention Process." Mind, 1906, xv. 329.
Especially 340 ff.

2 Op. cit. 341.

3 I take this opportunity of thanking those who, undeterred by the fact that these
were experiments in ' fatigue ' (although, it is true, of a rather less trying nature than is
usual in such cases), were kind enough to offer themselves as observers. Where not
otherwise stated, these were all trained psychologists or advanced students of psychology.
My thanks are also due to Professor Spearman for much valuable help and advice through-
out the investigation.

62 Fatigue in Illusions of Reversible PersjJective

of resistance in the nervous paths corresponding to the alternative
perspectives, it would seem that, if by any means one path can be kept
in use for a longer period than usual, then when the reversal of per-
spective does at length occur, a correspondingly long period would be
required for the unusual resistance to become reduced to the normal
level, i.e. for the effects of the fatigue to wear off. In^ther words we
should expect an unusually long period of one perspective to be foHowed
by an unusually long period of the other. In the drawings and diagrams
most frequently used for the study of these illusions, it is difficult or
impossible to produce an e.xceptionally long period of one perspective.
This can be done however with comparative ease in the case of those
illusions of reversible perspective which occur with objects in three-
: dimensional space. Under these circumstances it is found (as was
shown also by our own observations with the model of the prism figure
recorded in the previous paper, p. 394) that the ' illusory ' perspective
occurs much more frequently when the object is regarded with one eye
than when it is regarded with two. In the latter case the ' normal '
perspective can frequently be held for long periods without reversal.
If after such a long period of ' normal ' perspective with binocular
regard, we close one eye and thus produce a state of more equal oppor-
tunity as regards the two perspectives, we should expect in turn an
unusually long period of ' illusory ' perspective ; and this is in fact what
occurred in McDougall's observations.

For the purpose of these experiments McDougall made use of the
' windmill ' illusion or revolving cross. If a windmill or revolving cross
be regarded uniocularly and obliquely, it can be observed that the
movement of the sails or arms of the cross apparently changes its
direction from time to time, the movement of the sails in the upper
part of their orbit being sometimes towards the observer, sometimes
away from him. This is, in reality, an illusion of reversible perspective,
the change in the direction of the movement being coincident with
a change in the apparent spatial relations of the windmill, that side of
it which before reversal appeared the neai'er to the observer appearing,
after reversal, to be the further from him, and vice versdK
I In our own experiments a cardboard cross was used, the four arms
/. of which were each 30 cm. long and 2 cm. broad, and which was

/ rotated by an electric motor at the rate of about one revolution in five

* The illusion can of course be observed when the windmill or cross is stationary, but
the reversals are in this case less marked, probably because the changes of perspective are
no longer reinforced by apparent changes in the direction of the movement.

J. C. Flugbl 63

seconds. The observer sat at a distance of 2^ ra. and at an angle of 30°
from the centre of the cross. The changes of perspective were recorded
on a smoked drum in the manner described in the previous paper
(p. 360). Care was taken that the illumination of the cross and of the
whole room should remain uniform in all the experiments. Each com-
plete experiment consisted of three parts: — (1) the 'preliminary' or
' normal period,' made uniocularly and in an unfatigued state ; (2) the
' fatiffue-2nducmg_^eriod,' during which the subject regarded the cross
binocularly ; and (3) the 'test period,' in which the conditions were
similar in all respects to those of the preliminary period, except that
the psycho-physical system corresponding to the ' normal ' perspective
was now (we may suppose) in a fatigued state owing to its having been
in exercise during the whole or the greater part of the preceding
fatigue-inducing period. The preliminary period occupied one minute, i
the fatigue-inducing period five minutes; the test period, like thei
preliminary period, occupied one minute, except when it appeared
specially desirable to extend it over a longer time. In some of the
earlier experiments (those recorded in Table I) an interval of ten
seconds w£is allowed to elapse between the end of the fatigue-inducing
and the beginning of the test period, during which the subject looked
away from the cross ; but in the majority of the experiments the test
period followed immediately upon the fatigue-inducing period. The
manifestations of fatigue do not however seem to have been appreciably
affected in any way by the presence or absence of this interval.
A number of preliminary experiments had shown that the direction of
the attention was effective in controlling the perspective in the 'wind-
mill' illusion in the same way and to about the same extent as with
the models of the prism figure employed in the previous research. It
therefore became important that the conditions of attention should be
as far as possible the same in the pj;eliminary and in the test periods.
To facilitate this affixation mark was provided, in both cases, and the
subject was told to attend principally to the immediate neighbourhood
of this fixation point. The position of the fixation mark was not always
the same with different subjects, as it was desirable for the present
purpose to allow the fixation (and consequently the attention) to be
such as to obtain a fair proportion of both perspectives during the
preliminary period, and this was found to necessitate a different position
with different subjects. The position of the fixation mark, when once
determined, remained however the same during the preliminary and the
test periods, and the subjects were instructed to make, as far as possible,
the same effort to fixate and attend in both cases.

64 Fatigue in Illusions of Reversible Persjyective

The results of the first experiments conducted in this way are
shown in Table I. In this table, as in those of the previous paper, the
figures represent the proportion of the total period of observation
during which the cross was seen in either perspective. For ease of
comparison these proportions are in each case expressed as percentages
of the total period of observation. The two perspectives are here
described as ' right forward ' and ' left forward ' respectively. By ' right
forward ' is meant that perspective in which the right hand end of the
cross is seen nearer the observer, by ' left forward ' that in which the
left hand end is seen nearer. In this case the ' right forward ' was
the ' normal ' or ' real,' the ' left forward ' being therefore the ' illusory '
perspective.

TABLE I.

Subject

C. S.

N. C.

A. W.

W. G.

C. R.

J. C

F.

F. A.

C. L

B.

M. W.

Perspective

L R

L R

L R

L R

L R

L

R

L R

L

R

L R

Preliminary period

44 56

61 39

48 52

37 63

40 60

57

43

55 45

59

41

37 63

Test period
(after fatigue of R)

81 19

58 42

62 38

37 63

62 38

58

42 47 53

87

13

41 59

/.

Figures represent percentage of total period of observation during which either
perspective was seen. L = 'left forward,' R = 'right forward.' In this case 'right
forward' was the fatigued perspective, so that fatigue, when present, should manifest
itself in an increase of L and decrease of R during the test period as compared with the
preliminary period.

Of nine subjects four showed fatigue much in the same way as
McDougall's observers. The other five however showed no signs of
any such fatigue, the relative amounts of the two perspectives in the
test period being much the same as in the preliminary period. In view
of these apparent individual differences it seemed desirable to repeat
the experiments with a larger number of subjects. As it was not found
possible to obtain further trained subjects for this purpose and as in
these observations, which in this respect were unlike most of the
experiments recorded in the previous paper, delicate introspection was
not essential, arrangements were made with a number of boys of
from 12^ to 14 years of age, who came to the Laboratory from a neigh-
bouring schools After a little practice these boys appeared to become

^ I am greatly indebted to Mr E. Rendell, Headmaster of Stanhope Street School,
and to Mr W. Welborne, Master at the same school, for their kindness in selecting
suitable subjects from among their pupils, and for making the necessary arrangements
with them as regards visiting the Laboratory.

J. C. Flugel 65

quite capable of making and recording their observations with suflBcient
accuracy, and with scarcely an exception approached the task with
considerable interest and goodwills They were however not informed
of the purpose of the experiments nor in any way led to conjecture as
to the nature of the results to be expected.

At the same time the experimental procedure was elaborated byi
the addition of a second fatigue-inducing period, during which the
' left forward ' perspective was fatigued, the effect of this fatigue bein^
measured in a second test period, just as the fatigue of the 'right
forward ' perspective had been measured before. This was done with
the help of a second _cross of similar construction to the first, driven by
a separate niotor and placed in such a position that, when viewed by
the subject, it presented precisely the same appearance when in its
* normal ' perspective as did the original cross in its ' illusory ' perspec-
tive. For the sake of uniformity, the second cross was also used when
fatigue was induced for the * right forward ' perspective. The relative
position of the two crosses in both cases will be made clear by the
accompanying simple diagram, where the arrangements correspond to
those represented in a and b'i Thus in a the cross used in the fatigue-
inducing period (' fatigue-inducing cross ') is so placed as to appear
similar to the cross used in the test period (' test cross '), when the
latter was seen ' normally.' On turning from the former to the latter
cross at the end of the fatigue-inducing period, the condition of the

^ In estimating the reliability of the results obtained with these subjects, we need not
however trust only to the general impression of the experimenter. An objective guarantee
of the reliability seems to be afforded by the two following circumstances : — (1) the
experiments were performed at least twice with each subject at an interval of over a week
and the results obtained from the two sittings are (as will be seen from Table II), with
only two exceptions, remarkably consistent with each other; (2) the same individual
differences and peculiarities as are exhibited by the boys are shown also by the adult
trained observers.

2 In this diagram it is of course in every case the real position of the crosses which is
indicated. By the aid of the diagram it is therefore easy to apprehend the appearance of
the crosses when in their normal perspective ; to realise their appearance during the
illusory perspective it should be borne in mind that both the perspective itself (i. e. right
or left side forward) and the direction of the movement appear reversed. It was however
only the test cross which could be seen in this illusory perspective, since this cross alone
was regarded uniocularly; the fatigue-inducing cross, being always viewed binocularly,
was (for all practical purposes) seen only in its normal perspective. As regards the mani-
festations of fatigue, it is evident that where the normal perspective of the two crosses is
the same (as in a and d), fatigue must show itself by an increase of illusory perspective of
the test cross; on the other hand, when the normal perspective of the two crosses is
different, i.e. when one is ' right forward ' and the other is ' left forward ' (as in b and c),
fatigue must be manifested by an increase of the normal perspective of the test crosB.
J. of Psych. VI 5

k

66 Fatigue in Illusions of Reversible Perspective

observer is such that the system corresponding to the ' normal ' per-
spective of the test cross has been in use for some time, while the
system corresponding to the ' illusory ' perspective is unfatigued. In h
the fatigue-inducing cross is placed so as to appear similar to the test
cross when the latter is seen in its 'illusory' perspective: it is now
therefore the ' illusory ' system which has been used and the ' normal '
which is unfatigued at the beginning of the test period. The complete
series of observations, including one preliminary, two fatigue-inducing,

o

a

1 '

O
h

O

0

1 //

o
d

% 1

o

e

A A

o

/

Diagram to show the various arrangements of the crosses used in the experiments.
1 =• Preliminary' and 'Test' Cross. || = 'Fatigue -inducing' Cross,

f = Direction of movement of upper part of Cross. O = Observer.

and two test periods, was made twice with each subject. At each
sitting there was an interval of not less than 35 minutes between the
end of the first test period and the beginning of the second fatigue-
inducing period. At the first sitting the ' right forward ' perspective
was fatigued first, at the second sitting the ' left forward.'

The results of these experiments are recorded in Table II. This
Table shows that the results obtained from subjects 1 and 2 were
irregular, but that the other subjects exhibit quite a high degree of con-
sistency between the results of the two sittings^ After the completion

* This consistency holds not only for the percentage of the two perspectives but also for
the average duration of the individual periods (not shown in the table).

J. C. FLuaEL

67

of the whole series of experiments, the first two subjects were each
given two more sittings, both of which produced results consistent with
themselves and with the second of the two previous sittings, so that we
may conclude that it was only in the first sitting that these subjects
showed insufficient training for our purpose.

TABLE II.

Subject

1

2

a

4

5

6

7

8

Perspective

L

R

L

R

L

R

L

R

L

R

L

R

L

R

L R

Preliminary

62

38

36

64

69

31

33

67

38

62

13

87

43

57

63 37

Test (after fatigue of R)

100

0

55

45

67

33

64

36

65

35

68

32

42

58

48 52

Test (after fatigue of L)

100

0

86

14

43

57

10

90

22

78

0

100

34

66

43 57

Preliminary

70

30

38

62

51

49

30

70

55

45

32

68

47

53

68 32

Test (after fatigue of R)

67

33

38

62

42

58

59

41

100

0

55

45

43

57

49 51

Test (after fatigue of L)

67

33

12

88

30

70

22

78

15

85

15

85

41

59

54 46

Subject

9

10

11

12

13

14

1

5

Perspective

L

R

L

R

L

R

L

R

L

R

L

R

L

R

^Preliminary

50

50

38

62

45

55

46

54

36

64

56

44

48

52

Test (after fatigue of R)

43

57

52

48

46

54

49

51

72

28

56

44

52

48

^Jlest (after fatigue of L)

47

53

43

57

48

52

35

65

41

59

54

46

58

42

Preliminary

55

45

53

47

42

58

44

56

43

57

48

52

50

50

Test (after fatigue of R)

42

58

54

46

45

55

46

54

62

38

48

52

45

55

Test (after fatigue of L)

42

58

54

46

60

40

27

73

46

54

49

51

51

49

Figures and indications as before. Fatigue of R corresponds to a in the diagram, fatigue of L to b.

As a result of these experiments, the fifteen subjects seem to fall
into three_fairly distinct, jclasses. The first and largest class, which
includes eight subjects (Nos. 1, 7, 8, 9, 10, 11, 14 and 15), comprises
thfise-JsJio exhibit no appreciable or only very slight signs of fatigue. _^
Into the second class fall three subjects (Nos. 4, 5 and 6) who show -^

5—2

68 Fatigue in Illusions of Reversible Perspective

distinct signs of Jfetigue in the case both of a and of 6. The remaining
subjects (Nos. 2, 3, 12 and 13), who constitute the third class, show
the usual signs of fatigue with ona arrangement, while manifesting no
fatigue or distinctly less fatigue with the other arrangement. Of these
subjects No. 13 shows the greater fatigue with arrangement a, and the
other three with arrangement h. Taking the two last classes together,
the proportion of fatigued to unfatigued subjects is fairly comparable
to that found with the adult trained observers, and the results from
both sets of experiments agree in showing that the p^enomenon_of
local fatigue, as brought out in these and in McDougall's experiments,
is by no means manifested, by all persons, those who do not manifest
fatigue_being, at least in the present case, in the majority.

It is of course possible that if the fatigue-inducing period had been
still further prolonged, fatigue would eventually have been manifested
even by the latter. But even if this were the case, it is evident that
these subjects possessed an amount of resistance to local fatigue, which
would make it quite impossible to account for the ordinary fluctuations

\\oi perspective, as observed by these subjects, in terms of such fatigue :

it is clear in fact that, with these subjects, local fatigue can play no

appreciable part in the reversals of perspective under ordinary conditions.

It would seem therefore that we cannot assign to local fatigue any

[very important part in the fluctuations of perspective, as these are
generally observed, since it is probable that in many persons it is not
generally present to any appreciable degree, and comes into play, if at

, all, only on occasions when, for any reason, one perspective has been
retained to the exclusion of the other, for a period that is very much
longer than usual.

Although local fatigue seems thus to be a factor of minor importance

only in illusions of reversible perspective, the phenomenon is, as will be

seen, of considerable interest on its own account and in connexion with

^^,i>'' the whole theory of fatigue.

P^ *ixt^ Perhaps the most curious fact indicated by the experiments upon

1/^ ^' the boys, which we have just described, is the apparent co-existence in

i*^'^ tf the same individual of marked fatigability to one perspective with very

small^or inappreciable fatigability to the other. On repeating the

experiments upon fatigability to both perspectives with some of the

trained observers, upon whom we had experimented in the first place,

and with one other who had since become available, we discovered two

cases in which one perspective was more fatiguing than the other, and

one case of fatigability to one perspective combined with apparent

J. C. Flugbl 69

complete non-fatigability to the other. As in the case of the boys, the
individuals differed as to which_perspective was found the more fatiguing.
The fatiguing perspective however remained constant for the same in-
dividual, even when the experiments were repeated after a very con-
siderable interval of time (over two months in the case of two of the
adult subjects), showing that the greater fatigability to one perspective
was not due to any merely temporary cause bivL was a permanent
characteristic of Jbhe individual.

In view of this corroboration of the interesting phenomenon of one-
sided fatigue, there arises the question : — Is it possible to determine
more precisely the exact conditions of this difference of fatigability ?
Perhaps the most striking difference between the arrangements shown
as a and b in the diagram is that in the one case (a in diagram) fatigue
manifests itself in an increase of the ' illusory ' perspective of the test
cross, in the other (b in diagram) in an increase of the 'normal' perspective.
It might be supposed at first sight that the apparent differences of
fatigability are really due to differences in the ease with which fatigue
can be manifested, according as it results in an increase of * normal ' or
an increase of ' illusory ' perspective. The fact that some observers are
more fatigable with arrangement a, others with arrangement b, shows,
of course, that there is no universal tendency for an increase of ' normal '
to be produced more easily than an increase of ' illusion ' or vice versd.
It remains possible however that such a tendency may exist in anyf
particular subject, since we have not yet shown that fatigue may
result sometimes in an increase of ' normal,' sometimes in an increase of/
' illusion ' in the same observer.

The proof of this however is afforded by some experiments with the

two new arrangements depicted in c and d (diagram, p. 66), where the ^/^
test and fatigue-inducing crosses have been transposed, the test cross
being now to the left of the observer and the fatigue-inducing cross to
the right. As a result of these new observations it was found that the /
two subjects who displayed greater fatigue in b than in d also displayed I
greater fatigue in d than in c, though fatigue resulted, in the case of 6,
in an increase of ' normal,' in the case of d, in an increase of ' illusion.'
The remaining subject, who had shown distinct fatigue in a combined
with absent or inappreciable fatigue in b, showed, correspondingly,
distinct fatigue in c and little or no fatigue in ci : in this case the
fatigue manifested itself in a by an increase of ' illusion,' in c by an
increase of ' normal.' The numerical results obtained from the three
subjects with all four arrangements (a. 6, c and d) are shown in Table III.

70 Fatigue in Illusions of Reversible Perspective

TABLE III.

Arrange-

Test Cross,

Arrange-

Test Cross,

ment

1

right

ment

kit

Subject : J. K.

L

R

L

R

Preliminary

56

44

39

61

Test (after fatigue of K)
Test (after fatigue of L)

a

80

20

c

100 (78)

0

h

0

100 (162)

d

0

100 (240)

Subject : A. W.

Preliminary

31

69

45

55

Test (after fatigue of R)

a

67

33

c

62

38

Test (after fatigue of L)

b

0

100 (120)

d

0

100 (97)

Subject: C. S.

Preliminary

47

53

47

53

Test (after fatigue of R)

a

85

15

c

100 (131

0

Test (after fatigue of L)

b

45

55

d

50

50

(2nd series)

Preliminary

50

50

48

52

Test (after fatigue of R)

a

71

29

e

89

11

Test (after fatigue of L)

h

53

47

d

37

63

Main figures as before. The figures in brackets represent the actual duration in
seconds of the appearance of any particular perspecti%'e, when this perspective was visible
during the whole period of observation. In all such cases this period was prolonged until
a change of perspective did at length occur. The letters (a, b, c, d) refer to the arrange-
ment of the crosses with reference to the observer (see diagram p. 66).

These results seem to afford fairly conclusive evidence that the pheno-
menon we are considering is not due to any difference in the ease with
which fatigue can manifest itself according as it results in an increase
of ' normal ' or of ' illusory ' perspective during the test period. We
may therefore assume that our results are due to genuine differences of
fatigability.

Turning now to the actual perceptive elements concerned in the
reversals, it would seem that they can be split up into two mdependeot
factors : — (1) the perspective proper (i.e. ' right ' or ' left forward ') [that
this is an independent factor in the reversals is shown by the fact
already mentioned, that reversals can take place in the absence of any
movement of the cross] ; (2) the direction of^themoveraLeiit (e.g. the
top of the cross can be moving towards or away from the observer).

Let us consider first the part played by the perspective in the mani-
festations of fatigue. A study of Table III and of the diagram will
show that the relations of fatigue to this factor of the perspective

J. C. FLtJaEL 71

remain constant in each of the three subjects. With C. S. fatigue
is manifested on every occasion when ' right forward ' is seen during
the fatigue-inducing period (a and c), while fatigue does not appear
when ' left forward ' is visible during this period (b and d). With
J. K. and A. W,, on the other hand, greater fatigue is shown when * left
forward' is seen during the fatigue-inducing period (6 and d), less
fatigue when the ' right forward ' is seen (a and c). The presence oh
this constant relation between the fatigue and the perspective seems to '
indicate that this factor of the perspective is, as we might expect, an
essential element of the fatigable system.

As regards the second of the two above-mentioned factors, it will
be seen that there is no such constant relation to the fatigue as in theij
case of the perspective. Thus in a the movement of the fatigue-
inducing cross is towards the observer, in c it is away from the observer,
though C. S. manifests fatigue equally in both cases. The movement
similarly differs in the case of b and d, though these two arrangements
correspond in giving the greater fatigue with observers J. K. and A. W.
If the direction of the movement were the essential factor, we should
expect a to correspond with d and b with c, whereas with all three
observers the correspondences actually found are between a and c and
b and d, where (as we have already seen) the perspective and not the
movement is the common factor. Our observations seem then to
indicate conclusively that (in the present cases at least) it is the
perspective of the cross, and not the direction of the movement, which
is princijDally concerned in these differences of fatigability. t /

There remains the question whether the direction of the movementf f/^
gives rise to any fatigue at all, similar in nature to that brought about
by the perspective. The results recorded in Table III, as we have just fif

seen, fail to show any manifestations of such fatigue. It would appear
possible however to put the matter to a further test. If the direction
of the movement plays no part whatever in the production of fatigue,
we should expect that the manifestations of fatigue brought about by
the perspective would be independent of whether they resulted in an
increase of movement in one direction or in the other during the test
period.

For the purpose of applying this test the crosses were arranged as
in e and /, which are similar to a and c respectively, except that the
upper part of the test cross is moving away from, instead of towards,
the observer. In a fatigue was produced for the 'right forward'
perspective and manifested itself in an increase of ' left forward,'

72 Fatigue m Illusions of Reversible Perspective

together with movement away from the observer, both perspective and
movement being reversed. If now the perspective is the only element
concerned in the fatigue, we should expect a similar increase of ' left
forward ' in e. In the case of e however the reversal would affect the
perspective only and not the movement, since * left forward ' (which in e
is the 'illusory' perspective and therefore involves an apparent move-
ment in the opposite direction to that in which the cross is really
moving) must be accompanied by movement of the top towards the
observer, i.e. in the same direction as the movement seen during the
fatigue-inducing period. If, on the other hand, reversal of movement,
as well as of perspective, is in any way important for the manifestation
of fatigue, we should expect that no fatigue would be shown in e, or at
any rate that it would be exhibited to a less marked extent than where
(as in a) both perspective and movement can be reversed simultaneously.
Precisely similar considerations apply to /, except that in this case
fatigue of the perspective would manifest itself (as in c) by an increase
of the * normal ' instead of the ' illusory ' perspective in the test period.

Some results obtained with^subject C. S. with arrangements e and/,
together with (for the sake of comparison) a further record with arrange-
ment a, taken at the same time, are given below : —

a

e

/

L

R

L

R

L R

Preliminary

45

55

50

50

47 53

Test

94

6

46

54

49 51

It will be seen that no fatigue is manifested with e or /, though
there is, as before, well marked fatigue in the case of a. These results
seem to indicate fairly clearly that the different constituents of the
total perception corresponding to the perspective and the movement
respectively are both independently but simultaneously fatigued, and
that the effects of these two separately fatigued elements neutralise one
another in the test record. Thus in / fatigue to the perspective would
tend to make the test cross appear in the ' normal ' phase ; fatigue to
the movement, on the other hand, would simultaneously tend to produce
the * illusory ' phase.

It may appear perhaps at first sight that this interpretation is
contradicted by the results obtained from the same observer with
arrangements c and d, which, as we saw, failed to exhibit any mani-
festations of fatigue due to the direction of the movement, though in
these cases there was apparently nothing to prevent the manifestation
of such fatigue, had it existed. It must be admitted that the results

J. C. Flugel 73

obtained with arrangements e and / would have been more easily
explained, if there had been some manifestations of fatigue with h and
d also. The absence of such manifestations in the latter case is not,
however, incompatible with the existence of fatigue due to the move-
ment, such as seems indicated by the results with e and /, if we assume
(as would seem probable in any case) that there exists a definite and
more or less constant threshold, which must be passed, before fatigue
can manifest itself by an increase of one or other of the two perspectives.
We may then suppose that in the present observer the fatigue caused
by the direction of the movement is subliminal, and therefore fails to
manifest itself in h and d ; that the fatigue caused by the perspective
'right forward,' which, as we saw, is with this subject greater than that
caused by 'left forward,' is supraliminal, either by itself or when
combined with that due to the movement, and therefore manifests
itself in a and c ; but that this fatigue of ' right forward ' is no longer
supraliminal when, as in e and /, its effects are no longer strengthened
but opposed by the fatigue due to movement. This will be readily
understood, if expressed in actual figures. Let us suppose that the
threshold is passed, and fatigue therefore manifested, as soon as the
fatigue effects due both to inclination and to movement reach a total
of 30. Let us suppose further that the fatigue of ' right forward '
produced in five minutes amounts on any given occasion to 20, that
of 'left forward' to 10, and that of the movements towards and away
from the observer (between which there is, in the present case, no
reason to assume the existence of any considerable difference) in both
cases to 15. Then in the case of a and c the total fatigue will be
20 + 15 = 35, being above the threshold ; in the case of h and d it will
be 10 + 15=25, which is below the threshold. In e and /the effects
of the fatigue due to perspective and the movement respectively are
working in opposite directions ; therefore, instead of 20 + 15, as in
a and c, we get 20— 15 = 5, which is, of course, below the threshold.
The equation thus works out correctly in every case, affording very
considerable confirmation of the view here advanced. Our data of
course afford us no indications of the actual relative fatigability of the
different systems, except such as can be deduced from the facts that
(in this case) fatigue to ' right forward ' plus fatigue to movement is
above the threshold, while fatigue to ' right forward ' minus fatigue to
movement and fatigue to ' left forward ' plus fatigue to movement are
both subliminal ; there are, of course, a very large number of possible
relations between the systems which will meet these conditions.

/

74 Fatigue in Illusions of Reversible Perspective

JJ^aturally I am aware that this view may have to be greatly
modified as the result of further observations : I have advanced these
considerations chiefly with the object of showing that the results
obtained with h and d and with e and / are not necessarily mutually
contradictory, as they might perhaps at first sight appear.

Summing up the results from this series of observations, we may
then conclude : —

(1) That the phenomenon of one-sided fatigue is not due to
any difference in the ease of manifestation, according as it results in
an increase of * normal ' or an increase of ' illusory ' perspective.

(2) That the greater fatigability to one perspective is (at least in
our present three observers) chiefly connected with the perception of
the perspective of the cross {i.e. ' right ' or ' left forward ').

(3) That the perception of the direction of the movement is never-
theless, in all probability, independently fatigable.

One further experiment may be briefly referred to before concluding
this paper. Since our data all tend to indicate that the fatigue we are
here considering is extremely local and specific in nature, it would be
interesting to know how far this local fatigue is affected by other
simultaneous mental processes. We therefore determined to see
whether the amount of attention given to the cross during the fatigue-
inducing period had any influence on the amount of the fatigue. For
this purpose one or two distraction experiments were made, in which
the subject (in this case A. W.), while fixating the cross in the usual
way, was asked to give his attention to some other matter during
the fatigue-inducing period. Two kinds of distraction were tried: —

(1) attentive listening to a story read aloud by the experimenter,

(2) adding 3's aloud as quickly as possible. The first produced a mild,
the second a very high degree of inattention to the cross. The results
obtained in this way were as follows^ : —

Listening to

story Counting 3's No distraction

L E L E LB

Preliminary 65 35 66 34 58 42

Test (after fatigue of E) 90 10 77 23 85 15

Test (after fatigue of L) 13 87 0 100 (72) 0 100 (69)

It is fairly evident that the fatigue is unaffected by the distraction.

* It should be noted that these results agree with those for the same subject in
Table III in showing greater fatigue for ' left forward ' than for ' right forward,' though
these later- observations were made after an interval of over a month.

J. C. Flugel 75

We must conclude therefore that the local fatigue, with which we are
here dealing, is independent of the general direction of the cerebral
energy, and runs its course whenever the particular perceptive element
is called into play, without reference to the way in which the rest of
the conscious energy is occupied ^

Perhaps the chief point of interest that emerges from all our
observations on local fatigue, as manifested in the ' revolving cross '
illusion, is the extremely specific nature of this fatigue. This speci-
ficity is clearly iudicated by the following considerations: — (1) that)-
fatigue is manifested in some individuals and not in others : (2) that in
the same individual one perceptive system may be highly fatigable,
while another very similar system may be much less fatigable, or even'ii
show an altogether inappreciable degree of fatigue : (3) that individuals
differ as to which of these two perceptive systeuis is the more fatigable :
(4) that this fatigue is independent of the simultaneous activity of
other centres. It would seem that this extreme specificity of certain
kinds of fatigue cannot but be of very considerable physiological and
psychological interest. The fact that one perceptive system {e.g. ' right
forward ') can differ considerably from another apparently very similar
perceptive system {e.g. 'left forward') in such an important matter
as fatigability seems to indicate that there must exist important
differences of structure or function between neighbouring and closely
connected nervous paths. This again seems to point to the probably
very specific nature of certain functions and abilities — a conclusion
which, we may note, is in harmony with the results obtained from
much work that has been done on ' mental tests ' in recent years, and
which is also of importance in view of the problems connected with
the parts played by ' general ' and ' specific ' factors in these tests and
in mental performances generally I

As regards the question with which we are ourselves more im-
mediately concerned, i.e. the relation of this local fatigue to the

I It is interesting to compare this result with that obtained from similar observations
on the after effect of movement in a plane at right angles to the line of vision. (See
A. Wohlgemuth, " On the After Effect of Seen Movement." This Journal, 3Ionograph
Supplements, No. 1, p. 83.) In neither case has distraction any influence on the after
effect, though the three-dimensional space perception involved in the present experiments
is a ' higher ' process than that concerned in Dr Wohlgemuth's observations, in so far as
it is acquired by experience and not innately organised, as is probably the case with the
perception of movement in two-dimensional space.

^ See, for instance, Bernard Hart and C. Spearman, "General Ability, its Existence
and Nature." This Journal, 1912, v. 61.

76 Fatigue in Illusions of Reversible Fersjyective

principal factor in the reversibility of perspective, namely, the direction
of the attention, I am inclined to believe that the two factors work as
a rule in almost complete independence of one another, and that, at
any rate, they ha,ve little or nothing in common in their nature. As
already shown, the manifestations of local fatigue vary too much both
as regards the differences between individuals and the differences
between difiFerent perceptive systems in the same individual to allow us
to suppose that they underlie the involuntary changes in the direction
of the attention observed in every one under normal circumstances.
1 There is reason to believe moreover that the factors concerned in the
voluntary maintenance of attention are very general; while local
fatigue is, as we have seen, extremely specific. Nor is there any
indication that there exists any correlation between individual pecu-
liarities as regards the one and as regards the other. We have
shown in the earlier paper that there exist considerable individual
variations in the power of controlling the movements of attention,
which result in changes of perspective. Those who show local
fatigue are to be found both among those who exhibit high (e.g.
A. W.) and low {e.g. C. S. and C. R.) power of controlling the
attention. Similarly, as regards those who do not show local fatigue
{e.g. N. C. and J. C. F. respectively). In general, so far as the data at
present available enable us to judge, everything seems to show that
there is very little connexion between the two factors, which are
also, apparently, of very different importance as regards illusions
of reversible perspective. The one is, it would seem, continuously
operative in all cases of reversible perspective ; while the other, in
all probability, only comes to exert an appreciable influence with
certain persons, and under certain special conditions (such as that of
our long ' fatigue-inducing period ') not present in the majority of
observations on these illusions.

Although the phenomenon of local fatigue seems thus to play only
a minor part in illusions of reversible perspective, the facts concerning
this local fatigue brought out by the present observations are, I venture
to think, not without very considerable general interest. Besides the
facts indicating the extreme specificity of the fatigue, to which allusion
has already been made, there arise a number of further problems
connected with the subject of local fatigue, one or two of which may be
mentioned in conclusion. Thus it is evident that interesting problems
are afforded by the questions as to the frequency with which other
cases of this local fatigue are to be met, the levels of consciousness

J. C. Flugel 7?

at which they are most prevalent^ the importance and complexity of
the functions they affect and their relations to the wider phenomenon
of 'general fatigue.' These problems too will, in all likelihood, afford
no insuperable obstacles to experimental investigation, and, in so far as
they are solved, can scarcely fail to throw very considerable light upon
the whole nature of fatigue.

Summary.

Experiments on the ' windmill ' or ' revolving cross ' illusion confirm
McDougall's observation that a prolonged period of binocular regard
may produce fatigue of the aspect seen during this period, the fatigue
being manifested in an unusually long period of the opposite aspect as
soon as uniocular is substituted for binocular regard.

This fatigue however is not manifested by all the subjects of the
present experiments.

Among those subjects who manifest fatigue, there are some who
show greater fatigue with one aspect of the cross than with the other.

In these cases the differences of fatigability seemed to reside in the
perception of the perspective, though there is evidence to show that the
perception of the perspective and of the direction of the movement are
both independently fatigable.

The more fatigable aspect varies from one individual to another, but
remains constant for each individual.

The manifestations of fatigue are independent of whether they
result in an increase of the ' normal ' or of the * illusory ' perspective.

They are likewise independent of the amount of attention given to
the cross, being unaffected by other simultaneous mental processes
during the induction of fatigue.

These facts indicate that the local fatigue with which we are here
concerned is highly specific in nature.

It is probably unconnected with the factors determining the
direction of the attention, which (as was shown in a previous paper)
is the main condition of the reversals of perspective under ordinary
circumstances.

It is however of considerable interest on its own account.

^ As already mentioned McDougall has shown that they are to be found on the purely
sensory level.

{Manuscript received 25 March 1913.)

BINOCULAR AND UNIOCULAR DISCRIMINATION
OF BRIGHTNESSi.

By shepherd DAWSON.
(^From the Psychological Lahoratory^ University oj Glasgow.)

I. Description of a/pparatus.
II. Method of conducting the experiments.
III. Tabulated results, shouting the differences between binocular
and uniocular discrimination, as mea,sured (i) by the
frequency with which the grey to be discriminated was
located unth accuracy and certainty, and (ii) by the time
taken in discriminating it.
IV. Explanation of results; discussion of practice and the
summation of the brightnesses of the uniocular images
OS possible explanations; an introspective basis for an
eocplanation ; suggested explanation; additional experi-
mental evidence in support of it.

The following investigation has been made for the purpose of
finding whether there are any differences between binocular and uni-
ocular discrimination of shades of grey, and, if so, whether any experi-
mental data can be found which will explain them. The experiments
consist essentially in presenting to different subjects a grey ring of
variable and measurable intensity on an otherwise uniform disc, and in
asking them to discover its position and to describe as fully as possible
its appearance.

I. Description of Apparatus.

The apparatus consists of an aluminium sector of angle 60° and

radius 13 cms. rotating in front of a milk-glass screen behind which is

a Kamm's incandescent lamp. Parallel to the edge of the sector, at a

1 An abstract of this paper was read at Dundee before Section I of the British
Association for the Advancement of Science, September, 1912.

Shepherd Dawson

79

Figs. 1—5.

80 Discfimination of Brightness

distance of 2*1 cms. from it, is a groove in which slides a screw passing
through a rectangular strip A (see fig. 1) which can be placed anywhere
along the edge of the sector and made to project beyond it to any
extent up to 2 cms., and can be fixed in position by means of a screw-
nut (see fig. 3). The groove is covered in front by a thin sheet of
aluminium so that no light can pass through the sector. The adjusting
screw is, therefore, at the back of the sector. The width of the strip is
1 cm. It is set in position and the amount of its projection beyond the
edge of the sector is measured by means of the micrometer screw-gauge
shown in fig. 4. The screw-gauge is used in the following manner: it is
adjusted so that the distance of the surface D from the surfaces BB is
equal to the amount which it is required that the strip should project ;
it is then placed so that the surfaces BB rest against the edge of the
sector and the strip is pushed up to the face D and fixed there. In
the first few experiments the projection was always re-measured after
rotation of the sector, but after a little practice in the use of the
apparatus had been gained this was found to be unnecessary.

The sector is fixed to an axle rotating on pivot points and is driven
by a thin string passing round the groove G (fig. 2) and a driving wheel
(not shown in fig.). The axle is held in position by two rigid horizontal
bars, part of a solid framework of cast iron (fig. 5). On the side of the
axle opposite to the sector is a screw on which are double nuts which
act as a counterpoise {E, fig. 2). When the screws are carefully adjusted
there is practically no vibration. •

This framework is placed in a wooden box with a circular opening
in front, so that only the disc of light over which the sector rotates can
be seen. A milk-glass screen drops into a groove at the back of this
box, where it is held in position by small clips, and whence it can be
easily lifted out.

On the side of the projecting bar nearest the observer is gummed
an indented strip of black paper the teeth of which project above the
arm and stand out clearly against the bright background. The points
of the teeth are at distances 5, 6, 7, 8, 9, 10, 11 aud 12 cms. fi-om the
centre of rotation, and, to make them distinguishable from one another,
they are made alternately high and low.

In front of the screen is a long, five-sided rectangular box which
cuts off all light except that passing through the screen. The end
nearest the observer is pierced by an elliptical hole just large enough
to allow him to see the disc comfortably with both eyes: when only one
eye is used this opening is covered by a piece of cardboard in which is a

Shepherd Dawson

81

circular hole 2*5 cms. in diameter. This screens oflf the observer's face
from the rays passing through the disc, and so prevents reflection of light
by the face ; it has the further advantage of obviating the necessity of
using an eye-cover which is disconcerting to many subjects. From a
sheet of cardboard projecting over the head of the observer hangs a
curtain of black cloth, which entirely excludes all light except that
which passes through the screen. The whole of the apparatus — sector,
iron framework and box — is painted dull black.

If the sector be rotated at a high speed when the strip A projects
beyond its edge, there appears on the lighter background a grey ring
the relative brightness of which can be varied by varying the amount
of projection of the strip. By showing rings of different intensities
and asking the observer to detect and locate them, the threshold for
brightness discrimination can be found.

When we know the angle of the sector (60°), the distance of the
projecting strip from the centre of the disc, and the amount it projects
beyond the edge of the sector, the relative intensities of the ring and
the background can easily be found. They have been calculated for
the projections shown in Table I (the only projections used in the
experiments), but only their ' diflference-ratios ' are here shown, i.e. the
ratio of the difference of intensity of the ring and background to
the intensity of the background (lower difference-ratio) or to that of
the ring (upper difference-ratio).

p R

Q s

Fig. 6.

As one edge of the strip is parallel to that of the sector and the
others are perpendicular to it, the grey band that is formed by rotating
it is darker at its inner edge than at its outer, and these edges are not
so sharp as they would be if the strip were an arc of a circle. In fig. 6
let PMNQ represent the projecting strip, and OS the edge of the
sector, and let TP, MR, VQ, and NS be arcs of circles with centre 0 at
the centre of the disc. It is evident from the figure that the ring
formed by rotating PMNQ about 0 will be darkest along the circle of

J. of Psych. VI 6

82

Discrimination of Brightness

which MR is an arc, and will gradually increase in brightness to the
part formed by rotating VQ. From P to R and from S to Q there is a
very rapid fall in brightness. As, however, PR is never greater than
0*404 mm. and in these experiments did not subtend an angle of more
than 86" at the eye of the observer, this is unappreciable. It is slightly
greater than the maximum angle of discrimination, which Helmholtz
found to be 65""75. No blurring of the edges could be detected in the
largest and darkest ring, in which, of course, PR has its greatest
value.

TABLE I.

Radius

10

9

8

7

6

Lower ratios

Upper
ratios

A

B

C

Amount of
Projection

•10
•20
•30
•40
•50
•60
•70
•80
•90
1^00

•09
•18
•27
•36
•45
•54
•63
•72
•81
•90

•08
•16
•24
•32
•40
•48
•56
•64
•72
•80

•07
•14
■21
•28
•35
•42
•49
•56
•63
•70

•06
•12
•18
•24
•30
•36
•42
•48
•54
•60

•0019
•0038
•0057
•0076
•0095
•0114
•0133
•0152
•0171
•0191

•0017
•0035
•0052
•0069
•0087
•0104
•0122
•0140
•0156
•0174

• ^0019
•0038
•0058
•0077
•0096
•0116
•0135
•0155
•0174
•0194

All radii and projections are given in centimetres. On the top row are the distances
of the inner edge of the strip from the centre of the disc : the columns below show the
amounts of projection used at these distances ; in the last three columns are the ' upper
and lower difference-ratios ' obtained under these conditions. Column A shows the ratio
of the difference of intensities of background and of the darkest part of the ring {MR,
fig. 6) to the intensity of the background. Column B shows the ratio of the difference of
intensities of background and of the lightest part of the ring {VQ, fig. 6) to the intensity
of the background. Column C shows the ratio of the difference of intensities of back-
ground and of the darkest part of the ring to the intensity of the darkest part of the ring
{MR, fig. 6).

The principal advantages of this apparatus are the accuracy with
which the intensities of the light can be measured, the possibility of
varying the position of the ring, and the elimination of the inhibitory
effect of a dark background by placing the grey to be detected in the
centre of a large area of light.

The micrometer screw-gauge used for measuring the extent to
which the strip projects beyond the edge of the sector reads to 0*01 mm.
As the amount of projection used in these experiments varies from

Shepherd Dawson 83

0"06 cm. (with radius 6 cms.) to 1 cm. (with radius 10 cms.), if an
error of 0"01 mm. (a very large and unnecessary one) be made in
reading the screw-gauge, and if an error of 05 mm. be made in
adjusting the projecting strip to the lines on the sector marking points
6, 7, 8, 9, and 10 cms. from the centre, the largest possible error in the
values of the difference-ratios shown in Table I will be about 0*00009.
With moderately careful experimenting the values shown in the table
are, therefore, practically correct to the fourth decimal place : the fourth
figure cannot be more than one unit wrong.

The advantage of varying the size of the ring is so patent that little
need be said about it. In these experiments the moveable strip was
placed in irregular order in one of five positions, viz. 6, 7, 8, 9, or 10
cms. from the centre of the disc. The subject was not told that it
would be placed exactly at these points, nor did he know whether a
plain disc might be shown ; so that even if he did not see the ring, but
guessed its position, the chance of his guessing it correctly was very small.
Therefore, when he located it correctly, it was probably because of some
perceived darkening of the disc produced by the projecting strip.

The advantage of placing the grey to be detected in the centre of a
large area of light is well brought out by this apparatus. Along the outer
edge of the disc there is a perceptible brightening of the field due to
contrast with the dark background : it extends also along those parts
of the disc adjacent to the projecting bar which carries the axle on
which the sector rotates. This brightening has an interesting effect on
the appearance of the ring. All the observers remarked that when the
ring was very faint it could be seen most easily on the left side of the
field, and that unless it was very clear it could not be seen on that
portion of the disc which is adjacent to the bar. At first they generally
tried to find it on the right for the purpose of locating it correctly, but
on failing to discover it there they explored the whole of the disc and
often saw it on the left. For this reason after a few observations had
been made most observers looked first at the left half of the field, and,
on discovering the ring there, tried to follow it round to the right in
order to find its exact position. Now, on the left, there is a very large
area of light, while on the right the disc is crossed by the projecting
bar. The blackness of this bar stands out in marked contrast to the
light grey of the disc, and, not only does it make the adjacent parts of
the disc brighter than the rest, but apparently it exercises a strong
inhibitory influence on the ring, making it less clearly visible there
than anywhere else. One observer remarked that the ring seemed to

6—2

84 Discrimination of Brightness

pass behind a screen before it reached the scale on the cross-bar, so that
it appeared to end about half-an-inch above and below it. That this
effect was not due to any irregularities in the milk-glass screen was
proved by reversing the screen, so that the part which formerly stood
on the left was put on the right; under these conditions the phenomenon
still remained. Nor was it due to not placing the source of light on
the line running through the eye of the subject and the centre of the
disc, for the position of the lamp was carefully measured before each
series of observations was made.

An investigation of the exact extent of this inhibitory effect and an
attempt to explain it are unnecessary here and would unduly complicate
the subject of this paper; but, however the phenomenon may be
explained, it is one which must be reckoned with. In these experiments
it was allowed for by placing the projecting strip at points from 6 to 10
cms. from the centre of the disc, so that there was always a space of not
less than 2 cms. between the outer edge of the ring and the outer edge
of the disc, and a space of not less than 1 cm. between the inner edge of
the ring and the outer edge of the very faipt ring formed by that
portion of the screw which projected beyond the counterpoise.

II. Method of conducting the Experiments.

As the object of this investigation was qualitative as well as quanti-
tative, being the discover}? and explanation of any differences there
might be between uniocular and binocular discrimination of brightnesses,
full introspection was demanded of the subjects ; and as the dictating
and writing of these introspections took some time, only a few readings
could be taken at each sitting, viz., about 20 per hour. For this reason
and because the time of both experimenter and subjects was somewhat
limited, it has been impossible to make the large number of observations
required in using the method of constant stimuli, — the method which
is generally used in investigations on sensory discrimination. Only a
comparatively small number of short series has been made with each
subject, but as each observation was made with the utmost care and
described in great detail, it is hoped that the fulness of the intro-
spections will throw more light on the problem than the mere mathe-
matical treatment of thousands of affirmative and negative judgments.
The method of conducting the experiments was as follows :
The approximate position of the threshold of discrimination was
found by showing a few rings of different intensities. Then series of

I

Shepherd Dawson 85

projections were selected so that they gave eight or nine rings, the
intensities of which formed an arithmetical series, those of the highest
intensities being imperceptible and those of the lowest clearly perceptible.
These rings might be of any radius from 6 to 10 cms. In each series
the radii of the rings were varied in irregular order, care being taken,
however, that as far as possible rings of each intensity appeared as often
in one position as in another; so that the subject had no means of
finding the position of a ring other than from the visual presentation.

With three subjects the intensities in each series were gradually
decreased or increased. When this method, — a form of the method of
minimal changes — was used, the length of the series was varied from
time to time so that the subject should not learn to expect a change of
judgment at any particular point in the series. One advantage of this
method is that comparisons are avoided which are often made involun-
tarily when large and small differences are presented in irregular order.
Such comparison tends to increase the effect of large intensity-
differences and to decrease that of smaller ones. The gradual decrease
or increase makes possible an adaptation which, though perhaps not
entirely sensory, is comparable in its effects to the well-known pheno-
mena of adaptation to bright and dim illumination.

To one of the subjects the different intensities were presented in
irregular order, but in such a way that each was given approximately
as often first as second, third, etc., and as often of one radius as of the
others. This is simply the method of constant stimuli. More observa-
tions were made by this subject than by the others. The order of
presentation of the stimuli in all the series was determined after the
preliminary series had been made in which the approximate threshold
was found.

The objection which is usually raised against the method of minimal
changes, viz., that the subject's judgments are influenced and in some
cases entirely determined by the fact that he knows that the stimuli
are decreasing or increasing regularly in intensity, cannot be raised
against the modification of this method which has been used here. At
each observation the position of the ring has to be found, and the
knowledge that its intensity is regularly increasing or decreasing can
do nothing more than prepare the subject for the kind of ring to look
for. He was not told how the intensities would vary and, although he
soon discovered that they were decreasing or increasing, he thought
that the change was somewhat irregular. This was due partly to the
limited range of the intensities and partly to the necessity of giving

86 Discrimination of Brightness

a full account of what had been seen, which tended to distract the
attention from everything but the observation of the moment, and so
to some extent prevented side-comparisons and reflection on the method
of varying the stimuli. The general character of the judgments given
in the regular and the irregular series was the same.

The series were arranged so that the subject looked alternately with
both eyes and with one: when a ring of a given intensity had been
observed with both eyes, another, of different diameter but like physical
intensity, was observed with one eye (always the better one); then two
other rings of slightly less intensity were similarly observed, and so the
series was worked through. In this way the effects of practice, fatigue,
and other varying influences were equally distributed over the uniocular
and the binocular series, and a point to point comparison between them
became possible.

It might be urged against the alternation of uniocular and binocular
observations that it might be distracting to the subject. The intro-
spective records, however, show no evidence of this, and the subjects
were not aware of any distraction that could be attributed to this cause.
If such distraction existed, it would affect both series equally, and
would therefore be negligible.

The subject was directed to look at the disc when he heard the
signal ' Now,' and then to look for the ring and find its position as
accurately as possible by means of the teeth on the bar which stretched
half-way across the disc, the distance of each tooth from the centre of
the disc having previously been learned. He was asked to give a
signal when first he saw the ring, and to note its appearance as carefully
as possible. The experimenter fixed the projecting strip in the required
position and set the sector rotating. When it was rotating quickly
enough two signals were given, viz., * Ready ' and ' Now,' separated by
an interval of about two seconds. At the same time that the second
signal was given the experimenter set in motion the seconds-hands of a
stop-watch, by which he measured the interval that elapsed before the
signal was given which indicated that the ring had been seen. Thirty
seconds were allowed for each observation : at the end of that time
the experimenter called out ' Stop ' ; whereupon the subject ceased to
look at the disc and began to describe what he had seen. He continued
to examine the disc during the whole of the thirty seconds, whether the
ring had been located or not. This interval was chosen because it
allows sufficient time for searching the disc, and is not long enough to
produce any strain.

Shepuerd Dawson 87

During the first few observations the subject was only instructed to
locate the ring and to make what remarks he could about its appear-
ance ; but, as the work proceeded, his introspection became more
detailed and his attention was directed to the following questions: —

When did you first see the ring, or any part of it ?
Did you see it continuously, or did it fluctuate ?
Did you see the whole of the ring at any one time ? How often ?
Could you see any part of it you fixated ?
Was it clearly marked ? Were both edges distinct ?
Where did you see it first ?
Where did you see it most easily ?

How did it first appear ? Did it stand out suddenly and involuntarily, or did it
gradually develop as you looked at it ?

Are you sure you saw the ring, or part of it ?

Are you sure you have located it correctly 1 Why ?

He was allowed to give his introspection as he pleased ; no attempt
was made to direct it by requiring an answer to each of these questions
in turn. In this way it was hoped to bring out the parts of each
experience which struck him most forcibly. Occasionally after he
had given as full an account as he could, he was asked one or two
of these questions. On rare occasions all the questions were read
through before beginning a series : further direction of attention to
them was found unnecessary. The subject's attention was not directed
from the first to the above questions because that would have interfered
seriously with his observations : for him the most important part of the
observation was the detection and location of the ring, and it was
considered advisable in the first few experiments to let him attend
specially to that, and afterwards, as he became practised, to encourage
a more and more complete analysis of his presentations.

The subject sat at a constant distance (120 cms.) from the screen,
and the lamp which illuminated it was placed at the same distance
from it on the opposite side.

Each series lasted from three-quarters of an hour to an hour,
according to the length of the introspection. Usually only one series
was given at each sitting, but sometimes, when the introspections were
short, two were given. Generally, the hour at which the observations
were made was the same for the same subject. In two cases two
sittings were given per week, in the others more ; with all subjects the
interval that elapsed between two consecutive series varied.

The experiments were performed with four subjects, Messrs Paul,
Anderson, Robieson, and Craig, all graduates who had had considerable

88 Discrimination of Brightness

experience in psychological observation. To these gentlemen I am
deeply indebted for the time and care given in making these observa-
tions, and to Dr Watt for his invaluable advice and assistance.

III. Tabulation of Results.

The introspective records vary considerably in fulness of detail ;
they invariably increase in length as the subject gains experience in
the task before him : as a rule, they are shortest when the ring is very
clear and when it is not seen. The following are typical examples : —

Subject, R. Inner radius of ring, 6 cms. Diff.-ratio, 0"0095. Both eyes.
"When I called out (after 6 seconds) I thought I saw a circle about the middle
of the left side of the disc. It was very vague and indistinct, and I was not at all
sure that it was the circle but fixation made it clearer. I could follow it round
over the top, but with considerable difficulty, and, so far as I could make out, the
inner edge was at 6. I could see it only by careful fixation, and even then only
vaguely : still, I am practically certain it was the circle. I could see only very
little of it at a time."

Subject, P. Radius, 8 cms. Diff.-ratio, 0"0076. Right eye. (Signal given after
17 seconds.) "At 8 there appeared a very indistinct darkening about twice the
usual breadth of the ring, in the left-hand quadrant towards the top. I was not
at all sure of it, and did not think I would make anything of it, but a part of it
at the top of the disc defined itself and I saw the inner edge, then a part to the
right of that."

Subject, A. Radius, 7 cms. Diff.-ratio, 0-0076. Both eyes. (Signal given after
6 seconds.) "The ring was at 7. It was seen first just above the centre. I was
able to trace it round in the upper part of the disc. It was not quite so dark as
it was the last time I used both eyes. I did not see the edges very easily or
very clearly, but it was quite permanent, and I could trace it round continuously.
I am quite sure it was the ring."

The records of any one series of observations, either binocular or
uniocular, show that as the difference between the intensities of the
background and the ring is gradually decreased, there are well-defined
changes (i) in the appearance of the ring, (ii) in the accuracy of locating
it, (iii) in the interval which elapses before it is first seen, and (iv) in
the degree of certainty with which the judgments are made. When
this difference is appreciably above the threshold of dLScrimination, the
whole of the ring is seen simultaneously with its edges clearly defined ;
it remains steadily in view; there is no difficulty in finding its exact
position ; and it is seen at once. When the difference is less, the parts
of the ring that are not in the centre of the field of vision often vanish
and sometimes even a part that is in the focus of vision disappears ; the
edges are not well-defined ; the reaction-times are slightly longer ; and.

Shepherd Dawson 89

though the location is generally correct, the judgments are given with
greater hesitation. With still smaller intensity-dififerences steady fixa-
tion is necessary before the ring is seen at all, and then only a small
part is seen at a time ; it is generally seen first on the left or at the top
of the disc; and to find its position it is necessary to follow it round
gradually to the scale, fixating parts of the disc at the same distance
from the centre as the part seen : often only the approximate position
is given ; the reaction-time is comparatively long ; and the subject is
generally more certain of having seen the ring than of having located
it correctly. Finally, with the smallest differences, the ring, if seen at
all, is seen only in ' glimpses,' i.e. it appears for an instant and then
disappears. At this stage there is very great difficulty in detecting
whether what is seen is subjective or objective. As it is almost
impossible to follow the ring round to the scale in the manner described
above, its position has to be found by comparing its distance from the
outer edge of the disc or from the outer edge of the ring formed by the
rotating counterpoise with the distance of the teeth from those parts of
the disc ; the location is, therefore, generally only approximate. The reac-
tion-times are very long, and the judgments are given with the greatest
hesitation. It sometimes happens that a subject will say he has not seen
the ring, and will then go on to say that he thought he had a glimpse of
it, the correctness of the location showing that he probably did see it.

Correctness and Certainty of Location. In Table II is shown the
frequency with which rings of each intensity were correctly located by
each subject. Table III shows the number of times each ring was
located correctly and with certainty. From both tables are excluded
those observations in which rings were seen at several places : this is
necessary because little importance can be attached to observations in
which rings are located at two or three different distances fi:om the
centre of the disc.

These tables show that with binocular observation the ring was
correctly located more frequently, and the subjects were more frequently
certain that their localisations were correct. They show, too, that the
lowest intensity-differences which evoke correct locations are very
nearly the same in the binocular and the uniocular series, and that the
range of intensities over which the percentage of correct locations falls
from 100 to 0 is smaller in the former series than in the latter. On
comparing Tables II and III it will be seen that the number of ' correct
and certain ' binocular locations is greater than the number of ' correct '
uniocular locations.

90

Discrimiriation of Brightness

TABLE II. Table showing the namher of times each ring was
located correctly and in 07ie position only.

Intensity

P.

A. R. ' C.

1 i

Both

Eight

Both

Left

Both

Eight

Both

Eight

00191
0-0171
0-0152
0-0133
0-0114
0-0095
0-0076
0-0057
0-0038
0-0019

22
22
22
19
21
17
10
3

22
22
22
19
11
7
4

11
10
11
11
10
7
3

11
10
8
9
7
2
3

7
7
7
6
6
4
3

1

6
6
4
5
3
1

10

10

10

10

10

9

4

3

2

10

8
9
8
7
1
2
1

No. of series

22

11

7

10

TABLE III.

Table showing the number of times each ring was
located correctly and with certainty.

Intensity

P.

A.

E.

C.

Both

Eight

Both

Left

Both

Eight

Both

Bight

00191

0 0171
00152
00133
00114
0-0095
0-0076
0-0057
0-0038
0-0019

22
22
22
19
21
12
2

22

22

21

18

8

5

2

11
10
11
10

8
5

1

10
9
7
4
1

2

7
7
6
5
1
3

4
2
1
2

10

10

10

10

9

6

2

10

8
9

7

1

No. of series 22

11

7

10

It is worth noting that, with few exceptions, the subjectively
certain locations are objectively correct. Of the 834 observations made
in these experiments there were only six in which the location was
wrong, while the subject was certain that he had seen it and located it
correctly : this is 48 7o of the total number of observations in which
the ring was located incorrectly — a very small proportion.

Shepherd Dawson

91

The amount of the difference between binocular and uniocular dis-
crimination can be measured only very roughly, and will vary with the
method of measurement. A rough measure is given by the difference
between the intensities at which 50% of correct locations (or correct
and certain locations) were made. These intensities are : —

Subject

Correct

Correct and certain

Binocular

Uniocular

Binocular

Uniocular

P.

A.
R.
C.

0-0060
0-0050
0-0086
0-0080

0-0095
0-0070
0-0119
0-0108

0-0074
0-0060
0-0126
0-0090

0-0101

0-0105
0-0186
00127

The difference-ratio, then, which evokes 50 "/^ of correct locations is
about half as much again when only one eye is used. At frequency
75"/„ correct the diflference is greater still.

It ought to be noticed here that inaccuracy of location is not always
a proof that no appreciable effect in consciousness has been produced
by the projecting strip on the rotating sector. The task of locating the
ring is more complicated than one would expect. When the ring is
clear and well-marked there is no mistake except occasionally in the
first few series, when a subject may forget the numbers of the teeth
on the scale or when his attention is so much concentrated on examining
the ring that he either forgets to locate it or, having done so, forgets
afterwards where he has located it. But when it is less clear, errors are
frequently made in cases where there is reason to believe that the ring
has been detected. It is frequently remarked that when the ring is
seen only in detached pieces the position of a piece on the left is
different from that of one on the right. A small piece of the curve
seems farther out than a larger piece. Probably this is a form of
the well-known illusion that when two arcs of the same circle are
of different sizes, the smaller seems to be part of a larger circle. This
would account for a good many of the errors of localisation. Such
introspections as the following support this explanation : —

" I was surprised when it did fall on the sixth notch : I had thought it was
a little farther out." (It was at the sixth.)

"At first I thought it was at the eighth notch, but decided it was at the seventh.
I saw it first at the top of the disc." (It was at the seventh.)

" At the eighth. It appeared on the extreme left of the disc. I find I have not
been allowing enough curvature." (It was at the eighth.)

92 Discrimination of Brightness

On the other hand, when the part seen is above the scale of notches
the location is usually accurate, and the subject is certain he has located
it accurately.

When the ring could be detected only after careful and continued
fixation of various parts of the disc, the difficulty of finding its position
was very great, for any attempt to do so sometimes only led to its total
disappearance.

When only one glimpse was obtained of a small part not near the
scale, its position had to be estimated after it had disappeared by its
distance from the centre or from the outer edge of the disc, so that the
process of locating the ring near the scale was different from that
of locating it far from the scale.

Sometimes the position of the ring was found by noting its position
with regard to the inner and outer bright rings. Sometimes it was
known by the tendency to confuse it with the ring formed by the pro-
jecting screw; thus one subject said: "I think it was at 7, because
it was near the centre but did not tend to get confused with the
circle round the counterpoise as a ring at the sixth notch does." (It
was at 7.)

The localisation is only approximate when the ring is first seen late
in the experiment, or when it is not well defined, e.g. :

" As I gave the signal, the grey ring appeared about the top of the disc. I had
not time to refer it to the scale. It might be at 8 or 9." (It was at 9.)

" A greyness about twice the breadth of the ring seemed to stretch all the way
round. It was at the seventh or eighth notch ; I could not localise it definitely
because it was twice the breadth of the true ring. The inner edge would be about
7." (It was at 7.)

There can be little doubt, then, that the ring was sometimes seen
but not correctly located ; but it is improbable that in such cases the
location was more than one centimetre from the correct position. It
might seem advisable, therefore, if we wish to find how often the ring
was seen, to include with the correct locations those observations in
which the location was not more than one centimetre incorrect. Such
a procedure would, however, almost certainly lead to the inclusion
of subjective impressions, i.e. impressions not directly due to the
objective ring on the disc, and as the probability of hitting by chance
on the correct or nearly correct positions is very much greater than
that of hitting only on the correct position, it seems better, to base
our conclusions on those observations in which the ring was located
correctly.

Shepherd Dawson

93

Reaction-times. Another interesting difference between binocular
and uniocular discrimination is in reaction -times. The subjects were
instructed to give a signal as soon as they saw the ring, and the
interval between the beginning of the observation and the giving of
the signal was measured by means of a stop-watch. The averages
of the reaction-times with their mean variations are shown in Table IV.
In compiling this table only those observations have been included in
which the location was correct or within one centimetre of the correct
position.

TABLE IV. Reaction-times (in seconds).

P.

A.

Intensity

Both eyes

Bight eye

Both eyes

Left eye

No.

Av.

M.V.

No.

Av.

M.V.

No.

Av.

M.V.

No.

Av.

M.V.

0-0171
0-0152
00133
0-0114
00095
0-0076
00057
00038
0-0019

22
22
22
21
22
17
9

3-7
4-3
4-3
7-2
9-6
13-8
23-7

1-7
1-4
1-2
3-8
5-4
6-7
5-4

22
22
22
19
14
7
3

7-3
8-1
130
15-9
23-7
18-1
21-7

4-3
4-3
6-1
5-3
4-6
7-9
111

9
9
9
9
9
7
5
3

2-1

2-9

42

8-7

10-9

14-6

21-6

10-3

0-9
1-4
1-3
51
6-5
5-9
7-3
11-1

9
9
7
9
8
2
3

40
8-6
14-8
17-3
21-4
12-5
12-8

1-3
3-6
6-3
8-3
61
2-5
7-6

R.

C

•

Intensity

Both eyes

Right eye

]

Joth eyes

Right eye

No.

Av.

M.V.

No.

Av.

M.V.

No.

Av.

M.V.

No.

Av.

M.V.

0-0191

5

2-6

0-5

5

100

4-4

10

5-0

2-7

8

10-8

5-9

00171

5

31

1-5

4

7-7

2-4

9

5-2

1-6

9

9-5

3-9

0-0152

5

5-4

1-1

3

20-6

5-8

9

4-7

2-4

9

12-6

5-5

0-0133

5

5-8

3-4

4

13-7

2-8

10

8-8

4-2

8

12-5

6-8

0-0114

5

13-5

40

2

20-5

1-0

10

13-7

6-6

3

17-6

9-9

00095

4

9-7

2-1

1

27-5

—

6

12-3

4-2

1

30-0

—

0-0076

3

23-6

1-8

—

—

—

5

21-0

4-4

1

220

—

00057

—

—

—

—

—

—

—

—

—

—

—

—

00038

—

—

—

—

—

—

—

—

—

—

—

It will be seen that as the difiference between the intensities of the
ring and of the background decreases, the reaction-times increase, and

94 Discrimination of Brightness

that the mean variation (which measures the variability of the indi-
vidual reaction-times about their average) increases too. There is
also a well-marked difference in the average reaction-times of the
binocular and the uniocular series. If we exclude the lowest in-
tensity-ratios, the average reaction-time in the uniocular series is about
twice that in the other ; if we exclude these, it is about half as much
longer.

The instruction to give a signal as soon as the ring was seen seemed
at first to be quite clear and definite, but later proved to be not
sufficiently explicit when the ring was so faint that it was seen only in
glimpses, for then it might mean " React as soon as you see any indica-
tion of the ring," or, " React as soon as you are sure you have seen it."
No attempt was made to remove this ambiguity, for it was not discovered
until the experiments were well advanced. This may account for some
of the variability in the reaction-times, for it is quite apparent from the
introspections that the instruction was interpreted sometimes in one
way, sometimes in the other, even by the same subject.

IV. Explanation of Results.

Apart from stereoscopic differences binocular and uniocular dis-
crimination differs in several respects : small differences of intensity are
detected more frequently when both eyes are used, judgments are
given with more certainty and consistency, and their reaction-times
are shorter. How are these differences to be explained ?

The explanations which suggest themselves most readily are lack
of practice in uniocular discrimination and the possibility of summation
of brightnesses of the uniocular images. Neither of these, however, is
satisfactory.

Practice. Practice, no doubt, accounts for the apparent improve-
ment in both kinds of discrimination after the first series, but it does
not account for the differences enumerated above, for if it did, since
nonnal vision gives less practice in uniocular discrimination than in
binocular, we should expect more rapid improvement in the former
than in the latter, and there is no evidence of such improvement. In
Table V is shown the total number of correct locations made by all four
subjects during the first and second halves of the observations. In
making this table the first series in A.'s and R.'s set have been omitted
in order to make the number of series even. It will be seen that when
all the observations are put together there is in each case a slight

Shepherd Dawson

95

increase in the number of correct locations in the second half of the
experiments, the increase in the binocular observations is about 19 7o
while in the uniocular it is only 7 7o- This difference is hardly large
enough to be significant, but, such as it is, it is just the opposite of
what one would expect if the suggested explanation were correct.

TABLE V.

Both

eyes

One

eye

Ist half

2nd half

1st half

2nd half

0-0152

24

24

22

23

00133

23

23

23

22

0-0114

20

24

18

18

0-0095

21

23

10

12

0-0076

13

21

7

9

0-0057

9

11

3

4

0-0038

2

7

1

2

Total

112

133

84

90

Increase

19%

7°/o

Summation of Brightnesses. The other explanation, that the ring
seen by both eyes is relatively darker than that seen by one because
of a summation of the brightnesses of the uniocular images in the
binocular, is even less satisfactory.

There is great difference of opinion as to whether an illuminated
surface appears brighter in binocular vision than in uniocular.
Valerius S Aubert", McDougall'' and others say it does. Valerius^
says the amount of the difference is -^ of the brightness of the
uniocular image. Aubert* puts it at from zero to ^, and says there
is no perceptible difference "with brightnesses greater than that of
white paper in diffuse daylight indoors." McDougall'' thinks that
"dimly illuminated surfaces appear brighter in binocular than in
uniocular vision, but that with brightly illuminated surfaces no such
difference is perceptible." Sherrington" could detect no difference

1 Ann. d. Physik u. Chemie, 1873, cl. 323.

2 Physiologic der Netzhaut, 1865, 286.

■' Brain, 1910-11, xxxiii. 372. * Physiologische Optik, 1865, 500.

5 Loc. cit. « This Journal, 1904-5, i. 50.

96 Discrimination of Briglitness

within the limits of brightness with which he experimented. Even
if there be any such difference, it cannot be invoked to explain the
diflFerences between binocular and uniocular discrimination, for we must
suppose that all parts of the illuminated binocular field (ring and back-
ground) will be brighter than similar parts of the uniocular field by the
same proportion, and consequently the ratio of the intensity of the ring
to that of the background will be the same in both cases.

If the ratio of the difference of the brightnesses of the binocular
and uniocular images to the brightness of one of them be not the same
for all intensities of the stimulus, it will still be impossible to explain
our phenomena by this means, for the brightness of the ring is so
very nearly the same as that of its background that any such law
of differential increment will be practically inoperative. What little
evidence there is tells against this explanation. Aubert, Piper^ and
McDougall say that the increase of brightness of the binocular image
over the uniocular is greater with dim illumination than with bright.
This, if it has any differential effect at all, ought to make the binocular
threshold of discrimination higher than the uniocular and not lower, as
it really is, for binocular vision should increase the brightness of the
ring more than that of the relatively brighter background.

Finally, the strongest objection to this explanation is that direct
comparison shows no really satisfactory evidence of any difference of
brightness-differences, and introspection is after all the final test. Such
comparison was seldom invited but was frequently made. It shows
that when the brightness-differences are well above the threshold of
discrimination there is no apparent difference between the impression
received by one eye and that received by two. It is difficult to see
how this fact can be reconciled with the suggested explanation. When
the brightness-differences are near the threshold, as they are in these
experiments, there is a difference between the uniocular and the
binocular images, but it is not a difference of brightness. My subjects
after some practice were able to distinguish the grades of intensities
with very great accuracy, so that although they sometimes said that
a brightness-difference was equal to one which was really slightly less
or slightly greater, they rarely said that it was greater than one which
was really less than it (only twice, in fact). One would expect, there-
fore, that if any difference in the brightness-differences of the uniocular
and binocular images existed, they would have been noticed. Yet only
once did subject P. remark that a ring seen with both eyes was darker

^ Ztsch.f. Psychol, v. Physiol, d. Sinvesorg., 1903, xxxii. 176.

I

Shepherd Dawson 97

than a similar one seen with one eye, and subjects R. and C. not at all.
Subject A. noted a difference of brightness ten times, but it is possible
that he may have confused steadiness with darkness of grey. It may
be worth noting that it was he who on two occasions said that a grey
was darker than one which was really brighter. With such small
intensity-differences as were used here, the comparison of intensities
is very difficult, and very little importance can be attached to the few
cases in which a difference of brightnesses was recorded. Any un-
certainty, whatever be its cause (length of time allowed for observing,
intermittence of the impression, or lapse of attention), may be attri-
buted to a decrease in the brightness-differences of the impression.
Indeed it is a tribute to the care with which the observations were
made that such confusion did not take place more frequently. Con-
sidering the smallness of the difference-ratios used and the great
possibilities of error, there is, then, little reason for supposing that in
the binocular image there is a summation of the brightness-differences
of the uniocular ones.

Introspective basis for an explanation. In order to arrive at a
satisfactory explanation it will be necessary to examine the detailed
introspections of the subjects. It has already been remarked that
as the difference between the intensities of the ring and the back-
ground decreases there are clearly marked changes in the appearance
of the former which affect the accuracy and quickness of location, the
inverval that elapses before its first appearance, and the degree of
certainty with which judgments are made. When the difference is
above the threshold of discrimination the ring is seen continuously
as a whole and there is no difficulty in locating it ; when it is smaller
the ring is seen in fragments which disappear and reappear, and on
this account location is somewhat difficult; finally, when the smallest
differences are used very steady fixation is necessary, and even then the
ring, if seen at all, is seen in glimpses. At this last stage only very
small portions of the ring are seen at a time; sometimes they are fairly
clear, sometimes they take the form of an undefined blur often twice
as broad as they should be, sometimes only one edge is seen and
sometimes the edges are seen alternately. There is consequently great
difficulty in deciding whether what is seen is objective or subjective.
(I shall restrict the term ' objective ' to impressions of rings which are
formed by rotation of the projecting sta-ip on the sector.) These faint
objective impressions appear to differ very little from others which are
certainly subjective, as the absence of any objective cause for them

J. of Psych. VI 7

98 Discrimination of Brightness

in the rotating disc shows. Here are a few examples of purely sub-
jective impressions : —

Radius 9. Diff.-ratio, 0*(X)38. Signal given after 30 seconds. "At the sixth
notch. Just as you said * Stop ' it appeared quite clearly and suddenly at the top
of the disc. Its edges were well defined. I saw it only for a very short time, and
I saw only a very small part of it. It apjieared exactly like the ring ; had well
defined edges. It seemed to apjiear without any effort of attention. It was the
proper breadth, and looked just like a part of the ring when it is very clear. First
when I looked there was a greyness at the ninth notch, but it never a.«sumed the
form of a ring."

Radius 6. Diff.-ratio, 0-0095. "There was a greyness about the eighth notch
which looked like the ring and became clearer towards the end of the observation."

A blank disc was shown. " When I gave the signal (after 3 seconds) I saw the
ring as a whole. It was rather indistinct. It seemed to be near the eighth notch.
I saw it constantly throughout, but I was not always certain of the location. When
I looked near the scale it seemed to be at the eighth notch, but when I had only the
left-hand side in view it seemed to be closer in. I am quite sure that it was at the
eighth. The edges were never well defined. I did not see the ring at any other
part of the disc."

Some of these impressions — those of rings of radius 6 and 10 cms. —
may be due to contrast : the outer and inner edges of the disc were
brighter than the rest of the field because of contrast with the dark
background and the dark centre formed by rotating the counterpoise ;
the parts of the disc just inside the large bright ring and outside the
small one may have been slightly darker than the rest of the field
because of contrast with these bright rings. This would explain the
fact that rings were inaccurately located at 10 and 6 slightly more
frequently than at any other position. Yet it does not explain all
these impressions, for they were located almost as frequently at 7, 8
and 9. After-images of the outer edge of the disc may have produced
some of them, but other causes must have been at work too.

What the causes of these subjective phenomena may be does not
concern us now, but there can be no doubt of their existence ^ They
are detected in the observations of all the subjects, although they are
more common with some than with others. They are not the product
of continued observation in this class of experiment, for they are
present from the very first; nor are they due to carelessness. They
are a source of considerable disturbance in the detection of very small
differences of stimuli, so much so that a large part of the observation

1 Cf. Oswald Kiilpe, " Ueber die Objectivirung und Subjectivirung von Siunesein-
driicken," Philos. Stud., 1902, xix. 508,

I

Shepherd Dawson 99

seems to be given up to deciding whether what is seen is objective or
not. The following records show this very clearly :

Radius 8. Diflf. -ratio, O^OOSS. Signal given after 27 seconds. "At the sixth
notch there appeared a ring half as broad as it should be, with its inner edge well
defined. It appeared above the centre of the disc. I did not take it to the
scale but I could see that it was at the sixth. It was an arc of about 60° with
no outer edge : it just shaded off into the background. There were two other rings,
one at the eighth and one at the tenth. I should place the eighth ring next the
sixth one, but it was only a greyness ; one part of it about an inch long appeared
there for an instant, but it disappeared so suddenly that I thought it was not the
ring. The inner ring was much too narrow, and the outer one much too broad ; it
was twice the breadth of the proper ring; the edges of the outer ring faded off.
I cannot say which was the ring. The middle one appeared first, then the outer
one, then the inner one."

Radius 6. Diff. -ratio, 0'0095. Signal given after 25 seconds. "A very in-
distinct ring at the sixth notch. It was broader than usual by about one-half.
There were also other greynesses at about the eighth or tenth, but their edges were
not so well defined."

The reports of my experiments give some indication of the
standards by which the subjects decide whether their impressions are
objective. Several criteria are used, sometimes singly, sometimes
together. The chief are permanence of the impression, ability to
detect various parts of the ring, its curvature and breadth, clearness
of outline, relation to fixation and attention, and absence of similar
impressions elsewhere. A part of the ring that appears only for an
instant or that vanishes while it is being fixated and does not return
is usually classed as subjective. On the other hand, if it can be seen
at other parts of the disc, or, if it stands out steadily for some time, it
is supposed to be objective. A large part is more likely to be considered
objective than a small one. The following extracts from the reports of
observations in which the ring was located correctly will show the
importance of these considerations :

" It appeared first on the left, then it disappeared and I thought it was not the
ring, but it reappeared and I became sure."

" After giving the signal I thought I had made a mistake, but it returned."

" I did not see the ring. Once, about 8 cms. from the centre and above it,
a small part of what might have been a ring appeared, but it was just for an
instant." (There was a ring at 6.)

" A greyness about 7 cms. above the centre, but I should not care to say it was
the ring. It seemed a little broader than it should be. It appeared about the
middle of the observation. It never became defined. Another time I saw a grey-
ness about 11 or 12, but it was not so persistent as this." (It was at 7.)

7—2

100 Discrimination of Brightness

" I cjould not tell at first whether it was the ring, but when I saw it at three
parts of the disc all the same distance from the centre, I became sure it was."

" I was not at first sure it was the ring and did not give the signal till I had
traced it round from left to right."

"I had glimpses of what may have been rings at 6 and 11, but that at 6 was
more permanent and I seemed to see more of it. I think it was at 6." (It was.)

When the edges are blurred and hazy there is often considerable
doubt about its objectivity ; if, however, they are or become clear, the
subject is fairly sure that he has seen the ring. Sometimes rings were
seen that were broader or narrower than usual, or not quite circular ;
in such cases there was always some doubt as to their objectivity.
When after searching the whole of the disc, only one ring could be
found, the absence of other impressions would lead the subject to
decide in favour of this one. On the other hand, after concentrating
attention on an undefined blur to the neglect of the rest of the field,
he was often doubtful about the objectivity of this impression both
because of its indefiniteness and because of not looking elsewhere. A
few more extracts will make this clear :

" I saw a greyness about twice the breadth the ring generally is, whose inner
edge would be about 6 cms. from the centre. I should not like to say it was the
ring, because I did not give the rest of the disc much attention."

" I saw two rings of greyness, one at 7, the other at 9. I think the ring itself
was at 7, because its inner edge was more indistinct than that of the other. It was
clearer and not so broad."

" There was a very indistinct ring at 6, but it was broader than it should have
been by about one-half. There were also other greynesses about 8 and 10 but their
edges were not so well defined." (It was at 6.)

When there is any doubt about the objectivity of an impression, it
is because some or all of these marks of objectivity are missing. This
conclusion is corroborated by an examination of those cases in which
the ring was located so far from its true position that the impression
could not have been due to any differences in the intensity of the light;
and at the same time it partially explains those illusions, for in these
cases there is an illusion of objectivity just because the impression
shows some of the above-named marks of objectivity. A few examples
from the introspections will make this clear :

Radius, 10 cms. Signal given after 22 seconds. " I saw what I thought was
the inside edge of the ring on the left of the disc. It persisted, then disappeared
and I saw nothing else. It was so well defined that I am almost certain it was
there. There was a faint suggestion of the rest of the ring. It would be about 8.
Yes, I am sure of the location."

Shepherd Dawson 101

Radius, 9 cms. Signal given after 30 seconds. "At 6. Just as you called
' Stop ' it appeared quite clearly and suddenly with its edges well defined. It would
be closer in than 6 ; 5, I think. I saw only a very small piece of it — about a quarter
of an inch of it. It seemed to appear without any effort of attention. It was the
proper breadth and looked just like a part of the ring when it is very clear. I had
explored along the vertical line above the centre. First when I looked there was
a greyness at 9, but it never assumed the form of a ring."

There seems to be considerable difference between subjects in their
attitude towards these criteria. Some appear to be satisfied with only
one of these marks of objectivity, while others are not satisfied unless
they detect several. Subject R., for example, sometimes looked for
several, and because he could not find them he was doubtful of the
objectivity of his impressions.

This difference of attitude may account for a characteristic difference
between, subjects. Some are frequently doubtful of the objectivity of
impressions the correct location of which shows that they have been
due to differences of intensity in the light, and when the intensity-
ratio falls below a certain amount, they almost invariably state that
they have not seen the ring. Others not only do not subjectify their
impressions, but even obtain glimpses of rings the positions of which
are so far from that of the ring formed by the rotation of the strip
on the sector that they are obviously not objective : they even continue
to see rings or parts of them when the strip is entirely behind the
sector so that a uniformly bright disc is shown. These subjective
impressions are sometimes described in as great detail as objective
ones, and frequently they are believed to be objective. Subjects of
this type seldom say they have seen no ring. Subject A. is a good
example of this class, subject R. of the other ; P. and C. are more
mixed in type.

Suggested Explanation. The brief account that has just been
given of the difficulties experienced in making these observations will
suflSce to show that the process of discrimination is fairly complex,
involving in some cases a very careful examination of the shape, size,
distinctness, brightness and steadiness of our impressions. It is to
one of the causes of this complexity — namely, the intermittence of
a presentation — that we are to look for the explanation we are
seeking.

Oscillation seems to be a feature of our mental life. Visual
presentations, in particular, are constantly fluctuating : the visual
appearance of even a well-defined object will change if the latter be
looked at very steadily ; and, when the brightness of the object

102 Discrimination of Brightness

differs very little from that of the rest of the field (as in the greys
used in these experiments) these fluctuations are very frequent and
well marked.

All my subjects agree that the most pronounced difference between
the uniocular and binocular images evoked by the same stimuli is in
steadiness. A just discernible ring as seen by one eye was more
intermittent than a similar one seen by both ; it was sometimes
described as having the elusive, fluctuating character of the clouds
of light that are seen when the eyes are closed and the eyeballs
pressed, or as being hazy, filmy and ghost-like ; it was seen and located
less easily, the edges were sometimes less distinct, and less of it
was seen at once. Frequently it was remarked that the uniocular
observations were more difficult to make than the binocular, and more
careful observation showed that this difficulty was due to the greater
fluctuation of the uniocular ring. Now it is evident from what has
been said about the criteria of objectivity that an increase in the
steadiness of an impression will produce an increase in the number
of recorded objective impressions, and in the number of ' correct
and certain ' locations ; hence if we can account for the above-men-
tioned difference in steadiness, we shall have the explanation we
are seeking.

This explanation is, I suggest, to be found in the way in which
the sensations evoked by the stimulation of each eye are integrated
in binocular vision and is as follows: the sensations due to the
stimulation of each retina are developed and fluctuate independently,
their periods of fluctuation are not synchronous, and therefore in
binocular vision there is an overlapping of the periods of appearance
of ' corresponding ' parts of the two images, and hence a relative
increase in the steadiness of the impression.

Let us consider in turn each of the premisses on which this
argument rests.

Sherrington and McDougall have called attention to some facts
of visual experience which are explicable only on the assumption
that the cerebro-retinal mechanisms of the eyes are independent.
Sherrington found that when corresponding points of the retinae are
stimulated by intermittent light and dark stimuli either synchronously
or asynchronously the rapidity of alternation of stimuli required to
extinguish binocular flicker is just the same as that required to ex-
tinguish uniocular flicker. He concludes that these experiments "show
that during binocular regard of an objective image each uniocular

Shepherd Bawsoi?

103

mechanism develops independently — at least as to steadiness of
brightness and intensity of brightness — a sensual image of consider-
able completeness. The singleness of the binocular perception results
from the combining of these elaborated uniocular sensations : it is
the product therefore of a psychical synthesis that works with already
elaborated sensations contemporaneously proceeding\" McDoiigall has
adduced other observations which are most readily explained in the
same way.

Whatever be the nature of the physiological processes which
accompany binocular experience, there certainly are features of that
experience which indicate that even in many so-called cases of
binocular fusion the uniocular images retain — at any rate as regards
steadiness and brightness — the characteristics they have when they
are present singly. An analogy may help to make my meaning
clear. The sensations of sight and sound are to some extent in-
dependent: they are evoked by the stimulation of different sense-
organs, they are independently variable and, although they may fuse
to a greater or less degree or may partially inhibit one another, yet
they are each capable of analysis from the rest of experience and
each on analysis has the same attributes as when the other is absent.
So it is with uniocular images : each appears to be evoked by the
excitation of different cerebro-retinal paths which do not terminate
in a common cerebral centre, and each is capable of analysis in
binocular perception ; although they fuse more intimately than do
modally different sensations, yet they do not lose their identity in the
resultant binocular experience : each preserves the attributes it has
when the other is absent. I do not mean that the reciprocal influence
of one image on the other is impossible ; the facts of binocular rivalry
disprove that. I mean merely that each image is relatively complete
in itself and that the mode of integration of these images is a problem
for the psychologist.

This theory appears to be the only one which can adequately
account for Sherrington's and McDougall's observations and for all the
observations described in the preceding pages. It receives further
support from observations made to establish the second premise —
that the right- and left-eye images fluctuate independently and
asynchronously.

After the completion of the experinaents already described, another
series of observations was made in which the periods of intermittence

' Loc. cit.

104 Discrimination of Brightness

of uniocular and binocular images were measured. In these experi-
ments the subjects were instructed to press a reaction-key when they
saw the ring and release it when it disappeared. The movements
of the key were recorded on a rotating drum. Each observation lasted
about 75 seconds, and each sitting began with one or more practice
series. To facilitate the task of the subject he was directed to confine
his attention to the part of the disc directly above the centre and to
record the fluctuations of the ring there. To simplify the work still
more, all the rings used in these experiments were of the same radius,
and this was made known to the subjects before the experiments
began. Such a procedure would be very unreliable with unpractised
subjects, as their judgments might be very much influenced by the
knowledge of the position of the ring, but with practised subjects the
danger is not so great.

These observations are not easy to make, for it is difficult some-
times to know whether anything is seen or not, and the reaction
occasionally distracts the attention from the disc. Some observers
thought they discovered a tendency to tap rhythmically which deter-
mined the times of appearance and disappearance, but this did not
often happen, and the introspections were not always corroborated by
the tracings : if such a tendency did exist, there is no reason for
believing that it seriously disturbed the experiments. There is another
difficulty arising from the degrees of fluctuation. Sometimes the ring
in the area under examination would disappear entirely, leaving a
uniformly white field : sometimes different parts would alternately
disappear and reappear. The subjects were instructed not to record
the latter kind of fluctuations. This instruction may not have been
carried out in every case, but after some practice all the subjects felt
that they could keep their attention fairly constant : if any partial
disappearances were recorded they were not numerous enough to affect
the results seriously. Other disturbing conditions were a tendency to
look for the ring on other parts of the disc, and the difficulty of
pressing and releasing the reaction-key at the right moment. Yet,
in spite of all these difficulties, the records are probably significant,
for all the subjects were well practised, every possible precaution
was taken, the observations were repeated many times, and the records
agree in general tendency, although the subjects themselves during
the experiments had no idea of the relative lengths of the periods
of intermission.

Table "VI gives a specimen record. The figures in the columns

Shepherd Dawson

105

show the intervals (in seconds) during which the ring was reported to
have been seen or not seen. Each series began with an interval during
which it was not seen and ends with one during which it was seen.
This method of dealing with the results seemed better than that of
allowing a fixed time for observation and stopping exactly at the end of
it, because it made easier a comparison of the intervals during which it
was seen with those during which it was not seen.

TABLE VI. Subject P.

Diff. -ratio

0 0133

0-0095

0 0057

Not seen

Seen

!
Not seen • Seen

Not seen

Seen

5-2

4-7

11-8 I 1-2

32-5

2 0

4-4

7-2

81 1 3-6

14-2

0-8

6-8

21

3-3 11

11-8

2-8

Bight eje

6-9 . 121

60 i 2-8

91
71

2-5

7 0

15-7

2-4

11-4

5-6

7-6

21

7-7

0-9

1-8
10-5

3-6
3 0

48-4

35-4

59-3

241

74-2

8 0

5-0

17 •»

5-3

8-4

7-9

11

4-4

31

5-8

51

3-6

2-4

- 3-4

■ 5-7

5-2

41

2-3

2-8

5-3

17-6

10-5

8-2

4-9

6-4

6-8

1-9

7-3

8-0

7-1

6 0

Both eyes

4-2

9-7

7-5

7-0

6-3

2-9

2-8

0-7

8-6
4 1

1-5
10 1

•'

32-3 56-5

41-6

40-8

44-8

33-2

This record, like all the others, shows marked irregularity in the
times of appearance and disappearance in both the uniocular and the
binocular observations : the ring may be seen for a fraction of a second,
disappear, and then reappear to remain for five or ten seconds more.
The only point of resemblance between successive series of observations
made with the same eye on the same ring is in the ratio of the total
time during which it is seen to the time during which it is not seen ;
this remains fairly constant. In other respects they differ : the interval
that elapses before the ring first makes its appearance and the time it
remains before disappearing again vary considerably.

106

Discrimination of Brightness

It cannot be demonstrated by direct observation that the right- and
left-eye images fluctuate independently ; we must therefore look for
indirect evidence. This is to be found in a comparison of the rates of
fluctuation of binocular and uniocular images. If uniocular imagoes
fluctuate independently one must suppose that sometimes the ring will
be seen by one eye, sometimes by the other, sometimes by both and
sometimes not at all. The time the ring is seen when both eyes are
used will, therefore, be relatively longer than when only one is used.
Observation shows this to be the case.

In Table VII is shown for each subject and each intensity the ratio
of the total time during which the ring was seen to the total time

TABLE VII. Table shoiving ratio of intervals during which ring
was seen to intervals during which it was not seen.

Intensity

P.

K.

A.

C.

Both

Bight

Both

Bight

Both

Left

Both

Bight

0 0133

0-0095
0-0057

2-786
1-535
0-898

1-598
0-624
0-275

3-408
2133
0-543

0-828
0-610
0-197

2-227
1-734
0-652

0-907
0-635
0-326

1-661
0-991
0-178

0-915
0124
0-032

during which it was not seen. This ratio is in every case greater with
binocular than with uniocular observation. It also decreases as the
intensity-ratio decreases.

Here it might be asked : even if the uniocular images be developed
independently and fluctuate asynchronously, what reason is there for
supposing that in binocular vision the overlapping of the times of
appearance of the ring in each image has the effect of increasing the
relative interval during which it is seen ? Is it not as reasonable to
suppose that rivalry takes place so that sometimes the blank field seen
by one eye inhibits the ring seen by the other ? This, certainly, is
logically possible, just as possible as the account we are giving; it does
not, however, square with the facts. In the first place the phenomena
described above are not conditioned in the same way as rivalry.
Binocular rivalry is produced by presenting to the right and left eyes
objects which differ very considerably in form or colour or both and
evoke sensations which are equally different. In our observations the
stimuli are alike, but the sensations sometimes differ slightly. In the
first case the difference of sensation is due to differences in the stimuli ;

J

Shepherd Dawson 107

in the second it is due to the oscillatory character of the psycho-physical
processes. It would be unjustifiable to suppose that, because rivalry
sometimes (not always) occurs between sensations evoked by very
different stimuli, it will therefore occur between slightly diflferent
sensations evoked by the same stimuli. Again, there is seldom rivalry
when one eye is stimulated by a blank field and the other by objects
with well-marked contours, as when a blank card and one on which a
letter is printed are seen in the stereoscope. It is only with very steady
fixation that the letter will disappear; and even then it is doubtful
whether the disappearance is due to rivalry or merely to the well-known
fact that steady fixation produces fluctuations in any visual presentation.
There is, therefore, no reason for supposing that the absence of a dis-
tinguishable ring in one retinal image inhibits its appearance in the
other.

If our account of binocular perception be correct, if visual images
fluctuate, if corresponding elements of the right- and left-eye images
do not always appear synchronously, and if there is a consequent over-
lapping of the periods of their appearance which has the effect of
increasing the interval during which each element in the binocular
image is present in consciousness, we can account for all the observed
facts — the frequency of correct locations, the certainty with which the
judgments were given, and the reaction -times.

More negative judgments, i.e., those to the effect that no ring was
seen, were given in the uniocular observations because the ring being
more intermittent there was a greater possibility of its being regarded
as subjective. More correct locations were made in the binocular
observations because, the impression being steadier, it was easier to
find its exact position, and for the same reason more ' certain locations '
were made in the binocular observations. The smallest differences of
intensity that can be detected are nearly the same with binocular vision
as with uniocular because the binocular image is an integration of the
uniocular images and does not exist in their absence.

The reaction-times are similarly explained. It has been shown that
these may measure either the interval that elapses before a ring is seen,
or the interval that elapses before the subject is certain that he has
seen the ring. The greater steadiness of the binocular image will,
therefore, in a large number of experiments have the effect of reducing
the reaction-time, for certainty comes more quickly with a steady
impression than with one which comes and goes. The increase in the

108 Discrimination of Brightness

mean variation of the reaction -times with increase of intensity-ratio is
probably due to the greater difficulty of keeping one's attitude constant
in reacting to impressions evoked by low ditference-ratios ; the difference
between the mean variations of the binocular and the uniocular
observations is to be explained in the same way.

{Mannscinpt received 21 Februai'y, 1913.)

I

THE QUANTITATIVE INVESTIGATION OF
HIGHER MENTAL PROCESSES.

By STANLEY WYATT.

{From the Fsycholoyiccd Laboratory^ University of Manchester.)

I. Objects of the Investigation.

II. Historical.

III. The Subjects and General Conditions of the Present Investigation.

IV. The Tests and their application to the Subjects.

V. Statistical Methods employed and Correlation Results.

VI. Conclusions.

I.

The primary aim of this investigation^ has been to ascertain to
what extent ditfereut tests correlate with a subjective estimate of
intelligence, and to select those tests which give the highest coefficients
of correlation. For this purpose, the most satisfactory methods of
previous investigators have been followed, refinements, however, upon
earlier procedures being introduced whenever possible.

Indeed, the methods of applying mental tests to school children
have left much to be desired. Too often the children have been
subjected to distractions in the form of strange experimenters and
unusual conditions of work. This division of the child's interests must
unquestionably have affected the results obtained. Further, the
actual application of the tests has often been placed in the hands of
laboratory assistants who did not know the children, or in those of class
teachers, persons usually untrained in psychological methods.

In the part of this research which was conducted at the Fielden
Demonstration Schools, Manchester, the writer was the only supervisor

1 It was cond acted under the guidance of Mr T. H. Pear, to whom the writer is
especially indebted for the advice and assistance so freely given.

110 Quantitative Investigation of Mental Processes

present, and no one was allowed to enter or leave the room while the
tests were in progress. The aim throughout was to secure the most
favourable experimental conditions possible, so that the results would
afford a reliable estimate of the mental traits under investigation.
Consequently the conditions under which the tests were performed
differ very considerably from those of most previous investigators.

Much uncertainty still exists regarding the hierarchical arrange-
ment of the coefficients of correlation of each test with every other.
At the present time opinion is about equally divided and it is hoped
that the results of this investigation will assist in defining the correct
view.

Recently the question of a ' general factor ' underlying mental pro-
cesses has been the subject of much discussion. Many investigators
deny its existence altogether and those holding the opposite view are
not agreed upon its constitution. It was hoped that evidence would be
adduced here which would help to solve this problem.

Many of the tests in this research involve what may be termed
' general ability ' rather than knowledge which has been acquired in
school. Some of them bear no resemblance whatever to school work,
and hence they afford a means of testing the real (in opposition to the
apparent) intellectual capacities of the child.

II-

An excellent summary of the work done in this field is to be found
in a book by William Brown ^ Since the publication of this book,
Burt^ has issued the results of a research on " Experimental Tests of
Higher Mental Processes and their relation to General Intelligence."
He concludes that " those tests involving higher mental processes such
as Reasoning, vary most closely with Intelligence, and are least vitiated
by variations with irrelevant conditions, such as Sex, Social Status,
Training of the experimenter, and mass-measurement of numbers of
children at once^"

More recently there has appeared a paper on " General Ability, its
Existence and Nature ^" The authors utilise all available data for the
establishment of their conclusions ; data collected both by supporters

^ The Essentials of Mental Measurement. Cambridge, 1911, 81 — 97.

2 J. ofexper. Fed., 1911, i. No. 2, 93—112.

3 Op. cit. p. 112.

* Bernard Hart and C. Spearman: "General Ability, its Existence and Nature."
This Journal, 1912, v. 51—84.

Stanley Wyatt 111

of their theory and by those antagonistic to it. In all these cases they
consider that the 'general factor theory' is supported. They conclude
that " the fact of correlation existing between quite different intel-
lectual performances seems to be fundamentally identical with the fact
that any such performance inhibits quite different simultaneous ones.
Both phenomena are explicable by conceiving that every performance
depends partly on some common fund of energy. This, then, is the
required General Factor'." Further, "every performance depends, not
only upon this General Factor, but also in varying degree upon a factor
specific to itself and all very similar performances-."

Simpson^ on the other hand, maintains that " there is no justifica-
tion for the view that ' general intelligence ' is to be explained on the
basis of a hierarchy of mental functions," which are correlated owing to
their common connexion with a central factor.

III.

The subjects who were examined during the present investigation
comprised :

(a) A group of 34 children of both sexes (Group I) attending the
Fielden Demonstration Schools, Manchester^ Their ages ranged from
11 to 13 years. In social status they were superior to the average
elementary school-child. All had attended the school for several
years ; hence they had long been subjected to the same environmental
influences so far as school-life was concerned. The tests were ad-
ministered during the school hours of thb mornings of Tuesdays,
Thursdays, and Saturdays, the time occupied each morning by the
tests being 40 minutes. The entire series of tests was given by the
writer. Any distraction which, under ordinary circumstances, might
have been caused by the presence of a stranger, was eliminated, for
during the past two years the writer had frequently acted in the
capacity of teacher to the same children. The school presented peculiar
facilities for an experimental investigation of this kind. The children
were thoroughly accustomed to the introduction of new methods of

1 Op. cit. 2 Ibid.

^ Correlations of Mental Abilities. New York, 1912.

* The writer here desires to express his grateful thanks to Mr W. J. Deeley and
Mr A. S. Harrison, successive Senior Masters of the Fielden Schools, for the generous
manner in which they made possible the application of the tests ; also to Br P. Sandiford,
the Superintendent of the Fielden School, for his permission to carry out the tests in the
Schools.

112 Quantitative Investigation of Mental Processes

teaching and hence mental tests were not entirely foreign to their daily
experience,

(6) A group of 41 children (Group II) at the Manchester High
School for Girls^ Their ages ranged from 10 to 12 years. Owing to
the inconvenience caused by dislocation of the curriculum, the number
of tests had here to be curtailed. The tests employed were carried out
during school hours, and the children were led to believe that they
formed an integral part of the school curriculum. The tests at this
school were administered by the Headmistress of this section of the
school, the method of procedure being similar to that adopted with the
previous group.

In all these tests, the eagerness of the children was remarkable.
Often, at the end of each performance, the cry for more was raised.
The children entered into the work with far greater zeal and interest
than was shown in the case of ordinary subjects of the curriculum.
During the whole of the session this spirit was maintained ; and when
the tests were concluded, a general feeling of profound regret was
expressed.

At the Fielden Demonstration Schools, a special room was set aside
for the purposes of this investigation. It was situated in a quiet and
undisturbed part of the building; in fact it was practically isolated
from the rest of the school. The children were arranged around three
sides of a rectangle, the experimenter being at the fourth. Each in-
dividual was thus under his direct supervision, and he was in full
view of the class. Such an arrangement was found to be the most
economical, especially with regard to the distribution and collection of
papers. The experimenter could move rapidly round the inner side of
the rectangle with a minimum of effort and a maximum of effect.

For demonstration purposes, a revolving blackboard was used. It
could be distinctly seen by each child seated normally in his position
in class, and was well suited to class experiments when a definite time
of exposure was required. The stimulus word could be written on one
side of the blackboard, and, by a quick rotation of the board, could be
almost instantaneously exposed to the class.

During the tests the experimenter noted any peculiarities of the
children as shown by their outward expressions. Throughout the
whole session the proceedings were most automatic ; the children

1 To Miss Sarah A. Burstall, the Headmistress, for her kindness in allowing the nse
of the School, the writer tenders his best thanks ; he is also greatly indebted to Miss
Harrison of the Junior School, for her invaluable assistance in the api^cation of the tests.

Stanley Wyatt 113

always awaited eagerly the signal to start; perfect stillness prevailed
whilst the tests were in progress, and there was an immediate response
to the command ' Stop,' Thus the conditions appeared to be as perfect
as it was possible to make them, and the chance of obtaining accurate
and reliable results was thereby increased.

At the Fielden Demonstration Schools, a classification of the
children according to the order of their ' intelligence ' was obtained
from the master in charge of the class. This classification was based,
not upon the results of any class examinations, but upon the master's
opinion of the intellectual capacities of each of the children. Such a
method of classification may be expected to result in coefficients of
correlation which have a higher numerical value than those obtained
from the High School for Girls, where the results of class examinations
were used as data from which the correlations with intelligence were
calculated.

IV.

In the case of every test, prolonged preliminary trials were made,
and as a result some tests were discarded as impracticable whilst others
were omitted as irrelevant to the aims of the investigation. Those
were retained which showed indications of being representative of the
higher intellectual capacities; a few memory tests were also included \
Eventually the following tests were adopted, and were applied in the
order given (for the sake of brevity the usually accepted names of these
tests are given here, details are given later) :

Rearranged Letters.

Memory (immediate reproduction) for nonsense

syllables.
Memory (delayed reproduction) for nonsense

syllables.
Memory (immediate reproduction) for letter

squares.
Cross-Line Test.
Interpretation of Fables.

In the case of Group I the tests were repeated after an interval of
six weeks. Only the first eight tests were applied to Group II.

1. ' E, R' Test. A number of letters of the alphabet (600) were
printed on papers, each paper containing twenty-one each of the letters
E and R, and twenty-one each of the letters A, N, O, S for the next

1 The tests carried out at each sitting never exceeded three in number, and in the case
of the longer tests only twQ were given.

J. of Psych. VI 8

1.

' E, R ' test.

10.

2.

' A, N, 0, S ' test

11.

3.

Word-building.

4.

Sentence Construction.

12.

5.

Analogies.

6.

Completion.

13.

7.

Missing Digits.

8.

Part-wholes.

14.

9.

Dissected Pictures.

15.

114 Quantitative Investigation of Mental Processes

test. The remaining letters were drawn as equally as possible from the
rest of the alphabet. An example (using different letters) was first
given on the blackboard, and then the papers were distributed face
downwards. The children were told that they must cross through
every E and R, beginning at the top of the page and proceeding down-
wards line by line. They were advised to work as quickly as possible
but at the same time to avoid passing over any of the presented letters.
At a given signal they turned the papers over and began to cross
through the letters. At a second signal, two minutes later, they turned
the papers face downwards again, and these were then collected.

System of marking and results: 1 mark for every letter crossed
through correctly; — 1 mark for every letter omitted or crassed through
incorrectly.

Group I Group II

Mean 17-75 12-3

Standard Deviation {a) 6-4 12-8

Reliability Coefficient -72 —

Correlation with Intelligence (r) -40 '37

P. E. of r -097 -098

2. ' A,N,0,S' Test. The material and method of procedure were
similar to the last test, but here the children were required to cross
through the letters A, N, O, S.

The time allowed was three minutes ; different children having
been previously tested both in this and the ' E, R, ' test, in order to
ascertain the time required to complete the test. The three minutes
and two minutes were found to be just insufficient for the completion
of the respective test. The letters A, N, 0, S were written on the
blackboard to act as a reminder in cases of forgetfulness.

System of marking and results: the system of marking was that
adopted in the previous test.

Group I

Group II

Mean

41-2

37-0

Standard Deviation (a)

12-1

16^4

Reliability coefficient

•64

—

Correlation with Intelligence (r)

•45

•32

P. E. of r

•09

•099

3. Word-building Test. The children were provided with papers
containing the letters A, E, O, B, M, T typed on the top, and they were
required to construct, from the letters given, as many words as they
could. The words were to contain any number of the letters from two
to six (inclusive) but letters other than those given were not to be used.

Stanley Wyatt

115

Further, no letter could be used more than once in the same word, and
the words allowed as correct must be found in a standard dictionary,
A preliminary test was given with the letters E, A, I, R, L, P in order
to acquaint the children with the nature of it, and to encourage
questioning on doubtful points. The test papers were given out face
downwards, and at a given signal the children turned them over and
commenced to work. The time allowed was five minutes, at the end
of which period a second signal was given and the papers were again
turned over and collected.

System of marking and results : 1 mark was given for each word
correct.

Mean

Standard Deviation (o-)
Eeliability Coefficient
Correlation with Intelligence (r)
P. E. of r

Group I

10-5

4-08

•88

•58

•07

Group II
12-5
4-42

•50

•08

For the second application of this test the letters were rearranged
thus: 0,A,E,M,T,B.

4. Sentence Construction^. A list of ten words, expressing either
concrete or abstract ideas, was given. The children were asked to
construct sentences each containing a pair of successive words. They
were told that the sentences must be constructed so as to show the
closest possible connexion between the words used. Only the exact
word given might be used ; if singular, then the use of the plural was
forbidden. A trial series of ten words was first given, so as to make
the children familiar with the nature of the test.

The words of the test proper were printed in large, plain type on
the unexposed side of the revolving blackboard. After warning the
children to be on the alert, the board was quickly turned so as to
expose the list of words, and the stop-watch was set going. On perceiving
the words, the children immediately began to construct sentences con-
taining them ; writing down the results on papers. At the expiration
of 2^ minutes the children ceased work and the papers were collected.
The list selected was such that each word was in some manner related
to the succeeding word, e.g. Circle, Moon, Night, Sleep, etc.

System of marking and results: 5 marks were given for each
sentence showing the closest possible connexion between the words ;
4, 3, 2, 1 or 0 marks were given according to the degree of deviation
from this standard.

1 This ia a modification of a test devised by Mr H. S, Lawson of Buxton College.

8—2

116 Quantitative hivestigatioii of Mental Processes

The following are quoted from actual examples :

(1) The full moon is like a circle. (5 marks.)

(2) The moon is circular in shape. (4 marks.)

(3) A ball is a circle, I thought it was the moon. (3 marks.)

(4) The circle moon is round. (2 marks.)

(5) A circle is round and the moon is bright. (1 mark.)

(6) The circle and moon. (0 marks.)

Group I Group II

Mean 25-4 25-0

Standard Deviation (c) 8-45 8-12

Reliability Coefficient "83 —

Correlation with Intelligence (»•) '62 -60

P. E. ofr -07 -07

For the second test a different list of words was used, but presenting
about the same amount of difficulty. In this test an irrelevant factor
is the time taken in writing down the sentences ; it varies much ac-
cording to the length of the sentence to be written, and the subject's
speed of writing. A sentence may be perfectly correct, but of greater
length than another of equal value ; the time taken in writing the one
will be much greater than that of the other, a fact which will certainly
tend to make the results misrepresentative.

A child who is obviously inferior to another in mental ability may,
by constructing shorter sentences and by writing more quickly, equal
or even surpass the latter in this test. The sentences of the one may
be very simple but correct; those of the other may be longer but
involve deeper thought.

5. Analogies Test. This test is based on the principle of pro-
portional parts. Three terms were given, and the children were
required to find a fourth. Thus in the following examples, which were
used to illustrate the test, the children were presented with a definite
relationship as expressed by the first and second terms ; a third term
was given, bearing the same relation to the unknown term as the first
term bore to the second.

(1)
(2)
(3)
(4)

(5)

Much time was spent in making clear the nature of the test, the
children themselves being asked to construct examples in order to show

Storm

: Calm :

: War

: x'i

Arm

■ Elbow :

: Leg

: xl

Good

Better :

: Much

: x'i.

Sound

Echo :

: Seedtime

: x'i.

Known

: Unknown :

: Present

: x'i.

Stanley Wyatt 117

how far the instructions had been understood. Twenty-five such ex-
amples were printed and the papers given out face downwards ; at a
given signal the children turned them over and began the test. On
a second signal being given, five minutes later, the papers were again
turned face downwards and immediately collected. The children were
allowed to pass over any example which presented difficulties. Examples
requiring general knowledge only were given.

System of marking and results: 4 marks were given for each correct
solution ; 3, 2, 1 or 0 marks were given for partially correct solutions
according to the degree of correctness.

Group I

Group II

Mean

45-7

54-7

Standard Deviation {cr)

22-8

161

Reliability Coefficient

•92

—

Correlation with Intelligence (r)

•80

•62

P. E. of r

•04

•07

The correlation with Intelligence is remaFkably high (80 in Group I)
and individual differences were fairly well defined. In many cases in-
ability to perceive relations was obvious, words being supplied which
were distinctly irrelevant. In this test the examples may be constructed
so as to present varying degrees of difficulty and consequently may be
applied to children of widely different ages and ability.

6. Completion Test. This test consists in supplying the missing
words in a passage of prose from which a number of words had been
omitted. In order to acquaint the children with the requirements of
the test, a preliminary trial was first given ; this brought to light any
cases of misunderstanding of the directions.

The papers were given out face downwards, and at a given signal
the children turned the papers over and commenced to read through
the piece carefully with a view to grasping its general meaning.
Nothing was written during this period, which lasted for five minutes,
but at its termination a second signal was given and the children then
began to fill in the blanks. It was suggested that if a certain elision
presented difficulty, it could be passed over for the time being, as the
succeeding context would probably supply the necessary cue. This
period lasted for ten minutes, at the end of which a signal was given
for the children to stop.

The passage used was the following* :

* Given in Whipple: Manual of Mental and Physical Tests. Baltimore, U.S.A., 1910,
448.

118 Quantitative Investigation of Mental Processes

One eagle with the '. birds

see.. could highest agreed

he who fly should called

strongest All started same and

away among clouds . One by they

weary re , but eagle upward and

until was mere speck heavens. When he

back, others were for him; and

touched a linnet off back where

hidden and that himself

strongest " stronj^er

;" said the , "for not did I

as high, but he began downward

, I my hiding and up little

this boastful the

their heads and council to the matter. After

long decided the

the bird, not only he

80 high, but the as well.

To day plumes

are emblems of str and con

System of marking and results: each blank filled in correctly
was awarded 1 mark ; each blank filled in incorrectly was awarded
— 1 mark ; each blank passed over was awarded — | mark.

Group I Group 11

Mean 34-6 42-0

Standard Deviation (o-) 13-8 11-6

Reliability CoeflBcient '89 —

Correlation with Intelligence (r) -85 '61

P. E. of r -04 -07

For the purpose of repetition, the same passage was given and the
method of procedure was the same except that the five-minute period
for reading through the passage was omitted. The test thus occupied
ten minutes, the children starting to fill in the blanks immediately.
The correlation with Intelligence is unusually high {'8o in Group I)
and hence this and the previous test would seem to offer very accurate
indications of the child's level of intelligence.

This test correlates highly with the Analogies Test ('85) ; this being
the highest correlation of any one of the tests with any other.

7, Missing Digits. This is a test which involves an entire modi-
fication of the habitual methods of working. An unfamiliar situation
is presented to the subject, and the facility with which he can escape
from his fixed habits in order to cope with this novel situation affords
some indication of the degree of intelligence he possesses. The examples

Stanley Wyatt llO

given were illustrations of the four common arithmetical processes of
Addition, Subtraction, Multiplication, and Division; they were of the
most elementary type, and no special knowledge was required to effect
their solution.

Example : 2»94

•867
781 •

•42«6

Various digits in each example were omitted, the omissions being
denoted by a dot. Papers, which contained one example each of the
four classes named, were given out face downwards, and at a given
signal the children turned them over and began to insert the missing
digits. Five minutes later a second signal was given, whereupon the
class again turned the papers over. It should be noted that no in-
dication was given of the class to which each example belonged, except
that a preliminary test, containing one example of each type, was first
given.

System of marking and results : + 4 was given for each example
wholly correct ; — 4 was given for each example wholly incorrect and
+ 2, 0, or — 2 was given for each example partially correct according to
the extent of error.

Group I

Group II

Mean

2-85

6-4

Standard Deviation (cr)

5-76

6-65

Reliability Coefficient

•69

—

Correlation with Intelligence (r)

•65

•50

P.E. ofr

•07

•08

The inserted digits were often incorrect, and as a result negative
scores were obtained in many cases, thus making the average very low.
Often the children failed to recognise the type of example given ;
frequent attempts being made to solve an example in Addition by
means of the Multiplication process.

8. Part-wholes. In this test a list of ten words was given, and the
children were required to write opposite each word the name of the
whole of which the word given was a part. In order to familiarise
them with the nature of the test, replies to the following examples
were first elicited: eye, page, spire, wing, pillow.

The papers, on which the test words were printed, were given out
face downwards, and at a given signal the children turned them over
and began to write the appropriate words as quickly as possible. The

120 Quaiititative Investigation of Mental Processes

time allowed was thirty seconds, at the end of which the papers were
again turned face downwards, and then collected.

System of marking and results : 2 marks were given for each correct
associate ; 1 mark was given for each associate partly correct, and
0 marks were given for any incorrect associate.

Group I Group II

Mean 13-35 1310

Standard Deviation ((r) 4-6 4-36

Keliability Coefficient '65 —

Correlation with Intelligence (?) -67 "56

P. E. of r -07 -08

Unless the time allowed for this test had been just insufficient for
anyone to complete it, the correlation with Intelligence would have been
lower. The test is so easy that it could have been successfully accom-
plished by almost all the children if only sufficient time had been given,
but the different rates at which the associates were evoked caused a
differentiation closely related to the respective mental abilities of the
children when the time allowed was only thirty seconds. Probably a
longer list of words would make a better test.

9. Reconstruction of Dissected Pictures. This test was devised on
the lines adopted by Burt\ Each subject was provided with a picture
postcard containing a reproduction of Gilbert's " Shylock, Salanio, and
Salarino." This was divided into fourteen parts which were arranged
in haphazard order, the same order, however, being maintained for each
subject. This minimised the possibility of any chance selection of an
easy or difficult fragment. The fragments were so arranged before the
children entered the room, and were covered by papers. An intact
postcard of the same scene was turned face downwards on the table.
At a given signal this was turned over, and the children examined it
carefully for thirty seconds. A second signal was given, on hearing
which the children again turned the card face downwards, and then
began to build up the picture from the disorganised fragments. The
results were measured in terms of the time taken to reconstruct the
pictures ; a rather difficult procedure when only one experimenter was
present. However, the following method, which enables measurements
correct to two seconds to be taken, was devised.

The experimenter was provided with a stop-watch and a sheet of
squared paper. The horizontal axis of the paper was marked off into

1 Op. cit. 102.

Stanley Wyatt 121

divisions representing seconds, and the vertical axis into divisions re-
presenting minutes. Each subject was told to put up his hand on
completing the task; the experimenter, being able to see each in-
dividual in the group, could note immediately such a movement, and
at the* same time could note the time on the stop-watch. If the
experimenter throughout the test move his hand, holding a pencil,
horizontally along the paper from left to right, at such a rate that the
square on the paper at which the pencil is pointing is always equivalent
to the time of the stop-watch, then on noting the raising of a hand, he
can simultaneously place a dot on the paper, and the position of the dot
represents the time correct to two seconds. The difficulty of knowing
which subject corresponded to each dot was overcome by calling out a
number as each hand was raised, according to the number of subjects
who had already completed the test. Thus when the first to finish put
up his hand, the experimenter called out ' one,' and the subject to
whom it referred put this number down on his paper. The next to
finish was given number ' two ' and so on until all had completed the
reconstruction. The experimenter could then read oflf at leisure the
time taken by each subject, many of the times by this method being
correct to one second. The method certainly involves alertness and
coolness on the part of the experimenter; but when these are assured,
it is quite successful.

Group I
Mean 144 sees.

Standard Deviation {a) 42'15

Reliability Coefficient -92

Correlation with Intelligence (r) '63

P. E. of T -07

From the results with earlier trials with fewer fragments it would
appear that the correlation with Intelligence is dependent upon the
number and complexity of the cuttings. When the postcard is divided
into eight rectangular pieces, the correlation with Intelligence is much
lower. This is probably due to the fact that in such a case the times
are too short for definite differentiation and the factor of fortuitous
selection becomes relatively more pronounced. It is conceivable that
such a test can be made to suit all grades of intelligence, from the most
backward person to one of highest intellectual ability. The correlation
with Intelligence is fairly high ('63), and a higher correlation might be
expected by making the test more complex. The correlation with some
of the other tests was not so high as one might have expected, this
' dissected pictures test ' occurring rather low down in the ' hierarchy.'

122 Quantitative Investigation of Mental Processes'

10. Rearranged Letters. In this test the letters of any number
word (one, two, etc.) were arranged in haphazard order, and the children
were required to find the particular number which could be spelt from
the letters given. The words were printed on papers thus :

VFEI =
RFUO =

and opposite each word the children wrote the number down as a
figure, thus:

VFEI =5

RFU0 = 4

Ten such words were used and the children were informed that only
numbers from one to twenty (inclusive) would be given. Demonstrations
on the blackboard were first given, and then the papers, each containing
the ten words, were distributed face downwards. At a given signal the
papers were turned over and the children began the test. One-and-a-
half minutes later a second signal was given for the children to stop,
and the papers were collected.

System of marking and results : 1 mark was awarded for each
correct solution and — 1 mark for each incorrect solution.

Group I

Mean

4-75

Standard Deviation {<t)

1-99

Reliability Coefficient

•65

Correlation with Intelligence (r)

•72

P. E. of r

•06

11. Memory {immediate reproduction) for Nonsense Syllables. The
object of this test was to ascertain how far immediate mechanical
memory correlates with Intelligence. For this purpose a list of ten
nonsense syllables was printed in large, plain type on the revolving
blackboard and was exposed to the view of the children for a period of
three minutes. At the end of the three minutes the children counted
from twenty backwards at the rate of one per second (the pace being
set by the experimenter) and then began to write down the syllables
remembered. By this means, attention was distracted from any primary
visual memory image of the syllables that might otherwise have
intruded.

No directions were given, as to how the syllables were to be learnt
except that they must be reproduced as far as possible in the right
order.

Stanley Wyatt 123

System of marking and results : 1 mark was awarded for each

letter correctly reproduced ; 1 mark was deducted for each error of

transposition or of insertion of letters within syllables ; a half-mark

was deducted for each error of transposition of the syllables.

Group I

Mean

140

Standard Deviation {<t)

8-7

Reliability CoefBcient

•76

Correlation with Intelligence (r)

•59

P. E. of r

•07

Diflferent lists of syllables were used in the two applications of this
test. Among the results of the second test it was noticed that syllables
were occasionally reproduced which had been given in the first test,
generally in the correct position in the series, and yet the interval
between the two tests was six weeks. .

12. Memory {delayed reproduction) for Nonsense Syllables. In
order to test retentiveness after an interval of exactly two days, the
children were asked to write down the syllables they had previously
learnt. No warning had been given that such a performance would be
required of them, consequently, it was unlikely that any effort had been
made in the interval to retain the series.

System of marking and results : the system of marking adopted
was the same as in the last test.

Group I

Mean

5-25

Standard Deviation {a)

71

Correlation with Intelligence (r)

•74

P. E. ofr

•06

In connexion with the second series of nonsense syllables given,
the amount retained after two days was not tested because some of
the children, remembering the procedure in the earlier case, had en-
deavoured to retain the syllables learnt.

The difference between the means of this and of the previous test
(5*25 and 140 respectively) indicates how rapid is the rate of forgetting
in its earlier stages. In this test the correlation with Intelligence is
very high, "74, while in the previous test it was "oO ; thus providing a
measure of the degree to which intellectual capacities are more con-
cerned in prolonged retention than in immediate recall. It is reasonable
to assume that a longer interval between the learning of the series and
their recall would give even a higher correlation with Intelligence.

124 Quantitative Investigation of Me^ital Processes

Some children scored more in this test than in the immediate
memory test ; more was actually remembered after an interval of
two days than immediately after the exposure of the syllables, thus
indicating that 'consolidation' had been at work in the meantime.

This is a test which correlates highly with those tests high in
the ' hierarchy ' (the correlations ranging from '5 to 7) and also with
' Immediate Memory for Nonsense Syllables ' ('71). Thus those children
who were best able to retain senseless material also did best in those
tests which give the highest correlations with Intelligence.

13. Memory {immediate reproduction) for Letter Squares. The
material for this test consisted of all the consonants of the alphabet ;
these were plain letters cut out of white paper and gummed on to stiff
dark grey paper. The letters were oj inches high and 3^ inches wide,
and the grey background of each letter was 7f inches by 7 inches, — a
size sufficiently large for all to see conveniently. The letters being
separate allowed of any possible combination being readily formed.
For the purpose of exposure the revolving blackboard was used. It
was found convenient to use this test as the first test of the sitting ;
the letters could be fastened on to the blackboard by means of
drawing pins before the children entered the room, thus no waiting
was necessary.

The nature of the test was first explained to the children but no
directions were given as to how the letters should be learnt. The
number of letters exposed at a time was twelve, and the children were
provided with papers ruled into twelve squares for the purpose of
reproducing in the right order, as far as possible, the letters learnt.

The children were warned to look at the blackboard ; this was then
rotated quickly so as to expose the letters, and at the same time the
stop-watch was set going. At the end of twenty-five seconds the black-
board was again turned over and five seconds were allowed to pass
before the children began to fill in the squares; thirty seconds being
allowed for the actual reproduction.

The letters used for the first and second tests were :

1st test 2nd test

MTDX KXPE

VLPN CFLZ

SZQR HWQU

System of marking and results : 3 marks were given for each letter
in the correct position ; 2 for each letter one remove to the right, left.

Stanley Wyatt 125

above, or below ; 1 for each letter two removes to right, left, above, or
below, e.g.

M D

X

T^

(3) (2)

(2)

(1)

S L

P

R

(2) (3)

(3)

(2)

■ 24 marks out of f

I possible a

V H

R

Q

(2) (0)

(2)

(2)

Group I

Mean

17-4

Standard Deviation {a)

4-06

Eeliability Coefficient

•75

Correlation with Intelligence

(r)

•18

P.

E. of

r

•11

This test gives the lowest correlation with Intelligence ('18) and is
also lowest in the 'hierarchy.' It thus appears that intellectual pro-
cesses function only slightly in this teat, the process of imagery being
relatively more prominent.

With some tests such as Word-buildiug, Completion, and Missing
Digits, it shows scarcely any correlation. The correlation with In-
telligence is remarkably diflFerent from that of Memory for Nonsense
Syllables and Intelligence ("SO). Considering that both are 'memory
tests,' the results seem to point to the fact that different mental
processes are involved in the two methods. In addition to the im-
portant part played by imagery in the Letter-square test, the recall
was more immediate than in the Nonsense Syllable test. Further, a
relatively longer time was allowed for the learning of the syllables, and
from introspections obtained, the memory of the backward children
suffered more at the hands of the 20 seconds period than did that of
the intelligent. The learning of the nonsense syllables requires more
effort and concentration of attention than the learning of the letters,
and as the duller children have not the same power of keeping their
attention focused on a task over a considerable period, this factor may
have been instrumental in causing the widely different results.

14. Gross-Line TestK As a preliminary test the following figure
was drawn on the blackboard :

X

' "Tests for Practical Mental Classification," by Wm. Healy and Grace M. Fernald.
Psychol. Rev. Psychol. Monogr. 1911, xiii.

126 Quantitative Investigation of Mental Processes

It was pointed out to the children that the figure was made up of
four compartments, each containing a number. These compartments
were then drawn separately thus :

and the children were asked to give the number to which each com-
partment corresponded. When it was apparent that the children
understood the requirements of the test, the figure proper (Fig. 1),

12 3
4 5 6

7 8 9

Figl.

having been previously drawn on the reverse side of the black-
board, was exposed for a period of twenty seconds. During this period
the children examined carefully its constitution. Papers, on which the
separate compartments (Fig. 2) were printed, had been previously

Fig. 2.

given out face downwards. The children now turned these over and
commenced to write in each compartment the number corresponding to
it. When one subject had completed the test, the signal to stop was
given, and the papers were collected.

System of marking and results : 1 mark was given for each number
correct ; — 1 mark for each number incorrect.

Group I

Mean

3-7

Standard Deviation (a) '.

4-44

Beliability CoeflBcient

•82

Correlation with Intelligence (r)

•46

P. E. ofr

•09

For the second application of the test the numbers in the com-^
partments were redistributed^

1 This of course involves the difficulty that a previously formed association has to be

dissolved before the new one is formed.

Stanley Wyatt 127

The correlation with Intelligence ('46) is lower than usual ; this is
explicable upon the supposition that the successful interpretation of
this test is closely related to vividness' of visual imagery. The corre-
lations with the other tests are low, generally lying between 2 and '4
except in the case of the Dissected Pictures test, the correlation here
rising to '54. Since both these tests involve memory of spatial re-
lations and the recognition of parts, their comparatively high correlation
is quite conceivable. The Letter-Square test also gives a higher
correlation with the Cross-Line test than with any other ; but both
these tests involve imagery, and probably visual imagery in particular.

15. Interpretation of Fables. This test was introduced in the hope
that it would give some indication of the child's ability to grasp the
meaning of a passage of prose in the form of a fable.' The whole point
or moral of a fable can usually be expressed in a very few words, whilst
the inherent interest makes it easy to hold the whole story in the mind
when endeavouring to extract the meaning.

The following five fables were used^:

(1) The Boy and the Filberts.

(2) The Stork and the Cranes.

(3) Mercury and the Woodman.

(4) The Eagle and the Tortoise.

(5) The Ants and the Grasshopper.

Each fable was read twice by the experimenter, the second reading
following immediately after the first. At the end of the second reading
the children wrote down what they thought to be the moral of the
fable. This process was repeated with each of the remaining fables.

System of marking and results : 5 marks were given for each correct
interpretation; 4, 3, 2, 1, or 0 marks were given for incorrect inter-
pretations, according to the degree of deviation from the correct reply.

Group 1

Mean

11-5

Standard Deviation (ff)

5-7

Correlation with Intelligence (r)

•64

P. E. of r

•07

This test was not repeated, as afterwards some of the children
pondered over the fables given, and hence made the conditions for

1 These fables are part of a selection originally made by L. M. Terman, and are to be
found in " A Tentative Revision and Extension of the Binet-Simon Measuring Scale
of Intelligence," by L. M. Terman and H. G. Childs. J. of educ. Psychol. 1912, iii.
No. 3.

128 Quantitative Investigation of Mental Processes

a repetition unequal. The interpretations were very varied ; some
showing no connexion whatever with the original passage ; others
exhibiting a thorough grasp of the matter and displaying ability in
formulating the moral. None of the children had previously made the
acquaintance of any of the fables.

V.

Sixy)

Throughout this investigation the product moment formula ^ "

J.V 0"i (T<i

was employed. The coefficients of correlation were calculated from the
results obtained by amalgamating the two measurements made in each
test, and the reliability coefficients represent the extent of the corre-
lation between the series obtained by adding the first half of the first
performance to the second half of the last performance, and by adding
the second half of the first performance to the first half of the last
performance.

In Tables III and IV (see pages 131, 132) the coefficients of corre-
lation of each test with every other are given. These are arranged in
such a manner that the sums of the columns or rows of coefficients are
in descending order of magnitude from left to right or from above
downwards.

Comparison of Groups I and II. The following table gives the
mean, standard deviation (a), and correlation with Intelligence (r) of
each of the tests applied to the two groups.

TABLE I.

Test

E, R

A, N, 0, S

Word-building

Sentence construction

Analogies

Completion

Missing Digits

Part-wholes

Mean

Group I

Mean

17-75
41-2
10-5
25-4
45-7
34-6
2-85
13-35

23-9

Group II

Mean

6-4
12-1
4-08
8-45
22-8
13-8
5-76
4-6

9-75

-40
•45
•58
•62
•80
•85
•65
-67

•63

12-3

12^8

37-0

16^4

12-5

4 42

25 0

8^12

54-7

161

42-0

11-6

6-4

6-65

13 10

4^36

254

10-05

-37
•32
•50
■CO
•62
•61
•50
•56

•51

In those tests which give the highest correlations with Intelligence
the means of Group II are higher than those of Group I, although the

Stanley Wyatt 129

average age of the former group is slightly less than that of the latter.
The children of Group II were drawn from the wealthier parents, and
consequently it appears justifiable to infer that in this case, at least,
the children of superior intelligence belong to the socially superior
parents. This view is in harmony with Burt's conclusions^

The rank correlation" of r for Groups I and II is "89 ; which in-
dicates that the relative value of the tests as a means of indicating
' general intelligence ' is approximately the same in both groups.

The correlations with Intelligence are unu.sually high in the case of
Group I. This is probably due to the following factors :

(1) The teacher's estimate of intelligence was not based on the
class lists, but on his own opinion of the respective intellectual
capacities of the children;

(2) the perfect conditions which prevailed during the application
of the tests ;

(3) prolonged preliminary trials ;

(4) absence of irrelevant factors.

All the tests in Group I were applied by the writer, and thus any
discrepancy due to the presence of an untrained experimenter was
eliminated. Also it is evident from the table that in some cases the
standard deviations of the same tests are widely different in the two
groups.

Comparison of Results with those of other investigators. In the
following table the results of this investigation are given along with
those of other investigators who have employed similar tests. The
values represent the coefficients of correlation between various tests
and the subjective estimations of intelligence.

The upper and lower lines of Brown's results represent his Groups
II and III respectively. Similar representation of the writer's results
indicates respectively Groups I and II.

Thus the coefficients obtained by the writer are generally higher than
those obtained by the other investigators quoted ; the difference in all
probability being due to the reasons already given (see p. 109 above).

Evidence of a General Common Factor. According to Hart and
Spearman', if a General Factor underlies all the performances tested,

1 "Experimental Tests of General Intelligence." This Journal, 1909, iii. 176.

2 An explanation of the method of determining the rank correlation is given by
Brown, op, cit. p. 50; also in an article by P. Sandiford : "Educational Measurements,"
J. of exp. Fed. 1912, i. 217.

3 Op. cit. p. 59.

J. of Psych. VI 9

130 Quantitative Investigation of Mental Processes

TABLE II.

Test

Burt'

Brown 2

Wyatt

Completion |

Analogies

Erasure of letters E, E ... j
Erasure of letters A, N, 0, S ]

Mechanical memory |

Sentence construction

Dissected pictures

•48
•52
•39

•43 »
•75

•62

[•72]

•43
•69

0
■28
•13
•10
•55
•49

•85)

•61

•801

•62 1

•40)

•37 1

•45

•32

•59

•621
•60)

•631

the correlation between the pairs of columns of coefficients in Table III
should be high and positive ; that is, the correlation must approach
unity in proportion as the coefficients admit of perfect hierarchical
arrangement. To correct for the bias due to the presence of sampling
errors they give the following formula*

R ab =

S (pxaPxb) — (n — l)rab(r!ga(rscb

^S (p^^) - (n - 1) a'^ JS (p^^) - (n - 1) <r\, '
where R'ab is the correct value to be obtained

S{p\a) is the sum of the squares of the differences of the co-
efficients of column a from their mean, and similarly

sip'^y,

(Txa is the probable error of Vxa divided by '6745 and similarly axb.
To exclude columns which are influenced by the presence of large
sampling errors, Hart and Spearman require* " that in each of these
columns the mean square deviation should be at least double the
correction to be applied to that deviation." In the present investiga-
tion only five pairs of columns reach this prescribed standard^; these are

^ J. of exp. Fed., op. cit. p. 111.
* This Journal, op. cit. p. 145.

° S IpxaPseb) ^^ obtained in the form

Sip'

2 Op. cit. pp. 115, 116.
* Ibid. p. 82.
^) + S(p'^)-S^p^~p^r

-^iPxaPxb)-

this equals

* Op. cit. p. 56.

7 The writer here desires to acknowledge his indebtedness to Prof. Spearman, who
kindly supplied some corrections to the method adopted in determining the correlation
between the columns of coefficients.

Stanley Wyatt

131

Analogies and Word-building = '93

Completion and Word-building = '97
Completion and Part- wholes = 1"05

Word-building and Part-wholes = '99
Part-wholes and Memory (delayed) = "92

Mean correlation between columns = '97

TABLE III. (Group!)

Test

Analogies

Completion

Word-building

Part-wholes

Bearranged letters ...

Memory (delayed)

Missing digits

Sentence Construction

Fables

Nonsense syllables ...

Dissected pictures

E, R

A, N, 0, S

Cross-line

Letter-squares

■04

•65
■07

■70
•06

•77
■05

•63
•07

•66
■07

•72

•06

•52

■08

•47
•09

•57
•07

•30
•10

•56
■08

•49
•09

•15
■11

•006
•12

•63

•07

•62
•07

•63

•07

•59

•07

•61

■07

•63

•07

•50

•08

•38
■10

•48
•09

•53

■08

•65
•07

•41
•10

•32

■10

•16
■11

•63

■07

•34
•JO

•43

•09

•71

•06

•53

•OS

•34
•10

•40
•JO

•33
•JO

•15
•JJ

•43

•09

•47
■09

•56

•OS

•43

•09

•65
■07

•34
•JO

•40
•JO

•30
•JO

•20
•JJ

•29
•JO

•28
•JO

•57
■07

•17
•JJ

•11
•JJ

•54

■08

•43

■09

•49
•09

•46
•09

•41

•JO

•40
•JO

•46
■09

•45

•09

•37
•JO

•06
•J^

•26
•JO

•57
•07

•20
•JJ

•07
•13

•40
•JO

•33
•JO

•15
•JJ

•24
•JO

•32

•JO

•33
•10

•42

•JO

•28
•JO

•23
•JJ

•29
•10

•54

•08

•17
•JJ

•20
•JJ

•32
•JO

•28
•JO

•03
■12

•006
■12

•09
•JJ

•16
•JJ

•15
•JJ

•03
•12

•14
•JJ

•31
•JO

•25

•JO

•18
•JJ

•11
•JJ

•07
•12

•32
•JO

182 Qumititative Investigation of Mental Processes

Thus the correlation between the pairs of columns is high and
positive, and the contention of Hart and Spearman that a General
Common Factor exists receives further support.

TABLE IV. (Group II.)

Test

Completion

Part-wholes

Analogies

E, B

Word-building

Missing digits

A,N, 0, S

Sentence Construction

•69
■06

•54

•08

•48
•08

•36

•10

•38
•09

•32

•10

•53

•08

•58

■07

•54

•08

•19
•11

•54

•08

•54

•08

•14
•11

•39
•09

•58
•07

•48
•08

•19
•11

^

•36
•10

•36
•10

•54

•08

•39

•09

48

•37
■09

•00

•54

•08

•32
•10

•14
•Ji

•57

•07

•29
•JO

•37
•0.9

•29
•10

•57

•07

•53

•08

•39
•09

•27
•JO

•39

■09

•00

•29
•JO

VI.

The following conclusions may be drawn from this investigation :
1. The Analogies and Completion Tests give the highest corre-
lations with the subjective estimations of intelligence^ The inter-
correlations of each of these tests with every other are also high.

^ The practical value of these tests as a means of classifying children according to the
teachers' estimates of intelligence has since been demonstrated at the Fielden School,
Manchester. The tests were applied to seven children who desired to be admitted to the
school, and the papers were sent to the writer to be marked. From the results the writer
was able to suggest that five of the children should be placed in certain classes and that
the other two should not be admitted to the schools. The Senior Master, who had
interviewed each prospective scholar, and also tested their knowledge of aritlimetic,
reading, and writing, entirely agreed with the writer's classification, and accordingly the

Stanley Wyatt 133

2. The correlational coefficients of each test with every other
admit of hierarchical arrangement ; the theory of the General Common
Factor thus receives further support.

3. Memory of the kind displayed in the retention of nonsense
syllables, seems to be a prominent factor in the mental processes of
the higher levels. The correlational coefficients between memory for
nonsense syllables and those tests which correlate highly with the
subjective estimation of intelligence range from "5 to '7.

4. The children who are lowest in the scale of intelligence
(according to the teacher's estimate) are least able to retain senseless
material over a period of two days.

5. Some of the tests which are closely related in content give
relatively low coefficients of correlation with each other.

6. The rank correlation between the coefficients of correlation
(with Intelligence) of Groups I and II amounts to '89. This signifies
that the order of the performances measured tends very largely to be
the same in both groups, and consequently the relative values of the
tests as a means of estimating intelligence shows a very high degree
of similarity in the two groups.

7. The Letter-Square Test gives a higher correlation with the
Cross-Line Test than with any other. As both these tests involve
imagery, and probably visual imagery in particular, the comparatively
high correlation between them may be due to this fact. The corre-
lation of the Letter-Square Test with the Intelligence Classification is
very low.

8. In those tests which give the highest correlations with in-
telligence, the performances of the children of Group II are superior
to those of Group I.

9. In a random group of thirty subjects, though there are well-
marked individual differences, yet there are no abrupt transitions
between them.

children were dealt with as snggested. The class-work of each of the children admitted
has since confirmed the accuracy of the classification. The writer was equally successful
in predicting each subject's approximate position in class at the end of the term. Other
results which will further demonstrate the utility of these tests as a means of indicating
the child's level of intelligence, will be published later.

(Manuscript received 13 March, 1913.)

PUBLICATIONS RECENTLY RECEIVED.

Experimental Studies of Mental Defectives : a critique of the Binet-Simon
Tests and a Contribution to the Psychology of Epilepsy. By Dr J. E.
Wallace Wallin. Pp. vi + 155. Baltimore : Warwick & York,
1912. $1.25.

This is one of the series of Educational Psychology Monographs, issued under the
editorship of Prof. G. M. Whipple. The scope of the work, which appears to have
been carried out with the necessary care and ability, is sufficiently indicated in the
sub-title. The data were obtained during an eight months' residence of the author
in the New Jersey State Village for Epileptics. Comparisons are attempted
between the intelligence of the 333 epileptics examined by him in the Village and
378 feeble-minded inmates tested in Vineland, New Jersey, by Goddard. Unfortu-
nately, as the author points out, certa,in differences in the methods of testing make
such comparison difficult. Indeed the chief value of this careful piece of work lies
not so much in the study of mental defectives as in demonstrating the need for
improvement in the nature and arrangement of the Binet-Simon tests. By a
comparison of his own results with those of other workers, he shows that some of
the tests are too hard, while others are too easy, for children of a given age. He
also gives reasons for preferring the 1908-form of the tests to the modified form
proposed by Binet and Simon in 1911. The latter form, for example, omits the
reading test, which the author shows to be a valuable test for differentiating the
mental capacity of epileptics. But despite its present drawbacks, the author insists
that the Binet-Simon scale " does enable us to grade and classify defective individuals
far more rapidly and satisfactorily than would be possible by the ordinary methods
of observation " (p. 104). He also insists that it is the function of this scale " to
give us a preliminary, and not a final survey or rating of the individual who may be
tested." It is to be regarded " merely as a point of departure for further diagnosis"
(p. 109). In America, and doubtless in this country too, " the idea, unfortunately,
seems to be gaining ground that anyone, be he grade teacher, introspective
psychologist, practising lawj-er or general medical practitioner, is able to make
psychological diagnoses by putting the child through a few stock psychological
tests. Nothing is more preposterous" (p. 110). When the tests end, the work of
the expert psychiatrist begins.

A Psychological Study of Religion : its Origin, Function and Future. By
Prof. J. H. Leuba. Pp. xiv 4- 371. New York : Macmillan Co., 1912.

The author's main position in this interesting book is that while Psychology can
say nothing as to the ultimate value of the metaphysical doctrines used as a setting
for and as an intellectual interpretation of religious experience, the religious
experience itself belongs wholly to Psychology, and is to be subjected to psycho-
logical analysis equally with the non-religious parts of conscious life. He denies
that the religious consciousness is unique in the sense that it has specific affective
experiences or psychologically distinctive desires and purposes ; for him it is
fundamentally the same as the non-religious consciousness. " Any impulse, any
desire may lead to religious activity, and in it no type of emotion is to be found
which is not represented also outside." Leuba contends that as the actual gods of
historical religion are empirical gods, they are legitimately and entirely objects of
science. Theology, on the other hand, makes its appeal to our inner experiences,
either using them as the material for a quasi-scientific inductive proof of the
existence of a divine power, or treating them as if they bore in themselves the
marks of transhuman origin, as being immediate revelations of God. Leuba lays
stress on the easy confusion between the factual existence and quality of a given
subjective experience, on the one hand, and its transsubjective meaning and
objective validity, on the other.

Publications Recently Received 185

Psychotherapy. By Prof. H. Munsterberg. Pp. x + 401. London:
Fisher Unwin, 1909. 8». 6<f. net.

In this work, Miinsterberg discusses the psychological basis, the practice, and
the place of psychotherapy — that kind of treatment which, in contrast to physico-
therapy, seeks to treat the sick by influencing their mental life. The aim of the
book is " to counteract the misunderstandings which overflood the whole field " and
" to strengthen the feeling that the time has come when every physician should
systematically study psychology, the normal in the college years and the abnormal
in the medical school."

The Game of Mind: a Study in Psychological Disillusionment. By
P. A. Campbell. Pp. 80. New York : The Knickerbocker Press,
1913. 75 c. net.

A bright but shallow philosophical jest, the purport of which may be gathered
from the concluding sentences of the book. " With a sort of satisfaction we should
put ourselves in the introspective frame of mind of admitting that our ever-changing
mentality has its full source and flow in the organic life of the body. Of admitting
that perceiving, conceiving, knowing, etc., are after all only finer sorts of bodily
living. That remembering means reconstitution. That feeling is a discussion.
That consciousness is self-analysis. That mental evolution, for its part, is a
selective and mutatory handing on from the past, as a mechanistic gift to the present
and future, of the great bodily GAME OF MIND."

Insomnia : its Causes and Treatment. By Sir James Sawyer. Pp. 107.
Biraiingham : Cornish Bros., 1912.

Eye-Strain in Everyday Practice. By S. Stephenson. Pp. viii-j-139.
London: The Ophthalmoscope Press, 1913. 3s. 6c?. net.

In the Abstract. By Norman Alliston. Pp. 126. London : Swan
Sonnenschein & Co., 1909. 2s. 6c?.

I

PROCEEDINGS OF THE BRITISH PSYCHOLOGICAL

SOCIETY.

March 8, 1913. The Psychological System of Sigm. Freud, as set forth in Chap.
VII. of the "Traumdeutung," by W. Bbown.

Stem's Tonvariator (Demonstration), by W. Brown.

The Analysis of some personal Dreams, with special reference to
Freud's Interpretation, by T. H. Pear.

May 3, 1913. Notes on a case of Morphomania, by F. Aveling.

Wonder, Fascination and Curiosity, by Carveth Read.

ERRATA.

Vol. V. p. 421, line 17, for " sff^" read "r^."
„ „ p. 421, line 29, for "V^" read "r„j."

Volume VI OCTOBER, 1913 Part 2

ARE THE INTENSITY DIFFERENCES OF
SENSATION QUANTITATIVE?^ I.

By CHARLES S. MYERS.

/. Introduction.

§ 1. Initial assumptiotis.

§ 2. ' Intensiveness ' ^ extensiveness' and ' protensiveness.'

§ 3. Intensity and movement.

§ 4. Mental ^processes' and mental ^products.'

II. The Nature of Intensity Changes.

§ 1. The biological conditions of consciousness.

§ 2. The * all or none ' principle in spinal reflexes.

§ 3. The same jjrinciple in muscle-fihre.

§ 4. The same principle in nerve-fibre.

§ 5. The same principle in the heat and cold spot system of

sensibility.

§ 6. Pain in relation to other forms of cutaneous sensibility.

§ 7. The grading of ^clonic' spinal reflexes.

§ 8. The grading of auditory sensations.

§ 9. The grading of ' tonic' spinal reflexes.
§ 10. The grading of '^ tonic' sensations.

§ 11. Intensity, quality, and extensity in graded 'tonic' sensibility.
§ 12. Suggested relation of the attributes of colourless and colour

sensations.

III. Conclusions.

I.

§ 1. I ASSUME at the outset that the three following propositions
will meet with general acceptance^. The first is that, whereas Weber's
law is a direct expression of the data of sense experience, Fechner's

^ A contribution to the Symposium presented at the Joint Meeting of the British
Psychological Society, the Aristotelian Society, and the Mind Association in London,
7 June, 1913.

^ Reasons for accepting them will be found in m.y Textbook of Experimental Psychology,
2nd ed. 1911, i. 249-253.

J. of Psych. VI 10

138 The Intensity Differences of Sensation

law has no such immediately psychological basis, being derived from
the application of mathematical symbols and processes to those data.
The second is that we are incapable of measuring the intensities of
sensations, qua sensations; we cannot say that the sensation evoked
by an arc light contains so many units of the sensation evoked by a
candle light, — we can only range sensual intensities in a graded series.
The third proposition is that intensity experiences and intensity differ-
ence experiences are fundamentally similar in their dependence on past
experience and (unconscious or conscious) comparison, and in the con-
ditions governing their respective (absolute and differential) thresholds.
What may follow from the discussion on intensities in this paper will
therefore apply also to intensity diflferences which are the subject of this
symposium.

§ 2. From the acceptance, however, of these three propositions,
it by no means follows that differences in sensual intensity (or that
sensual intensities themselves) are qualitative and not quantitative.
It may well be that sensual intensities are dependent for their increase
or decrease on an increase or decrease in the quantity of sensation,
notwithstanding our inability to isolate the component units of which
any intensity may really be the sum.

Even in the case of a given experience of extensity of sensation, we
can only say that one sensation is more or is less extensive than another.
We can hence only arrange sensations in the order of their extensity ;
qua sensation we cannot say that one is twice as extensive as another.
But we can conclude that one line or area (not the sensation of a line
or area) is twice as extensive as another; and here lies the difference
between extensity and intensity. Whereas, on the basis of extensity,
we can directly measure the 'extensiveness' of objects^, we can only
measure their 'intensiveness' indirectly in terms of 'extensiveness' — e.g.
we express the intensity of visual or auditory stimuli in terms of
units of amplitude (extensiveness) of wave movement, and the intensity
of a weight or taste stimulus in terms of units of matter (extensiveness)
of the lifted or tasted object.

Now just as sensual extensity, when conjoined with movement,
assumes the form of quantitative extensiveness (spatiality), so sensual
protensity, though immeasurable qvd protensity, is capable of giving
rise to a quantitatively measureable 'protensiveness' (time). Indeed

^ I agree with Ward: "...before and apart from movement altogether, we experience
that massiveness or extensity of impressions in which movements enable us to find
positions, and also to measure " (Encycl. Brit. Article).

Charles S. Myers 139

extensity and protensity differ strikingly from intensity and quality,
in that the experiences of space and time, to which the former give rise,
are lived through, whereas those corresponding to intensity and quality
are lived in; the one pair of experiences are essentially transitional, the
other pair punctual. We can at the outset integrate extensity and
protensity with movement in a way impossible in the case of intensity
and quality.

§ 3. But nevertheless the connexion of intensity with movement
is obviously intimate. Power and intensity are practically synonymous.
The very word inteusity means a state of tension or strain. What,
then, is more natural than to suppose that our experience of intensity
depends ultimately on the muscular strains exerted to resist force —
or even, w^ith Mtinsterberg^ to deduce intensity experiences from
kinaesthetic sensations?

But if differences in intensity, say of visual sensation, are to be
traced to differences in intensity of kinaesthetic sensation, it is difficult
to see why we should be able to appreciate, as we can, a difference
between 100 and 101 units of intensity of two lights, whereas we are
hardly able to appreciate the difference between 100 and 103 units of
mass in two lifted weights I Moreover^ the explanation is only pushed
a step further back; we have yet to consider how sensations from the
locomotor apparatus can vary in intensity.

§ 4. It is true that the sensations of muscular strain have in the
past been held wholly responsible for the 'sense' (or 'feeling') of effort
in all self-activity. Unquestionably afferent impulses of motor origin
are of prime importance for our experience of activity, just as they and
other {e.g. visceral) impulses are of prime importance for our experience
of emotion. But there is now, I think, a growing consensus of opinion

1 Beitr. z. exp. Psychol. 1890, Hft. 3, 23.

2 I have perhaps unconsciously adopted this argument from A. Aliotta's La misura in
psicologia sperimentale (Firenze, 1905, second footnote, page 106), which since writing the
above I have re-read after a long interval. I may add that Aliotta criticizes Bergson's
contention in Les donnees iminediates de la conscience (Paris, 1889, 32 fif.) that to deter-
minate values of a stimulus correspond particular qualities of sensation, and that from
experience we are led to associate with each such quality of the effect (sensation) the idea
of an intensive quantity of the cause (stimulus). But, asks Aliotta (pp. 105, 106), how do
we know that the intensity of the stimulus is changing except by previous sensations of
different intensity ? How can we transfer the magnitude of physical objects to our mental
experience unless the idea of intensity has already arisen in consciousness? Cf. also
Knight Dunlap {Psychol. Rev. 1912, vi. 425, 426) for his criticism of Titchener's view
{Textbook of Psychology, Pt, 2, 140) of the connexion between kinaesthesis and the
perception of relation.

10—2

1 40 The Intensity Differences of Sensatioii

that the acts of the self are, in themselves, conscious, apart from
sensations of peripheral origin, and that we should make two main
divisions of consciousuess — the consciousness of 'acts' or 'processes'
(e.g. the 'acts' of attending, imagining, remembering, thinking, willing)
and the consciousness of ' contents ' or ' products ' (e.g. ' what ' we attend
to, ' what ' we imagine, ' what ' we remember, ' what ' we think, ' what '
we will).

11.

§ 1. Consciousness makes its appearance in life whenever a choice
on the part of the organism is possible between two or more reactions
to a given stimulus. So long as behaviour is fixed (in living matter
it is of course never absolutely fixed), there is no consciousness — or at
all events no consciousness in which the organism's 'self shares. But
with the development of instincts, fixity gives places to plasticity;
a certain choice of reaction is open; a certain im provability is possible
by practice and imitation ; whereupon (as I have before insisted^) a
rudiment of intelligence at once comes into play.

Now the correlate of differences in quality of a sensation consists in
differences in type of reaction. A sweet taste corresponds with one type of
reaction, a bitter taste with another; similarly with the sensations of colour
and pitch, different types of reaction are evoked from longer or shorter
waves. In their most primitive form, reactions are mainly those of
approach and withdrawal. Certain stimuli cause positive, others cause
negative ' taxis.' So long as the taxis is fixed, sensation is useless.
But with the evolution of the nervous system, as soon as plasticity
replaces fixity, vague ' affections ' of pleasure and pain enter, followed
at a later epoch by the evolution of ' sensations,' the number of possible
reactions to the same class of stimulus being simultaneously increased.
At bottom, differences in type of movement must be the cause of dif-
ferentiation in the quality of sensation; it would be of no advantage
for the organism to experience different qualities of sensation, unless
those differences were serviceable in promoting different types of
response ^

1 This Journal, 1909-10, ni. 209-218, 267-270.

2 Of course, in the developing adult we have to distinguish between the inherited
physiological basis of sensual qualities (and intensities) and his successful dififerentiation
of those qualities (and intensities) which is improvable by practice. That is to say. we
have to distinguish between the primary influence of heredity and the secondary influence
of education (environment), though the latter is ultimately dependent for its effects on
heredity.

Cfiarles S, Myers 141

§ 2. If the qualities of sensation correspond to dififerent types of
reaction, we may be inclined to attribute the intensities of sensation
to different degrees — moreness or lessness — of the same reaction. This,
however, leads to the consideration of what is involved in moreness or
lessness of a reaction. There are some reflex actions of the cord which
apparently cannot be graded in strength. For example, by pressing or
stretching the skin between or beneath the pads and cushion of the dog's
hind foot, a reflex known as the 'extensor thrust' reflex is evoked; the
leg is reflexly extended. (Reflexes can be best studied in the ' spinal
animal, in which the connexion between the cord and the higher parts
of the central nervous system has been severed.) Now the extensor
thrust reflex of the spinal preparation is little, if at all, altered by the
strength of the external stimulus. So long as the external stimulus is
adequate, whether it be relatively weak or strong, it produces practically
the same strength of response \ It is an instance of 'all or none' in
reaction. The reflex thrust is of very short duration, being easily
fati gable.

Hence, in the case of the extensor thrust reflex, a stimulus of
suitable strength and of suitable character, applied in a given situation,
evokes in the nervous tissue within the cord a certain pattern of
response, the issue of which is the discharge of efferent impulses along
certain nerve fibres supplying certain muscles ; and within fairly wide
limits, the movement resulting from this efferent discharge is inde-
pendent of the strength of the ingoing stimulus.

Now if in the spinal animal these limits are grossly overstepped, we
do not get an increased extensor thrust reaction ; quite another type of
reaction appears, — an immediate and well-marked flexion-reflex. In
other words the afferent impulses, on reaching the cord, evoke quite
another ' mechanism.' They evoke quite another pattern of nervous
response within the cord, so that we get a very different discharge,
causing the contraction, not of extensor, but of flexor muscles.

As we have seen, the extensor thrust reflex is an instance of an ' all
or none ' reaction. The reaction is either obtainable or unobtainable :
the stimulus is either effective or not ; there is no grading of the re-
action comparable to the grading of the stimulus. Nor is this ' all or
none' effect manifested in a reflex only. There can be little doubt
that muscle fibres and peripheral nerve fibres follow the same principle.

§ 3. In the case of the cutaneus dorsi muscle of the frog (which is
particularly suitable for investigation owing to its peculiar nerve
^ C. S. Sherrington, The Integrative Action of the Nervous System, London, 1906, 74.

142 The Intensity Differences of Sensation

supply), Lucas^ has been able to prove that the force with which the
muscle contracts as a whole is due simply to the number of the indi-
vidual muscle fibres involved in the contraction.

§ 4. Next, as regards nerve fibres (which are of more immediate
interest) — thanks mainly to the work of Gotch^ and Adrian* — there can
be little doubt that the principle holds good for them as it does for
muscle fibres. Gotch found (i) that the rate of propagation of the wave
of excitation (as shown by the concomitant electrical changes within the
nerve trunk) is the same whether the excitation is strong or weaker ;
and (ii) that the effects obtained by exciting only a portion of the nerve
fibres of a nerve trunk closely resemble those evoked by a weaker
stimulus applied to the entire nerve trunk. These results suggest that
a stronger stimulus to a nerve does not increase the strength of the
impulse passing down any one nerve fibre but merely leads to a greater
number of nerve fibres being involved. It is corroborated by Gotch's
observation that the electro-motive force in a stimulated nerve trunk
always takes the same time to show itself and to reach its height and
to disappear, however it be made to differ in amount by employing
different strengths of stimulation.

Adrian's experiments point to the same conclusion. He finds that
the time needed to narcotize a nerve trunk by alcohol or by other
means, so that the passage of the nervous impulse is blocked, depends
not merely on the length of the nerve trunk which is narcotized, but on
the disposal of that length. For example, if in one preparation 9 mm.
of nerve be narcotized, and if in another two lengths of 4*5 mm. of nerve
be narcotized, these shorter lengths being separated by a non-narcotized
interval of normal nerve, the times taken to narcotize one and to
narcotize both of the 4*5 mm. lengths are the same, while they are
considerably longer than that taken to narcotize the 9 mm. length.
" The disturbance [corresponding to the nervous impulse] has much
greater difficulty in passing one length of 9 mm. of affected nerve than
it has in passing two lengths of 45 mm." We must assume " that the
disturbance [corresponding to the nervous impulse] increases in size in
the normal area between the two lengths of 4o mm." And Adrian
brings forward evidence that " the increase of a subnormal disturbance
on entering normal nerve tissue is certainly complete before the disturb-
ance has travelled 5 mm. in the normal region, and it may be instan-
taneous " ; and that " the disturbance increases to a fixed size on entering

1 J. of Physiol. 1908-9, xxxviii. 113-133. ^ Ibid. 1902, xxvii. 395-416.

3 Ibid. 1912, XLV. 389-412.

Charles S. Myers 143

normal tissue no matter what its size may have been at the end of the
region of decrements" For these reasons he concludes that the nervous
impulse in the normal nerve fibre must obey the 'all or none' principle.

§ 5. So far as regards muscle and nerve. Their increase in intensity
of function seems to depend on a greater quantity of elements (muscle
fibres or nerve fibres) taking part in the action. With a weak stimulus
only a few elements respond : with a stronger stimulus, other (less
sensitive) elements are also involved. Each element follows the ' all or
none ' principle, which we have seen exemplified in the extensor thrust
reflex.

How now in regard to sensations ? Have we any sensations which
behave similarly ? Owing to the careful work of Head, Rivers and
Sherren, the sensations afforded by the heat and cold spots of the skin
can now be shown to afford an excellent instance of this behaviour^
Different heat and cold spots are in different degrees sensitive to heat
and cold. But each apparatus acts explosively and is easily fatigued.
If the stimulus is strong enough, it reacts ; if a still stronger stimulus is
employed, it does not react differently. Hence we have here a clear
instance of the ' all or none ' principle. We have relatively sensitive
and relatively insensitive heat and cold spots, and presumably these
reflexly produce relatively considerable and relatively weak reactions.
The differences in sensual intensity are correlated on the reflex level
with differences in the strength of individual reactions. We may
suppose that the nervous impulse from a more sensitive heat spot
spreads centrally and hence efferently to a greater number of nerve
fibres than are reached by the stimulus of a less sensitive heat spot.

§ 6. Thus the heat and cold spot sensations (and probably the
other sensations belonging to the ' protopathic ' system of sensibility),
appear to be analogous instances of the.' all or none' principle. Each
heat or cold spot, like the provocative of the extensor thrust reflex,
gives a reaction which, within wide limits, is independent of the
strength of the stimulus, provided that it is effective. But suppose
these limits are overstepped. We have seen (page 141) that, with
excessive increase in the strength (or with change in the character)
of the stimulus, the extensor reflex suddenly gives place to the flexor
reflex. So too, with like changes in the thermal stimulus, the quality
of the sensation alters. The threshold of a new quality of sensation
is overstepped : pain enters and quickly suppresses the temperature

• Op. cit. 399, 402, 412. 2 Brain, 1905, xxvm. 105.

144 The Intensity Differences of Seiisation

sensation^ We cannot stay here to discuss whether the same end
organs can give rise to pain and, with weaker strengths of stimuli, to
heat, cold and touch. The important point now to observe is that the
qualit}' of the thermal or tactual sensation at once changes, just as
the quality of the spinal reflex changes. A relatively innocuous
'exteroceptive' reaction, to use Sherrington's useful terminology, gives
place to a 'nociceptive' reaction. And, just as in the case of the
reflex, the pattern of impulse within the cord is different, so in the case
of the sensation, the paths of the impulses as they pass up to the brain
are (sooner or later) changed. At all events we have a hint that a
difference of ' specific energy,' as it used to be called, may depend not
(or not only) on peripheral conditions at the sense organ, but on more
central conditions, say at the first synapses with which the impulses
meet; in other words, difference of sensual quality may depend not
merely on functional differentiation of end organs and peripheral fibres,
but also on the mapping out of different patterns of response within
the nervous system.

§ 7. But it will be objected that neither all reflexes nor all
sensations obey the ' all or none ' principle ; and that what we have
been considering are exceptions to the general rule. The question
at once arises whether — inasmuch as muscle and nerve obey this
principle, inasmuch as the extensor thrust reflex follows it, and
inasmuch as the lowly protopathic system of sensibility shares it —
we are not justified in assuming that higher reflexes and higher forms
of sensibility are really governed by the same principle. Let us take,
for instance, the ' scratch ' reflex. Clearly this reflex, at first sight
at least, disobeys the 'all or none' principle. The strength of the reflex
can be graded according to the intensity of the stimulus. As the
stimulus increases in strength^ the clonic twitching movements of the
scratching leg increase in amplitude, force, and number, although their
frequency is practically unchanged. On the afferent side we may
conjecture that more and more nerve fibres are being stimulated ; on
the efferent side that more and more nerve and muscle fibres are being
stimulated. But what is going on at the spinal centres, that receive
the afferent and emit the efferent impulses? May we not suppose
that, shooting across the synapses, the disturbance preserves a constant
'type' of pattern of excitation, that pattern constituting a functional
unit for evoking the reflex? When the stimulus is increased, the

1 Cf. liivers and Head, "A Human Experiment in Nerve-Division," Brain, 1908, xxxi.
381.

Charles S. Myers 145

pattern of excitation changes. The additional afferent impulses shoot
across into new synapses, which become integrated with the old. The
old pattern becomes, so to speak, the nucleus of a more comprehensive
pattern. But, despite these additions, the ' type ' of the pattern is
unchanged. The pattern functions in precisely the same manner as
before, i.e. as a mechanism adapted for evoking the same kind of reflex.
The central neural pattern changes, therefore, not in intensity but in
extent; and it changes in extent in such a way that the type of reaction
movement is still maintained. That even in graded reflexes the ' all
or none ' principle is followed is shown by the observation that the
latent time taken to manifest the efferent effects of a sudden increase
in an original afferent stimulation is practically identical with the
latent time taken to produce the efferent effects of that original
stimulation ^

§ 8. Corresponding to the graded strengths of such reflex twitches
we have, I suggest, such graded strengths of sensation as occur among
auditory sensations. We can imagine that the reflexes caused by a
tone of constant pitch of changing intensity vary in much the same
way as the scratch reflex changes with strength of stimulus. The
number of nerve fibres carrying the incoming nervous impulses
increases with the growing strength of stimulation. Each nerve fibre
follows the ' all or none ' principle, while the central pattern, despite
its growth, preserves its functional reaction unaltered in aim. So
long as the pitch is kept constant, the type of reaction is practically
constant, and hence the sensation is practically unchanged in quality.
Further, the reaction, whatever its extent, preserves its character of
indivisible integrity; for which reason, perhaps, an intensity cannot be
analysed into component units of lesser intensity.

In certain respects, however, auditory sensations (indeed all forms
of sensations other than protopathic) are distinguished from such a
reaction as the scratch reflex, viz., by the apparent absence of refractory
periods, and by the virtual lack of fatigability. The movements of the
scratching limb are clonic ; they recur with a periodicity which is
practically fixed and independent of the frequency of the stimuli
applied to the afferent path of the reflex. An additional stimulus,
inserted with the object of evoking an additional reflex movement
interpolated between these movements, is ineffective. The reflex
centre thus shows refractory periods, during which stimuli are
inoperative. The reflex centre is also fatigable. But auditory
1 C. S. Sherrington, The Integrative Action of the Nervous System, London, 1906, 24.

146 The Intensity Differences of Sensation

sensations, on the other hand, are not intermittent, there appears
to be no refractor)? period ; nor are they sensibly fatigable. A pro-
longed sound is heard uninterruptedly ; and it may be heard for
minutes, days, or years, without appreciable alteration, so long as
we choose to listen to it\ The reasons for such continuity and
indefatigability we shall examine more closely in the following
paragraph. Having regard to their wide difference in complication,
we should not expect too close a parallelism between the reflex spinal
mechanisms and the far higher mechanisms concerned in sensation.
It might be objected, too, that what holds on the ' physiological ' side
need by no means hold on the ' mental * side of a reaction. But this
paper has been written from a diametrically opposite standpoint. Based
on the hypothesis of psycho-physiological parallelism, it aims at indi-
cating the light that may be thrown on the fundamental character of
the psychical attributes of sensation by physiological considerations^.

§ 9. There remain a third class of reflex, and corresponding to it a
third class of sensation, which we have yet to consider. If we may
describe the extensor thrust reflex as a ' gradeless ' reflex, and the
scratch reflex as a ' clonic ' (or ' twitch ') * graded ' reflex, we may term the
remaining form of reflex a ' tonic graded ' reflex. The reflexes, termed
by Sherrington ' proprioceptive ' reflexes, afford excellent instances of
this form of reflex. They are based on the integration of a pair of
antagonistic reflexes. The two centres of these reflexes are fed simul-
taneously by the same unceasing proprioceptive impulses, i.e. by impulses
arising from within the body, especially from the muscles, tendons, etc.,
of the limbs concerned. These afferent impulses act reflexly so as
to keep the muscles in a postural condition of perpetual tonus, — a
condition which may be described as a state of active equilibrium of
the double reflex. In this condition it rests as on a knife edge, from
which it may be made to swing in one or other of two opposite

^ If very weak, a continuous sound stimulus is only heard intermittently. (Into the
responsibility of inhibitory processes or refractory periods for such fluctuations I cannot
here enter.) If strong overtones are present, one or other of them may successively attract
the attention (cf. C. Stumpf, Tonpsychologie, i. 361).

^ A caution may be added against supposing that in the complete differentiation of
sensations the efferent side of central reactions must necessarily show itself visibly, by
altered muscular contractions. In the first place the efferent side of a reaction may lead
not only to skeletal but to visceral movements ; in the second place, not only to muscular
but to glandular and other activity; and in the third place, not directly to outgoing
peripheral activity at all, but to further central changes in virtue of the connexions
of the main efferent limb of the reflex arc with other arcs.

Charles S. Myers 147

directions, as between two poles. Broadly speaking, afferent impulses
which cause (or are set up by) reflex contraction of a group of muscles
governed by one centre of the double reflex simultaneously cause reflex
inhibition of contraction in the antagonistic group of muscles governed
by the other centre of the same reflex. It is the special role of such
reciprocal graded inhibition to procure an exact adjustment between
the strengths of incoming stimuli and outgoing discharges.

The tonicity of this class of reflex seems continuous and within certain
limits indefatigable. The continuity, the lack of refractory period,
probably depends on the phenomenon of after-discharge. The efferent
impulses are not cut short (or inhibited) periodically as in the step
reflex. They outlast the stimulus, and so, if the stimulus be repeated
sufficiently often, a continuous instead of an intermittent reflex
movement, a condition of 'posture,' results.

The tonic reflexes are especially prominent in the ' decerebrate '
animal ; the refractory periods, so characteristic of * decapitate ' and
' spinal ' preparations, are suppressed. Decerebrate rigidity is entirely
proprioceptive in origin ; it is abolished by ' deafferentation V and is
produced by the action of the proprioceptive impulses on a central
mechanism, situated above the spinal cord, these impulses abolishing
the refractory periods which are so characteristic in the ' spinal '
animal^.

§ 10. Corresponding to this class of tonic graded reflexes, we have
a class of tonic graded sensations, well exemplified (i) in the sensations
of warmth and coolness obtained from the epicritic system of cutaneous
sensibility and (ii) in the sensations of light obtained from the cones
of the retina. Adaptation, contrast and indefatigability are the dis-
tinguishing characteristics of this class of sensation. Just as the
extensor thrust reflex is fatigable, so we have seen the heat and cold
spot system to be fatigable; just as neither the extensor thrust nor the
scratch reflex can exhibit a condition of equilibrium, so neither the
heat and cold spot system of sensation', nor an auditory sensation (nor
perhaps the retinal ' rod ' sensations), is amenable to adaptation or
contrast. There is no middle or ' indifference ' point in heat, cold, or

^ I.e. cutting ofif all afferent impulses.

2 C. S. Sherrington, Proc. Roy. Soc. 1906, B. lxxvii.; Quart. J. of e.vp. Physiol. 1909 ;
J. of Physiol. 1910, xl.

3 Cf. Rivers and Head, op. cit. 406-410. In point of fact, the scratch reflex is not a
pure example of its class ; the scratching leg assumes a definite ' posture,' besides execut-
ing a series of scratching movements.

148 The Tnteimty Differences of Sensation

sound intensity as there is in warmth and coolness, or luminous
intensity. In place of fatigue we have, in the class of tonic graded
sensations, a variable neutral point of adaptability, a ' tonic ' sensation,
which is neither warm nor cold, neither bright nor dark, from which it
is possible to change in the direction of warmth or coolness, brightness
or darkness, and to reach (within certain limits) a new state of
adaptation. In place of two isolated mechanisms for heat and cold, we
have an integration of the two ' incopresentable ' polarities for warmth
and coolness, within a single mechanism \

§ 11. In the case of the class of tonic graded sensations, it is far
more difficult to discriminate between changes in intensity and
changes in quality. Who shall say whether the gradations of the
warm-cool (or white-black) series of sensations change in intensity or
in quality?

Nor is the confusion merely between intensity and quality. With
the class of tonic graded sensations a new element enters, that of
' extensity.' One sensation appears to have more extensity than
another ; it appears to come from a wider sensory area. But, as on
the spinal reflex level, if the same stimulus is applied to a wider
sensory area the sensation must alter at the same time in intensity
and often in quality. A vessel of warm water, in which the whole hand is
immersed, feels warmer than when a single finger is introduced. The
intensity and hue of a luminous sensation depend not merely on the
strength of the stimulus at the point of stimulation but also on the
area, the number of points, of application of that stimulus.

So close a connexion between extensity and intensity has been
supposed to be correlated anatomically with the mode of termination

^ How sensation, like a reflex, leaps from plane to plane has been already exemplified
(p. 143) in the passage from heat, cold and touch to pain. The plane of the protopathic
system is in turn distinguished, as we have just mentioned, from that of the epicritic
system of cutaneous sensibility by its gradation, by its power of adaptation, by its (relative)
indefatigability, and, further, by its freedom from diffusion and radiation and by its power
of accurate localisation and of estimating relative size. So, too, the vibrations of a
tuning-fork applied to the skin (a fortiori when applied to the ear) evoke a central reaction
different from a mere series of touches. Thus Head and Holmes record a case of
Brown-Sequard paralysis in which " the vibrations of a tuning-fork were appreciated
badly or not at all over the right arm and leg, in spite of the complete integrity of tactile
and pressure sensibility" {Brain, 1911-12, xxxiv. 111). So, too, a flickering light, as soon
as the rapidity of the flicker is so great as to produce an uninterrupted sensation,
establishes a new sensation, behaving, as Sherrington has shown (this Journal, 1904, i.),
very differently from a series of flicker sensations ; that is to say, a new ' type' of central
neural pattern is at once initiated.

Charles S. Myers 149

of afferent nerve fibres at the periphery. Any given fibre divides
into fibrils and supplies a relatively wide sensory area, which is also
supplied by similar divisions of other neighbouring fibres on either
side of the given fibre. Hence it has been thought that extending the
area of stimulation involves stimulation of a greater number of fibrils
of the same fibre, and so leads to increase in the intensity of the
nervous impulse along the nerve fibre. But if, as we have seen reasons
for believing, nerve fibres act according to the * all or none' principle,
this explanation of the connexion between extensifcy and intensity falls
to the ground, and we must seek some other — preferably in the close
similarity of reaction, i.e. in the partial identity of neural pattern,
which a more intensive stimulus and a more extensive stimulus reflexly
call forth. Both the extensive and intensive series affect a greater
number of nerve fibres as they are increased ; both therefore call forth
at first almost the same central changes of neural pattern, and almost
the same efferent changes in outward reaction.

§ 12. We have seen that, so long as the reaction is of the same
* type,' pursuing the same plan and purpose, the corresponding sensa-
tion alters only in intensity ; but that, when it begins to alter in type,
differences of quality make their appearance. It is not surprising,
then, that when sensations are made to increase in intensity, they so
often show more than changes in mere intensity. Indeed it is probable
that no sensation can be increased in intensity without at the same
time undergoing some change in quality.

But this consideration alone hardly helps us to understand why the
warm-cool and bright-dark series of sensations approach so much closer to
changes in quality than do those belonging to the other two systems of
sensation we have considered. We have also to take into consideration
the manner in which the qualities of a given stimulus are differentiated
in the evolution of sensibility. We start, I think we may assume, with
a vague 'whole' of sensibility, which is differentiated into an increasing
number of constituent ' parts.' Auditory stimuli of different vibration-
frequencies at first give rise to a vague appreciation, merely perhaps of
high and low pitch; we come ultimately, by superadded reactions, to
have an enormous number of different types of reaction corresponding
to different vibration-frequencies of stimulation ^

In the case of colour sensations, it is, I think, generally agreed that
they have been differentiated from the colourless series of sensations.

1 Cf. H. J. Watt, this Journal, 1911, iv. 146.

150 The Intensity Differences of Sensation

But even prior to such differentiation, I venture to conceive a time
when the various colour stimuli produced vague sensations of colour
and of white (evolved from tHe still more primitive phenomena of
taxis) by their action on three different neural centres. These centres,
each with its appropriate pattern, correspond respectively to red,
blue and yellowish-green sensations when excited individually, and
to orange, yellow, purple, etc., or to white when two or more of them
are excited simultaneously. At such a stage we may suppose a fairly
close correspondence between visual and auditory sensibility. The
reactions of both are due to the excitation of a group of single centres.
Any one neural pattern is variable in a single direction only, the extent
of variation corresponding to the degree of sensual intensity.

In this system of three-colour sensibility sensual contrast, adapta-
tion and brightness are unknown \ Upon it a more complex system of
sensibility, giving rise doubtless to new and 'higher' sensual qualities,
has been erected. All colour stimuli come now to act on double centres
corresponding to reflexes of the tonic graded type and giving rise,
in the first place, to a series of sensations of graded brightness.
Brightness replaces intensity now that the reaction swings from pole
to pole, instead of, as before, from zero to the maximum of reaction.
The corresponding sensations vary from white through grey to black
and are characterized by contrast, adaptation and lack of fatigability.

From within this paired system of colourless sensibility two similar
paired systems of colour sensibility become differentiated. The
yellowish-green component of the lower three-colour system is now
divided into yellow and green elements, and the red and green form
the one, while the yellow and blue form the other, of the two new
paired centres. It is as if from an axis of up- and down-reactions (corre-
sponding to the white-black series of sensations) there had arisen two
opposite to- and fro-reactions, each in a plane at right angles to the
original up- and down-reaction (one to- and fro-reaction corresponding
to the red-green, the other to the yellow^-blue series of sensations).
We must suppose that each colour stimulus acts not only on the white-
black but also on the red-green or yellow -blue system (or on both
systems) of sensibility. We must further suppose that when either of
these two colour systems of sensibility is in equilibrium there is

1 Of course contrast (and adaptation) of a sort are never absent ; there is a contrast
effect in a heat-foUowed-by-cold experience, or in a noise-followed-by-whisper experience.
But such instances differ not in degree but in level from the sensual contrast (and adapta-
tion) we are here considering.

Charles S. Myers 151

no sensation save the grey which results from the ever simultaneous
action of the colour stimulus on the white-black system.

Thus we are able to start from a (variable) neutral balancing point
of equilibrium, a state of tone, from which we may progress in one
direction or the other, from any grey either towards red, green, yellow
or blue (or any intermediate shade of colour). Or, perhaps largely in
virtue of the more primitive three-colour system, we may start from
red and proceed to green by orange and yellow, or by carmine and
violet. This mode of evolution of the colourless and colour series of
sensation may help us to understand our difficulty in determining
whether gradations in either series represent variations in quality or
intensity. Quality and intensity appear to meet here ; it becomes as
difficult to be certain that, as we pass from red through shades of
orange to yellow, we are passing through changes not of intensity but
of quality, as it is difficult to be certain that, as we pass from black
through shades of grey to white, we are passing through changes not
of quality but of intensity. The same holds good for the evolution of
our sensibility to warmth or coolness, although here (and possibly also
in vision) the co-existence or the precedence of a more primitive ' spot '
system of varying sensibility may have been psychically helpful. As
Sherrington has said of the corresponding reflexes, — propriospinal reflexes
normally fuse with other reflexes as adjuvant to them.

The close connexion we have traced between the colourless and the
colour series of sensations is also shown in our ability to express the
latter in terms of the former, i.e. to estimate in terms of greyness the
brightness of a colour sensation. The intensities, on the other hand, of
two very different colours we can only vaguely compare, either by
their physiological action (so we may vaguely compare the intensities of
two tones of very high and very low pitch), or by objectively imagining
what amounts of colour there are in the respective stimuli (so we may
vaguely compare the intensities of two different qualities of gustatory
or olfactory sensation).

Our neurological considerations have led us to trace our experience
of intensity back to a stage when it was possible only by a comparison
of experiences corresponding to the neural patterns of two or more
quite separate sense apparatus. At this first stage there were in
question but two opposite qualities, two types of reaction, — one, e.g.,
a heat reaction obtainable from scattered heat spots, the other a cold
reaction obtainable from scattered cold spots, of varying sensitivity.

I

152 The Intensity Differences of Sensation

At this stage the gap between heat and cold sensations was not
bridged.

In the second stage of sensibility, we find both quality and intensity
far more advanced. Not only can intensity within a given sense
apparatus be delicately graded, but also quality admits of further
gradation by the simultaneous excitation of two or more elementary
reactions in different degrees. Thus, given two colour qualities red
and yellow, every shade of intermediate (juality is obtainable by
simultaneously presenting the two stimuli in different strengths.

In the third or last stage, quality is still further developed ; it rises
into ' modality.' It is no longer possible to pass from one quality
{e.g. bitter, red, or warm) to another {e.g. sweet, green, or cool,
respectively) without passing through a quality which belongs to
neither. Such antagonistic qualities, 'modalities' as Helmholtz called
them, are incopreseutable ; when simultaneously excited, the}" either
* neutralise ' one another (giving rise to a new sensation) or show
rivalry. What remains of intensity at this stage is a relic from the
second stage in the evolution of sensibility, when we are dealing
with one (or more) single centres in action, each starting from zero ;
whereas at this third stage we have one (or more) double centres in
action, inclining to one or other side of a variable point of active
equilibrium.

III.

The conclusion, to which we have been led, is that the ultimate
difference between the quality and the intensity of sensation depends
on the nature of the underlying reaction. Broadly speaking, when the
reaction changes its fundamental type, it alters in quality and the
sensation changes also in quality. So long as the reaction preserves
its fundamental type, it can be said to vary only in quantity, and the
sensation changes also in intensity.

In this sense sensual intensities are quantitative. But intensities
are not quantitative in the sense that there is a moreness or lessness of
excitation within the same anatomical area; for we have seen reason to
believe that any given neural tissue, central or peripheral, follows the
' all or none ' principle. Nor are intensities quantitative in the sense
that the stronger sensation contains the weaker, as a large quantity
may be said to contain a smaller ; this is obviously contradicted by
introspective experience. But in so far as they fail to answer to these
requirements, it does not follow that sensual intensities are qualitative,

Charles S. Myers 153

unless we merely mean that they resemble sensual qualities in behaving
differently from quantities (which after all are not in the same sense
sensual).

We cannot even say that intensities are quantitative, merely in the
sense, and for the reason, that they show a moreness or lessness of the
same sensory character. Otherwise we must term the differences in
any series of colour sensations, ranging say from red to yellow,
differences of intensity. We can speak of colour sensations as differing
in moreness or lessness of hue ; we can describe olfactory or auditory
sensations as having more or less of a certain quality. But we do not
in general confuse such variations in quality with variations in intensity,
of the sensation.

Sensual intensity thus means, at bottom, how much reaction ;
sensual quality means what type of reaction. That is to say (if only
we can legitimately translate neurological into psychological process)
intensity and quality of sensation are dependent on variations of
mental activity in one or other of two directions\

Now when the present activities of higher systems look back on the
past activities of higher systems, we get a differentiation of the activity
consciousness into consciousness of mental ' processes ' or ' acts,' on the
one hand, and of mental ' products ' or * contents ' on the other. If
the lowly sensual levels could but look back upon themselves, their
activities would be similarly differentiated into ' acts ' of sensing and
' contents ' of sensation. But this is not possible. We experience
' acts ' of apperception, thinking, willing, imagining, etc., in all of
which the self is involved ; but we have no experience of the ' act ' of
sensing. It is true that intensity and quality are derived from the
activity of the sensual level. But they come to us as contents of
consciousness, not merely by virtue of such lowly activity (itself
undifferentiated into act and content), but also by virtue of our
possession of higher and still higher forms of experience culminating
in comparison, relation and abstraction. We are able to compare
and to relate individual reactions differing in 'moreness' or 'less-
ness ' or in ' type ' ; and thus finally we reacli the abstract forms (the
Gestaltvorstellungen) of intensity, quality and quantity generally'^.

^ That these directions are fundamentally different is shown by the limitation of
Weber's law to intensities.

^ This is well shown in certain parietal cortical lesions, where our powers of unconscious
comparison, relation and attention seem at fault, and in consequence not only the
differential, but also the absolute, thresholds of sensations (as judged by the uniformity of
correct answers) suffer, although sensibility itself may be relatively little affected (H. Head
and G. Holmes, Brain, 1911-12, xxxiv. 102-254).

J. of Psych. VI 11

164 The Intensity Differences of Sensation

If pressed to give a definite answer to the question whether
intensity differences are quantitative or qualitative, I decline to be
bound to either horn of the dilemma. To me the question seems
mal posee, since it is capable of every conceivable answer. Thus
it is open for anyone to call a change in sensual intensity qualitative, —
in so far as it is not directly measurable ; in this sense, a change in any
other sensual attribute (in quality, extensity or protensity) is also
qualitative. Or a change in sensual intensity may as legitimately be
called quantitative, — in so far as (even like changes in sensual quality
as contrasted with sensual modality) it depends on ' more or less ' of a
given complex of reaction, and in so far as (even like changes in sensual
quality or modality) it is indirectly capable of measurement in spatial
and temporal terms based on sensual extensity, sensual protensity and
movement. On the other hand, since changes in sensual intensity
occur in a specific direction quite different from that of quality
changes, they may be called non-qualitative. Indeed it is thus
arguable that intensity differences are neither qualitative nor quanti-
tative, but strictly sui gene7'is, i.e. intensive.

(Manuscript received 20 March, 1913.)

ARE THE INTENSITY DIFFERENCES OF
SENSATION QUANTITATIVE?^ 11.

By G. DAWES HICKS.

§ 1. The relation of qualitative and quantitative.

§ 2. The physiological correlate oj differences in quality and
quantity of sensation.

§ 3. Some general points of psychological theory.

§ 4. Psychological criticism of Dr Myers's hypothesis.

§ 5. Bergson's explanation of the reason why we regard sense con-
tents as quantities^ and a criticism of that explanation.

§ 6. Meinong's account of intensive quantity.

§ 7. Meinong's interpretation of Weber's Law, and criticism thereof.

§ 8. Is quantitative comparison always an act of judgment ?

§ 9. Differences of intensity may he said to he magnitudes hut not
qu4xntities.

§ 1. With many of the concluding observations in the preceding
paper I am in accord. I agree that the question whether intensities
are qualitative or quantitative is badly framed, and admits of no
intelligible answer. "There is," as Mr Bradley puts it, "no such
thing as quantity merely extensive, or as quantitative differences with-
out quality. Because anything is qualitative, that is no reason why
it should not also have quantity ^" The quantitative, in other words, is
but an abstract aspect of what we are actually dealing with even in that
region where its importance and significance are indisputable, and
quantitative explanation has always the perfection and the imperfection
which attach to abstract treatment. Whatever be its nature, the
material world can never find complete expression in quantitative
terms. It has, and is bound to have, its own structure or collocation
of parts. Though we divest the parts of every shred of qualitative
distinctness, though we reduce them to what can be satisfactorily

1 A contribution to the Symposium presented at the Joint Meeting of the British
Psychological Society, the Aristotelian Society, and the Mind Association, in London,
7 June, 1913. 2 jifi^d, N.S. iv. 5.

11—2

156 The Intensity Differences of Sensation

rendered in quantitative formulae, yet, however far we push this method
of procedure, there will always remain at least one qualitative element
that cannot be eliminated, — the particular kind of distribution which
holds good at any given moment. Any such absolute separation as
that which Munsterberg' would institute between the physical and the
psychical, quality being excluded from the one no less rigidly than
quantity is excluded from the other, is bound, therefore, to lead to
a thoroughly false conception of the way in which the two disparates
are related the one to the other. I agree, also, with the remark made
in an earlier part of the paper (II, § 12), although I think it not a little
damaging to the author's main thesis, that probably a sense-presentation
can seldom be increased in intensity without at the same time under-
going some change in quality. Particularly in visual apprehension,
the sense-experiences we have to deal with are so complex, the factors
implicated are so numerous, that it is hardly conceivable there should
ever be changes in intensity only with qualitative constancy.

Further, though the considerations which Dr Myers has adduced
should appear inadequate to bear the weight of the hypothesis he
would base upon them, few will question the value and interest of those
considerations in themselves. Whatever else Munsterberg succeeded
in doing in the third of his Beitrdge, he certainly managed to make
manifest the important function fulfilled by kinaesthetic factors in the
development of sense-experience, and Dr Myers has accomplished a
most useful piece of work in the same direction by showing how from
the physiological side those factors call for recognition. He has made
it abundantly evident that the prevailing tendency to interpret sensa-
tion as though it were a purely cognitive process is a mistaken tendency,
and that the complexity of the total process, as involving not merely
awareness of a content, but a change in the state of feeling, and
consequently specific forms of movement, requires to have more justice
done to it.

§ 2. I am not, however, convinced that a case has been made out,
even from the physiological side, for supposing that " the ultimate
difference between the quality and the intensity of sensation depends
on the nature of the underlying reaction." Dr Myers, it will be
noticed, passes over with very scant mention the peculiar structure of
the end organs and the mode of their development. With "the
mapping out of different patterns of response within the nervous
system," he does, it is true, in one place (II, § 6) couple the " functional

^ Grundziige der Psychologies i. 2QQsqq.

G. Dawes Hicks 157

differentiation of end organs and peripheral fibres " as a condition upon
which the difference of sense quality may in part depend, but in the
rest of his investigation the latter seem to disappear from view and
exclusive prominence to be given to the former. He is in no way
wishful, I take it, to minimise the significance of the elementary fact
that the physical stimuli themselves are different, that, for example,
ether waves and air waves are essentially dissimilar, nor to suggest
that the gradual differentiation of the sensory mechanism has not been
largely determined by these external dissimilarities. But what I miss
in his treatment is any indication of the way in which he conceives the
two factors, (a) the mapping out of different patterns of response
within the nervous system, and (6) the functional differentiation of end
organs and peripheral fibres, to be connected with each other. The
argument certainly does seem to require that the latter be placed in a
position of such decided dependence upon the former as is hardly
reconcilable with the biological evidence. The broad general fact
that all the organs of special sensation are originally derived from the
ectoderm would seem to point in the opposite direction, and the
rudimentary stages of (say) the invertebrate eye, — the gradual forma-
tion, namely, of groups of pigmented cells, — appear to be naturally
explicable from the action of the physical stimulus upon the epithelium.
Moreover, I note that a significant change in the presentation of the
case with reference to the ' all or none ' principle occurs when Dr Myers
advances from muscle and nerve to sensation. In respect to the former,
the rule is laid down that with a weak stimulus, only a few elements or
fibres respond, whilst with a stronger stimulus, other elements or
fibres are also implicated. But in respect to thermal sensations,
allowance has to be made for the fact that the heat and cold spots are
relatively sensitive and relatively insensitive, so that " presumably
these reflexly produce relatively considerable and relatively weak
reactions," and " we may suppose that the nervous impulse from a
more sensitive heat spot spreads centrally and hence efferently to a
greater number of nerve fibres than are reached by the stimulus of
a less sensitive heat spot" (II, § 5). That is to say, when we take not
any one heat (or cold) spot, but any group of such spots, there is
a gradation in the intensity of the corresponding sensations, only the
important factor then is not increase or decrease in the strength of the
stimulus, but the relative sensitiveness or insensitiveness of the end
organs concerned. In other words, there has to be recognised at the
very outset, so far as sensations are concerned, the essential dependence

158 The Intensity Differences o/ Sensation

of their intensity upon peripheral conditions. Equally so is it in
respect to their quality. Dr Myers argues (ibid.) that just as with
excessive increase in the strength of the stimulus, the extensor reflex
suddenly gives place to the flexor reflex, so with like changes in the
thermal stimulus, the sensation alters in quality and becomes a sensa-
tion of pain. Curiously enough, however, he passes over the question,
which is surely relevant to his argument, — " whether the same end
organs can give rise to pain and, with weaker strengths of stimuli,
to heat, cold and touch." If different end organs come here into play,
and there seem strong reasons for thinking this to be probable, the
parallelism which it is sought to establish with what happens in the
case of the extensor thrust reflex so far breaks down.

§ 3. The criticism, however, which I am concerned to press is
criticism of a psychological kind. At the outset, I had better perhaps
refer to certain general points of psychological theory in regard to
which, if I understand him rightly, I should dissent from Dr Myers.
For the sake of brevity, I group them under the following five
heads : —

(a) We are at one as to the necessity of distinguishing in what is
called a ' state of consciousness,' the act or process of consciousness
from the content of consciousness. But this distinction I take to be a
distinction of aspects, and not a distinction of two existences. Act
and content are not, I should say, what, for example, the Herbartian
psychologists inclined to conceive them as being, two independent and
separate entities. The distinction does not, therefore, seem to me to be
accurately described as involving "two main divisions of consciousness,
— the consciousness of ' acts ' or ' processes ' and the consciousness of
'contents' or 'products.'" On the one hand, 'acts of consciousness'
'are not necessarily ' the consciousness of acts.' Constituting as they
do the very life of the conscious subject, they are rather ways in
or through which that subject is aware than objects of which he is
aware. On the other hand, the content, although fairly enough in
one sense called a product, is not necessarily a product in the sense
of being wholly a construction on the part of the mind itself. And it
may be, I think it is, an error to give to the content the position of an
independent object, and to picture the act of consciousness as a sort of
inner vision directed upon it.

(6) Although the duality of act and content be involved in the
most rudimentary phases of the mental life, it does not by any means
follow that even the crudest recognition of that duality is a primitive or

G. Da\ves Hicks 159

primordial fact of miud. To differentiate, for example, hearing from
the sound heard implies, I should say, an experience far more com-
plicated and the use of concepts far more abstract than we can ascribe
to the animal consciousness. But this does not, I think, constitute a
fundamental severance between sense and the other cognitive processes,
as Dr Myers seems to suggest. I cannot discover any ground for
saying that in the mature mental life there is no recognition of the
distinction between the act of sensing and the content sensed, any
more than I can find ground for saying that in the less mature mental
life there can be no imagining or apperception before the corresponding
recognition is reached.

(c) Those who lay stress upon the part played by sensations of
muscular strain or tension in the so-called ' feeling of effort ' do not
necessarily deny that consciousness is an activity, nor even that there
may not be awareness of such activity. What they do call in question
is the legitimacy of identifying that activity with the activity of which
there is awareness in the ' feeling of effort.' Certainly to me it seems
doubtful whether we can ever be said to be directly aware of the
activity involved in consciousness itself, — aware of it, I mean, after the
manner in which we are directly aware of a colour when we actually
see it or of a sound when we actually hear it. But, in any case,
I should be prepared to maintain that the ' feeling of effort ' evinces
itself as the outcome of a long repetition of experiences, and as having
acquired an appearance of simplicity which disguises from us its really
complicated character.

(d) It is not precisely clear what Dr Myers means by the
possibility of " a choice on the part of the organism between two or
more reactions to a given stimulus " as the condition of the appearance
of consciousness (II, § 1). But if he intends to suggest that conscious
choice, though it may be of the crudest kind, is an essential factor of
mental life, I should hesitate in following him. It cannot be sup-
posed that the consequences of specific reactions are in any way
prefigured or represented in anticipation by the primitive consciousness.
And until that in some vague way comes about, I do not see what
conscious choice can mean.

(e) I do not think we are entitled to lay down as a psychological
presupposition that " differences in type of movement must be the
cause of differentiation in the quality of sensation," on the ground
that *• it would be of no advantage for the organism to experience
different qualities of sensation, unless those differences were serviceable

160 The Intensity Differences of Sensation

in promoting different types of response." With the appearance of
conscious life, advantageousness for the organism may not be the sole
principle determining the course of evolution and may become less and
less so as mental evolution proceeds.

§ 4. I return then, now, to the hypothesis propounded by Dr Myers.
According to that hypothesis, if I interpret it correctly, the deter-
mining factors throughout, in regard both to the quality and the
intensity of sensation, are the nervous responses and the efferent
reactions which ensue upon the respective stimuli. The lower and the
higher forms of sensibility, not less than the lower and the higher kinds
of reflexes, are in reality governed by the ' all or none ' principle.
Within certain limits the type of reaction remains the same, and
the grading of the latter in strength according to the strength of
the stimulus (when it is so graded) means that as the strength of the
stimulus increases more and more nerve and muscle fibres are called
into play, the corresponding sensation meanwhile altering only in
intensity. If, however, those limits be exceeded, if the increase in
strength of the stimulus be excessive, the type of reaction changes, and
a difference of quality makes its appearance in the corresponding
sensation. Apart from the admittedly conjectural character of most
of the essential steps of the argument, the objections I feel inclined to
urge are these. In the first place, it is not easy to gather from
Dr Myers's exposition in what relation he conceives the sensation to
stand to the reaction. Which is the antecedent, and which the con-
sequent? The whole trend of the line of thought pursued would seem
to imply that when it is said that " the ultimate difference between the
quality and the intensity of sensation depends on the nature of the
underlying reaction," the dependence intended is a chronological
dependence, and that in the order of sequence the sensation follows
the reaction. But can it be maintained that this is in accordance with
fact? And if it can, what becomes, on such a supposition, of the
contention that "it would be of no advantage for the organism to
experience different qualities of sensation, unless those differences were
serviceable in promoting different types of response" (II, § 1) ? If, on
the other hand, the sensation precedes the reaction ; if one sensation
" leads to " one type of reaction, and another sensation to another type,
then how can we " attribute the intensities of sensation to different
degrees of the same reaction " (II, § 2), how can the circumstance that
a sensation is unchanged in quality be due to the fact that the type
of reaction is constant (II, § 8)? In short, how is the conception

G. Dawes Hicks 161

of different qualities of sensation promoting different types of response
reconcilable with the doctrine of " psycho-physiological parallelism,"
upon which Dr Myers tells us his view is based? In the second
place, I fail to see in what way the hypothesis helps towards the
solution of the problem which is the subject of this discussion.
Suppose it be true that the correlate of differences in quality of a
sensation consists in differences in type of reaction, and that the
correlate of differences in intensity of a sensation consists in different
degrees of the same reaction. Still, even then, the reaction forms no
part of the content of the corresponding sensation, and for the
experiencing subject no comparison of this content with the reaction
is any more possible than with the external stimulus. Howsoever it
may be for the scientific reflexion of the physiologist, for the
experiencing subject sensual intensity most assuredly does not mean
how much reaction, nor does sensual quality mean what type of
reaction. There is nothing in the redness of red or the blueness of
blue or in the sound of a tone heard to suggest to the experiencing
subject specific types of nervous and muscular reaction, nor is there
anything in the varying intensity of a colour, or in the faintness or
loudness of a note, to suggest to him a greater or less number of efferent
nerve and muscle fibres in a state of stimulation. And the question as
to the intensity of sensations has to do, I presume, primarily with
sense contents, — whether, namely, they warrant any definite assertions
as to difference of amount, or whether what are taken to be differences
of amount do not, in truth, indicate solely qualitative changes. I do
not find that the answer to that question is materially furthered by
correlating sense presentations with types and degrees of response
within the nervous system rather than with kinds and amounts of
external stimulation. When worked out from the psychological side,
Dr Myers's theory would, I imagine, lead him, in the long run, to a
conclusion resembling that of Miinsterberg in the Beitrdge, according
to which muscular sensations, or sensations of strain {Spannungsemp-
findungen), mediate as a tertium quid between the physical and the
psychical. Sensations proper, in MUnsterberg's view, vary only in
quality, and what is usually called a more intense sensation is, in
all cases, a content of consciousness numerically and qualitatively
distinct from the weaker sensation with which it is compared. Muscular
sensations, on the other hand, occupy a unique position in the mental
life ; although sensations, they share with physical entities the.
characteristic of varying only in quantity and not in quality. Since

162 The Intensity Differences of Se7iscUio7i

any physical stimulus necessarily evokes a change in muscular strain or
tension, the equivalent of which in consciousness is a Spannimgsemp-
findung, all sensations proper have as their accompaniments muscular
sensations, and these afford the basis of measuring differences of
intensity. Some such mode of translation into psychological terms the
theory before us would appear to demand, and Dr Myers himself almost
hints as much when he insists at the beginning, that " the very word
intensity means a state of tension or strain " (I, § 3) — a remark, by the
way, which seems to conflict with the concluding suggestion that
" intensity differences are neither qualitative nor quantitative, but
strictly sui generis." Ingenious, however, as Miinsterberg's treatment
of the problem undoubtedly is, it raises, I think, more perplexities than
it succeeds in removing. The peculiar character ascribed to muscular
sensations, as toto genere distinct from the character of all other sen-
sations, is eminently unsatisfactory, and no serious attempt is made to
show how it comes about that in comparing the intensities (say) of two
sounds, our judgments are directed upon the sense-presentations them-
selves, and not upon their assumed concomitants \ Perplexities of a
like kind would, I am persuaded, confront Dr Myers the moment he
attempted to specify the psychological equivalents of the physiological
factors which he takes to be involved in the differentiation of sensual
quality and intensity. I am far from wishing to dispute the contention
that kinaesthetic sensations are implicated, in some form or another, in
every mode of sense experience. I think it likely enough that they
are. But, after all, sensations of tension and strain have their own
content, and from that content to the neural responses and muscle
reactions themselves is a far cry. It seems to me that we have here a
problem thrust upon us precisely the same in character as that which
presented itself when the correlatives were taken to be the external
stimulus and the sensation.

§ o. Somewhat similar obstacles beset the path of those who, like
Bergson, maintain that the contents of mental states cannot rightly be

1 The stress of these and allied difSculties probably occasioned the change of attitude
observable in the Grundziige. At all events, in the later work, Miinsterberg disputes
apparently the possibility of any, even an indirect, measurement of sensual intensity, and
what he has to say about muscular sensations deviates markedly from his earlier mode of
dealing with them. See Grundziige, i. 263 sqq. Note, especially the remark on p. 280,
" Spannung und Streben bedeutet also Kraft fiir die vorpsychologische Wirklichkeit und
fiir die empirische psychophysiche Personlichkeit, im System der psychologischen Be-
wusstseinsinhalte bedeuten sie dagegen nur eine Erfahrung und stehen dem Probleme der
messbaren Wirkungen nicht naher als die Empfindungen blau und tonend und sauer."

G. Dawes Hicks 163

treated as magnitudes, that the relations of greater and less are not
properly applicable to them. Bergson admits that ordinarily we do,
without the slightest hesitation, pass judgments involving quantitative
comparison upon the contents of our experience. He has, then, to
explain how it comes about that into the field of what is purely quali-
tative the appearance of intensive magnitude intrudes and creates the
illusion of progressive increase and decrease. The explanation is ob-
tained by tracing back the appearance to the natural propensity of
consciousness to objectify mental states, to regard them, that is to say,
as extensive quanta. In the case of what he calls representative states,
we transfer he thinks the idea of the cause which is quantitative into
the effect which is purely qualitative, and the notion of intensive
magnitude is only a "perception acquise"; in the case of affective
states, we give the name of intensity to the larger or smaller number
of sensations which we associate with the fundamental sensation, and
the notion of intensive magnitude is here a " perception confuse."
When, for example, we experience a pain which becomes, as we say,
more and more acute, consciousness distinguishes, within the charac-
teristic sensation which gives the tone to all the others, a larger or
smaller number of sensations arising at different points of the periphery,
muscular contractions, organic movements of various kinds,, and the
totality of these elementary psychical states expresses the new exigencies
of the organism in presence of a new situation thus constituted for it.
We estimate the intensity of the pain by connecting the differences of
sensation with the reactions which usually accompany them, and which
are more or less extended ; by prefiguring, that is to say, the future
bodily movements in the very midst of the sensation which is being
experienced. When, again, we estimate quantitatively the loudness of
a sound, we take into account not merely the change or disturbance in
the vital condition of the organism, but also the fact that, by striking
some object and thus expending a definite quantity of effort, or by
exerting ourselves in the use of our vocal organ, we have repeatedly
produced a similar sound; and the idea of this effort immediately
presents itself when we transform the intensity of the sound into a
magnitude \

There are, indeed, certain portions of Bergson's analysis which seem

to me to be entirely on the right lines. In dealing, for instance, with

the so-called " sense of effort," he complains, not without reason, of the

crudeness of the conception of " a psychic force imprisoned in the mind

^ Les donnies immediates de la conscience, Ch, i.

104 The hUensity Differences of Sensation

like the winds in the cave of Aeolus, and only waiting for an opportunity
to burst forth," and of the will as watching over this force, and from
time to time opening a passage for it. The considerations which he
urges in favour of regarding experienced "effort" or "activity" as a
content of consciousness, and not as itself identical with the activity of
consciousness, although perhaps not in themselves sufficient to establish
this conclusion, can, when reinforced by others which he might have
used, be formed into a coherent body of evidence which it would be
extremely hard to resist. To isolate the act of apprehending from the
content apprehended, and to attribute to the latter a strength or
intensity of its own, which may vary independently, is a procedure for
which no justification is yielded, so far as I can discover, from psycho-
logical analysis. At all events, we are, I think, entitled to say that the
differences of intensity which we discriminate in the content appre-
hended are not to be regarded as equivalent to a greater or less amount
of apprehending activity. We by no means of necessity apprehend the
more intense better or more accurately than we apprehend the less
intense. We are more liable to overlook changes of loudness in the
roar of a cannon shot than those in the buzzing of a bee. Leaving on
one side the thorny issue whether, as it is misleadingly stated, the
intensity of a sense content may be increased by attention, I would
only insist that, in any case, such a definition of attention as " the
variously related degrees of psychic energy expended upon the different
aspects, elements, and objects, in the one field of consciousness'"
prejudges at the start the fundamental question which it is the very
business of psychological investigation to decide. For the increase of
clearness and definiteness which results from attention may depend not
upon a " focussing of psychical energy," — a conception which we shall
try in vain to render intelligible, — but upon the number and kinds of
discriminations we are able to make in the content attended to, the
distinguishable marks we are able to recognise in it, — a process which
would consist largely in connecting the said content with, in relating it
to, representations and ideas already possessed by the apprehending
subjects Clearness, certainly, is one thing, and intensity another, but
if through attending the content becomes more intense, an explanation
of that circumstance must be found that is consistent with the ex-
planation we are enabled to give of the increase in clearness and
distinctness.

^ Ladd, Psychology, Descriptive and Explanatory, 74-5.

^ Of. my paper on " TheNature and Development of AttentioQ," in this Journal, 1913, vi. 1.

G. Dawes Hicks 165

Bergson's arguments relate chiefly, however, not to the processes of
consciousness but to the contents apprehended thereby, and in this
reference the explanation he has to oflfer seems to me to fail. The
failure evinces itself, I think, in much the same manner whichever be the
department of experience with which he is dealing. Take, for instance,
his account of the, way in which we come to regard a pain as increasing
in intensity. He recognises that it will not do to say that the more
intense pain corresponds to a greater nervous disturbance, for these
disturbances are as movements unconscious, and their equivalent in
consciousness has no resemblance whatsoever to motion. But, he con-
tends, the automatic movements which tend to follow the stimulus are
likely to be conscious as movements, and the differences of sensation
are interpreted by us as differences of quantity because we connect
them with the reactions which usually accompany them and which are
more or less extended. Now, the obvious question which at once
presents itself is, why should the movements that accompany the
sensation be said to be unconscious, and those that follow it be said to
be conscious, as movements ? It becomes very soon apparent that the
latter way of speaking is no more than a metaphor. By movements
that are conscious as movements Bergson here means simply the sum
of sensations that arise from muscular contractions, organic conditions,
changes in the state of joints, tendons and skin, and so on. In other
words, the apprehension of these movements as movements is just as
distinct from the fact of movement itself as are all presentations from
the objective events giving rise to them. The factors called in to account
for the appearance of intensive magnitude ought, ex hypothesi, then,
to be as little capable of yielding it as the sensation of pain itself.
Moreover, the theory is not, I think, confirmed by the appeal which is
made to experience. When I become aware that a tooth-ache from which
I am suffering is becoming more acute, or that a headache is becoming
more severe, I fail to detect even by the most careful introspection any
reference at all to the "thousand different actions " I might take in order
to avoid either of these calamities. What introspection does seem to
testify is that the estimate of intensity is derived directly from the
experience of the pain itself. Now, I quite admit that introspection may
in this respect be deceptive, but if it is, the deception stands in need of
explanation, and the difficulty of finding one along the lines that have
been followed seems well-nigh insuperable. Finally, the criticism I am
urging may be summed up in more general terms. True though it
may be, that in mature experience our judgments as to the comparative

166 The Intensity Differences of Sensation

intensity of two sense contents are constantly aided by the knowledge
we possess of the physical world, yet it is impossible to suppose that
such judgments could have become possible, if ultimately such sense
contents are never directly apprehensible as standing to one another
in a relation of greater and less. What problematical knowledge of
the cause of olfactory sensations could conceivably have originally
induced us to pronounce one smell to be stronger than another, if the
contents of those experiences did not themselves furnish the data for
such comparison ? Our scientific modes of exact measurement, our
interpretation of the physical world in quantitative terms, itself pre-
supposes psychologically the more rudimentary comparison between the
contents of sense experience. Doubtless, such scientific knowledge
when attained facilitates and modifies the judgments we form of the
increase and decrease of sensual intensities, but in no case can the
latter be wholly dependent on that knowledge.

I 6. " The fact is," says Bergson, " that there is no point of contact
between the unextended and the extended, between quality and quantity.
We can interpret the one by the other, set up the one as the equivalent
of the other; but sooner or later, at the beginning or at the end, we
shall have to recognise the conventional character of this assimilation \"
The abstract severance thus formulated is the basis upon which the
refusal to recognise intensive magnitudes is rested. Extensive magni-
tude, so the argument runs, involves the relation of container to contained,
involves, in other words, the relation of whole and part, and from the
point of view of magnitude, there could be nothing in common between
the extensive and intensive, save the divisibility which the relation of
whole and part implies. But, since intensive qualities are indivisible,
to speak of them as magnitudes is a contradiction in terms.

In his elaborate essay, Ueber die Bedeutung des Weber' schen Gesetzes,
Meinong does not explicitly refer to Bergson's treatment of the subject,
but the essay contains what is still by far the most complete and
conclusive answer to the points that Bergson raises^ I note some of
the main features of Meinong's argument.

The first important fact he seeks to establish is that there are

1 Les donnees immediates de la consciejice, 52. I do not know how the contention in
Mature et Memoire, 242 sgg. et passim, that "all sensations are primarily extensive,"
is to be reconciled with the sharp antithesis that is drawn in the earher work, nor how the
non-quantitative character of sensations is to be sustained in the face of that contention.

2 The second volume of Meinong's Gesammelte Abhandlungen (Leipzig: Bkrth), 1913,
has just reached me. It contains the essay on Weber's Law which has long been out of
print.

Ct. Dawes Hicks 167

quantities which are not divisible, and that such quantities are not
confined to the class of what it has been customary to call intensive
quantities. Some relations are quantities, and relations are not even
conceivably divisible. For example, distance, the apartness of two
points in space, is undoubtedly a quantity, but it is not when rightly
regarded a divisible quantity. True, it is often mistaken for such,
because it is confused with length (Strecke); but the thought of the
length between two points in space is something quite different from
the separation or distinction of two points in space. Distance is a
relation, whereas a length is a whole containing parts. So again similarity
and dissimilarity may be quantities. We talk about a greater or less
degree of similarity, a greater or less degree of dissimilarity. But neither
the one nor the other of these relations is a collection of units. Quite
apart, then, from such intensive qualities as a pleasure or a pain, we
are bound to admit indivisible quantities.

The next thing to notice is a further distinction which has an
important bearing upon that between divisible and indivisible quantities.
Although in ordinary speech unlikeness or dissimilarity ( Verschiedenheit)
is often used as synonymous with difference (Unterschied), yet on purely
empirical grounds we are able to assert that when, as a result of com-
parison, we affirm or deny an unlikeness, we are not judging about
difference. Unlikeness is asserted not alone of quantities ; mathematical
difference can only hold between quantities, and moreover only between
divisible quantities. Thus, the difference between two lines is itself a
line, but the unlikeness between two lines, like every other unlikeness,
is a relation, and in no sense a length. Again, the difference may
remain the same, whilst the unlikeness is not the same. Thus between
1 and 2 there is a much greater unlikeness than between 100 and 101,
though the difference is the same. And in like manner, the unlikenesses
may remain constant, whilst the differences differ, — a condition of things
illustrated by Weber's law.

Now, all measurement rests upon the mental operation of comparison.
As Mr Russell puts it, "without the immediate comparisons, which are
necessary both logically and psychologically, nothing can be accom-
plished : we are always reduced in the last resort to the immediate
judgment that our foot-rule has not greatly changed its size during
measurement, and this judgment is prior to the results of physical
science as to the extent to which bodies do actually change their sizes^"
And Meinong takes pains to make clear that to whatever extent physical
' Principles of Mathematics, 178-9.

168 The Intensity Differences of Sensation

operations may be substituted for mental, yet there is no possibility of
basing measurement wholly upon the former. The process of super-
posing for example would have no meaning, did we not know that when
one thing exactly "covers" another, the result for the most accurate
comparison can only be equality. To look upon measurement as a
purely physical operation would be tantamount to supposing that
addition and multiplication had been converted into physical operations
because both can be carried out by a reckoning machine.

Measurement, however, in the strict sense of the term, is applicable
only to divisible quantities, in regard to which to say that A is double
of B means that it is the magnitude of two quantities together, each of
which has the magnitude of B. Such measurement may be either
immediate or mediate. The former, which Meinong insists is applicable
to time as well as to space, can be replaced by the latter when, as con-
stantly happens, it is more convenient to measure directly a substitute
for the object than the object itself. If, for example, it is a question of
determining the length of a line which forms one side of a square, the
problem may be solved by measuring any one of the other sides, if for
any reason it is easjier to do so. And the possibility of indirectly
measuring indivisible quantities depends upon an extension of this
method of substitution. In all such substitutive measurement (swr-
rogative Messung) that which is actually measured is always a divisible
quantity which serves as a substitute for the indivisible quantity. For
example, distance, as we have seen, is a relation, and as such indivisible.
But every distance, whether spatial or temporal, is associated with a
length, and every length is associated with a distance. A distance
may, therefore, be measured by measuring the length with which it is
correlated. Similarly, in the case of the thermometer, only the height
of the mercurial column can be, in the strict sense, measured, but we
can take that to be a measurement of the temperature, so soon as an
empirically determined regularity has been found to subsist between the
height of the' mercury and the states of temperature. So again, velocity
is not identical with a length and the time in which it has been traversed,
but we regard the velocity as measured when we have measured the
length and time, and divided the former by the latter. The legitimacy
of this process of substitutive measurement depends upon the extent to
which there may accrue to it the advantages which are obtained from
direct measurement. Three things, Meinong finds, give value to direct
measurement. In the first place, a discrete term, namely a number, is
substituted for an element of a continuum, and thereby the intractability

G. Dawes Hicks 169

of the latter is relegated to the unit. In the second place, this number
stands in the same relation of magnitude to other numbers as the given
quantity stands to the other quantities of the same continuum. And,
in the third place, the absolute limits, zero and infinity, which have
validity for indivisible no less than for divisible quantities, coincide for
the numbers and the quantities. Now, of the cases of indirect measure-
ment to which reference has been made, those of distances and velocities
participate in all three advantages, whilst to the measurement of
temperature by the thermometer there accrues only the first of them.
It appears, therefore, that some forms of indirect measurement are more
imperfect or more rudimentary than others.

In the discussion of what he calls " psychical measurement,"
Meinong assumes that by sensual intensity is to be understood
intensity not of the act of sensing but of the content, which he takes
to be no less than the former psychical in character I He points out,
what follows indeed at once from the prior investigation, that the
distinction between psychical and physical does not coincide with the
distinction between intensive and extensive. Some intensive quantities
are to be met with in physical nature, whilst extensity, he thinks, is a
characteristic of some psychical facts. Confining attention meanwhile
to intensive psychical facts, Meinong dismisses as self-contradictory the
conception, introduced by Fechner, of sensation-increments (Emp-
findungszuwuchse). Because, however, there are not, and cannot be,
units of sensation'^, it does not by any means follow that sensations are
jiot measureable, any more than it follows that temperature is not
measureable because there are no units of temperature. If regard be
had to the changes of sensation-intensities, the problem does not
present itself as in any sense a hopeless one. The thought of change
rests upon the thought not of difference ( Unterschied) but of unlikeness
{Verschiedenheity, and the measurement of change carries us back to
the measurement of distance. Change and distance do not, in them-
selves, imply increments and units. Whilst, therefore, the assertion
that the change of sensation from S^ to S^ is equal to the change
from S^ to S* is a perfectly intelligible proposition, the assertion that

^ I should differ from Meinong in this respect, but I am purposely avoiding that issue
in the present discussion.

2 Cf. Mr Bradley's article, "What do we mean by the Intensity of Psychical States?"
in Mind, N.S. iv. 7. Mr Bradley contends that such units exist, although we are not
able in fact to discriminate and fix them.

3 For such a phrase as eben merklicher, or gleich merklicher Unterschied, there ought
to be substituted the phrase eben merkliche, or gleich merkliche Verschiedenheit.

J. of Psych. VI 12

170 The Intensity Differences of Sensation

S^ is so many times greater than S- is not. If, then, the possibility of
treating physical intensities as quantities be admitted — and in regard
to some of them, at any rate, it cannot be disputed — the possibility
of treating sensation-intensities in the same way as quantities must be
conceded. In short, there is no theoretical difficulty in regard to the
measurability of sense-contents. What difficulty there is is a practical
difficulty, and arises from the circumstance that those operations which
give to physical measurement its exactitude are not, as a rule, available.
We are bound to have recourse to substitutive measurement, and the
substitute must be a divisible quantity. If, however, there can be
established on empirical grounds a definite series of correlations
between changes of sensation, which are not numerically determinable,
and changes of some extensive quantity, which are capable of
numerical determination, then we should be just as entitled to take
the magnitude of the latter as measuring the magnitude of the former,
as we are entitled, for example, to measure temperature by means
of the mercurial column of the thermometer. Only we must beware
of taking for granted that no degrees of intensity are possible unless in
fact we can measure them.

One of the chief features of interest in Meinong's analysis is the
clear way in which it is shown that " psychical intensity " is not
sui generis. Upon that assumption Bergson's argument throughout
proceeds. His contention amounts, in short, to this, — that in order to
be quantitative, a sensation would have to be built up, as Fechuer
supposed it was built up, of equal parts or increments. By bringing
"psychical intensity" into line with intensity that is certainly not
psychical, Meinong is enabled to free the former from an utterly
incongruous conception. And when that is done, Bergson's thesis falls.
It is perfectly true that the measurement of sensation-intensities is
possible, if at all, only by a convention. But then that is equally true
of the measurement of distances^ It is perfectly true likewise that
we are dependent upon the immediate apprehension of a change as
revealed by the subjective comparison. But then all measurements
depend in the long run upon immediate judgments of equality, and
these, as also the immediate judgments of greater and less, are still

^ The convention in the case of distances is, as Mr Eussell states it, the following.
It is agreed that, " when the distances a^ui, aia2 ... a„_ia„ are all equal and in the same
sense, then a^a^ is said to be n times each of the distances a^ai etc., i.e. is to be measured
by a number n times as great." This is a convention because "owing to the fact that
distances are indivisible, no distance is really a sum of other distances." Principles of
Mathematics, 180.

(x. Dawes Hicks 171

possible where measurement, in the strict sense, cannot be carried out\
The real question is, how far the immediate comparison of sensations is
reliable, — a question, no doubt, to which very varying answers will be
given, but which does not affect the issue raised by Bergson.

§ 7. I confess I feel less satisfaction with Meinong's solution of the
practical problem, acute and suggestive though his mode of handling it
must be admitted to be. He maintains that, whilst in regard to
sensation-intensities the appearance of equality can never be trusted,
there can in normal cases be no question of an illusory appearance in
the case of unlikeness. What to immediate apprehension appears
unlike is unlike, although what is unlike only appears as unlike down
to a certain limit — the threshold, namely — where the appearance of
equality supervenes. The inferiority of judgments of equality as
compared with those of unlikeness may lead to an apparent paradox,
but it does so not only in the field of psychological but also in that
of physical inquiry. And although at the disposal of the physicist
there are vastly greater facilities for surmounting this defect of the
faculty of comparison, it can never be completely overcome even by
him. From this it follows at once that just appreciable unlikenesses
need not be, as Exner, for example, assumed they must be, equal ;
but where there is equal sensitivity to unlikeness, there is a well-
grounded presumption in favour of their equality. Upon these
premisses Meinong rests the interpretation he has to offer of Weber's
Law. So far as extensive sensations are concerned, it can be said, he
thinks, that proportional sensations correspond to proportional stimuli
and vice versa. With respect to intensive sensations, however, it is
solely a question of equality of unlikenesses, and the law means that
if Ri, R^, R3, Ri be four stimuli and S^, 8^, S3, S^ the corresponding
sensations, then if the proportion i^i : iSa = -B3 : R4 hold of the stimuli,
the corresponding pairs of sensations exhibit equal unlikeness, and the
unlikeness of Sj, and S^ is equal to that of S3 and S^. Owing to his
confusing difference and unlikeness, Fechner assumed that just appre-
ciable differences were themselves sensations, and his logarithmic
formula calls, therefore, for unreserved rejection.

Far more uncertainty, however, attaches, I think, to Weber's Law
than Meinong seems inclined to admit. The assumption that just
appreciable unlikenesses can, even with his proviso, be regarded as
equal, is destitute of any suflBcient grounds. We know but very

1 Cf. BnsBell, ibid. 182.

12—2

172 The Intensity Differences of Sensation

little of the conditions upon which the appreciation of minimal uii-
likenesses depends. One factor at least is hardly amenable to control, —
the state, namely, of the adaptation of the end organs to impression.
In the case of vision, for example, the conditions on which such
adaptation depends are so numerous that they cannot be reduced to
uniformity, and certainly cannot be eliminated. Such a fact as this
alone would almost drive us to the conclusion that the minimal un-
likeness is of variable nature. Moreover, these difficulties would
recur in determining, as Meinong desiderates, where there is equal
sensitivity to unlikeness. And then, again, the number of deviations
from Weber's Law increases as investigation proceeds, so that it is fast
becoming doubtful whether any field for its applicability will in the
end remain. So far as cutaneous sensations are concerned. Rivers and
Head find that it does not hold in respect of protopathic sensibility^
and, indeed, it is hard to see how, except perhaps on Dr Myers's
hypothesis, it is compatible with the 'all or none' principle. As regards
taste and smell, the difficulties of obtaining experimental verification
would certainly in any case be considerable, but, in spite of Miss
Gamble's careful piece of investigation^, it cannot, I think, be claimed
that the applicability of the law to either of these senses has been
placed beyond the reach of doubt. With respect to vision, there does
not appear to be such independence of intensity and quality as would
be requisite for the establishment of the law, and the same is to be
said of hearing. At all events, the law cannot be taken to be more
than an interesting empirical generalisation, based upon experiments
that have not been purified from interfering circumstances, and must,
even where it would seem to hold, be an expression for an extremely
complex set of conditions.

§ 8. " It must never be forgotten," writes Sherrington, " that
Weber's Law deals with judgments. The comparison of one sensation
with a second of similar quale, but of dissimilar quantum, involves more
than the mere neural process concerned with a simple sensation. From
the very outset it works with ideas based on perceptions^" What is
liere said is certainly true so far as attention is confined to the elaborate
comparisons on which Weber's Law is rested. But Sherrington's
contention suggests an interesting question as to the ultimate psycho-
logical nature of the appreciation of unlikeness. The developed act of

1 Brain, Nov. 1908, 428-9.

2 Amer. J. of Psychol. 1898, x. "The Applicability of Weber's Law to Smell."
* Schafer, Textbook of Physiology, n. 932-3.

G. Dawes Hicks 173

comparison always involves a reference to the objective order of fact as
distinguished from the sense contents. It cannot, however, be supposed
that any such objective reference is present in the rudimentary sense-
experience out of which the recognition of an objective order has
gradually emerged. That rudimentary experience could only have
contained, at the most, the simple foundation on which the later
process of judging psychologically depends. The inference seems
inevitable that originally appreciation of unlikeness is itself a com-
ponent of sense-experience. And the inference certainly gains
confirmation from the consideration that the unlikenesses which we
discriminate, be they great or small, are as much given as are, for
example, the distinct sense contents which are pronounced unlike.
Thus, I think, we are enabled to see that the sense-data we are
supposed to compare in the developed act of judging degrees of
intensity are not, in truth, sense experiences in the strict acceptation
of the term. They are abstractions from sense-experience, and the
isolation we artificially produce by working on the given material
serves to disguise from us the actual nature of the experience we thus
manipulate. Unlikeness, that is to say, is not something added to the
contents of sense-experience from some other function of the mind ;
it does not arise for the first time when a complex act of judging comes
into play ; it is no less an element or aspect of that sense-experience
than the distinguishable contents themselves. Such unlikeness may be
of various kinds. The apprehension of quantity comes, we may agree
with Mr Bradley, later than that of quality, if, that is, quality be taken
at its crudest stage. But we have no ground for supposing that the
elementary discrimination of either the one or the other necessitates a
function of mind different in kind from that of sense-experience itself.

§ 9. If pressed to give a definite answer to the question whether
intensity differences of sensation are quantitative, the reply, I presume,
would have to be framed in some such terms as these. As the mathe-
matician conceives of quantity, the only quantities whose differences
may likewise be described as quantities are divisible quantities.
Consequently, the difference or unlikeness of two intensive quantities
is not itself a quantity, — which amounts, in other words, to saying that
these quantities are not multiples of an element or unit similar in
quality to themselves. Using for the moment the word difference in its
non-mathematical sense, one would assert that just as the difference
between two distances is not itself a distance, so the difference between
two sensations is not itself a sensation. To quote Mr Russell's well-

174 The Intensity Differences of Sensation

known dictum, " the difference between two intensive quantities, in
fact, differs from each as much as the difference between two horses
differs from a horse." The distinction, however, which Mr Russell, in
his Principles of Mathematics, draws between quantities and magnitudes
would, I gather, enable us to speak of intensity differences of sensation
as magnitudes. Magnitudes, as he would employ the term, are more
abstract than quantities. A specific magnitude is a common property
of a number of equal quantities. An actual foot-rule, for example, is a
quantity ; its length is a magnitude. A quantity is anything which is
capable of quantitative equality to something else — that is to say, which
is capable of possessing the same magnitude as something else. Properly,
one quantity ought not to be described as greater or less than another,
for the relations of greater and less hold between their magnitudes.
On the other hand, properly one magnitude ought not to be described
as equal to another magnitude, for the relation that would be really
meant in such a case would be the relation of sameness or identity.
Thus, for example, suppose a sound A possesses the loudness a and
a sound B possesses the loudness /S. A and B are each of them
quantities ; a and ^ are magnitudes. If A is louder than B, then the
difference a — y8, let us call it y, is not a sound possessing magnitude ;
7 simply is a magnitude. If A resembles B in loudness, then a and y9
are not two magnitudes, a is the same magnitude as, or is identical
with, y9. And the difference or resemblance of A and B in loudness is
a magnitude, because it is greater or less than other differences or
resemblances, such, for instance, as the difference or resemblance in
loudness of the sounds G and D. " Quantities not susceptible of
numerical measurement can," says Mr Russell, " be arranged in a scale
of greater and smaller magnitudes, and this is the only strictly quanti-
tative achievement of even numerical measurement. We can know
that one magnitude is greater than another, and that a third is
intermediate between them ; also, since the differences of magnitudes
are always magnitudes, there is always (theoretically, at least) an
answer to the question whether the difference of one pair of magnitudes
is greater than, less than, or the same as the difference of another pair
of the same kind. And such propositions, though to the mathematician
they may appear approximate, are just as precise and definite as the
propositions of Arithmetic'."

' Principles of Mathematics, 183. Cf. 159.

AEE THE INTENSITY DIFFERENCES OF
SENSATION QUANTITATIVE?^ III.

By henry J. WATT.

1 . Which differences of sensation do we call intensive ?

2. (a) What psychological place does intensity occupy amongst the

attributes of sensation ?
(b) In what relatio7i does intensity stand to those modes of
experience which hear a close psychological affinity to
sensation and its attributes ?

3. What is m,eant by the term ' quantitative ' ?

4. Is intensity a multitude or a magnitude ?

5. What other objects besides intensity are at least magnitudes ?

6. Can intensity possibly be treated as a multitude ?

7. The source of the confusion.

This question may be specialised into a series of questions. The
answers given to them will not only indicate the special points at which
differences of opinion may legitimately arise, but will also show that
certain differences are due to a confusion of ideas and may therefore be
eliminated.

1. Which differences of sensation do we call intensive ? It is
agreed, I think, by all that the classification of certain differences as
intensive cannot possibly be called in question. The cutaneous, mus-
cular, gustatory, olfactory, and auditory sensations all possess the
undoubtedly similar attributes of intensity. We may, of course, enquire
whether intensity is native to all these groups of sensations and, if not,
how they came to acquire it. But that it is there, is surely not disputed.
Nor does the absence of any marked degree of variation of intensity, as
for example in the articular sensations, really present a difficulty. The
only important problem in this connexion is whether the particular
case of visual brightness is to be classified as a form of intensity or as a
form of quality or the like. But we can afford to neglect this problem

1 A contribution to the Symposium presented at the Joint Meeting of the British
Psychological Society, the Aristotelian Society, and the Mind Association in London,
7 June, 1913.

176 The Intensity Differences of Sensation

here and to confine our attention to the accepted cases of intensity. If
visual brightness is to be considered intensive, the conclusions which
are obtained for accepted forms of intensity will apply to it. But it is
of interest to notice in passing that the proper classification of the
attributes of sensation is not a perfectly simple task. There is room for
serious divergence of views even at the present time^ Introspection
is, of course, the only ground upon which a true classification can be
founded. But it must be granted that the first, or in fact, any single,
deliverance of introspection about the inherent nature and connexions
of experiences is not necessarily irreproachable. We have to learn to
think truly about simple experiences, just as about the objects of the
physical world we live in.

2. (a) What psychological place does intensity occupy amongst
the attributes of sensation ? In discussing whether intensity is quanti-
tative or not, reference is frequently made to extensity, as if the latter
were undoubtedly quantitative. A certain amount of prejudice against
a negative judgment regarding intensity is thus created. If this pre-
judice is misleading, it must be removed. I do not think that extensity
can legitimately be considered to be a variable attribute. It is invariable.
It is not really less present in the sensation from a ' spot ' than in that
from an area ; there is not more of it in a square inch than in a square
centimetre of colour. Nor is a low tone properly more voluminous than
a high one. What there is more of in these cases is extent or volume,
not extensity or voluminosity. We have indeed said for long enough
that low tones are more voluminous than high ones. We had perhaps
good reason to fear a confusion between the volume of a tone and the
volume of the physical material, if we had used the same term for both.
But nowadays this confusion can hardly occur in reference to the study
of sensation. It is no longer from without, but within the field of
psychology that the danger appears.

The variant commonly referred to under the name of extensity,
voluminosity, and massiveness, then, is not an attribute of sensation.
It is a derivative, a higher product, a Gestalt, like that of a line or a
curve, and it is variable in the sense of being greater or less, like these.
The attribute of extensity^ is the common basis of extensiveness, the
real ground of fusion and continuity of sensation in the midst of
differences of local sign and its analogues, position and pitch, which
I prefer to group under the generic name of orders If the same

1 Cf. my discussion of pitch and other cases in this Journal, iv. 843 ff.

2 I hope to deal with this attribute more fully at another time. ' Cf. op, cit.

Henry J. Watt 177

distinction is applied to the attribute of temporal extensity or duration,
we obtain the following grouping of the attributes. Quality and
intensity stand apart from the others, which fall into two pairs. Each
pair comprises an extensive and an ordinal member and the two pairs
may be named temporal and systemic. But, however tempting it may
be for the purpose of systematic appearances, it is impossible to treat
either quality or intensity as extensive or as ordinal in character. They
are both ordinal in the sense of being self-disposing, but this peculiarity
of them cannot be identified with ordinality; for upon differences in
quality or in intensity none of those Gestalten or modes of sensory
experience are founded which grow upon ordinal contents, viz. distance
or interval, motion and others ; and besides, quality and intensity are
both more than merely self-disposing.

(b) In what relation does intensity stand to those modes of
experience which bear a close psychological affinity to sensation and its
attributes ? These modes of experience have been forcing themselves
with ever increasing insistence upon the notice of psychologists. There
can be no doubt about their enormous variety and importance. Since
the leading paper by Chr. v. Ehrenfels in 1890, by whom they were
called Gestaltqualitdten, a large number of studies have been made of
them^ Such modes of experience are said to be founded upon contents,
which may either be other modes or in many cases elementary
sensations. I believe that there is always a certain amount of
resemblance between the founded mode and its founding contents or
some aspect or attribute of the latter, as well as an objective psychical
dependence of the mode on its founding contents. These relations seem
to me to form good ground upon which a body of pure psychological
theory concerned with the interconnexions and development of experience
may be built up^. Many modes are variable and self-disposing, as being
greater or less than one another in respect of their own peculiar pheno-
menal content. Distance and interval of time and motion are amongst
the simplest of them, but there are many others^. The full and adequate
study of these modes, their variety, relations, and theoretical explanation,
is one of the newest forms of the psychological task, and will undoubtedly
show itself to be one of its bulkiest parts.

In this connexion I see reason to differ from certain views indicated

1 The first volume of a most valuable and important work by Karl Biihler on Die
Gestaltwahrnehmungen has just been published (1913).

* Cf. my paper on the " Psychology of Visual Motion," in this Journal, vi.
^ This Journal, iv. 157 ff. For other modes cf. Biihler, op. cit.

178 The IiUensity Differences of Sensation

by Dr Myers in i. § 2. The modes which stand next to elementary
sensation seem to be, first distance and time-interval, and then, as a
combination of these two, motion. Spatiality, if it is merely simple
distance, seems to me to be in the matter of psychological origin
independent of motion ; if it is complex, such as is the spatiality of
binocular vision, it does not seem to me to involve motion as a necessary
psychological antecedent at all. Nor do I see any evidence for the
existence of a psychological antecedent to intensity, simpler than
intensity, from which intensity might arise by the integration of two or
more of its varieties, as distance may be said to be integrated out of
differences in the attribute of order. Any other speculations regarding
the origin of intensity seem to me to be either inventions or to rest
upon mistaken correlations.

3. What is to be meant by the term ' quantitative ' ? It seems to
be agreed that there are two possible meanings. A quantitative object
is either,

(a) A collective object, whether real or ideal — a number of material
particles, persons, states of mind, events, or a number of ideal numbers,
lengths, forces, universals. Let us call this kind of object a multitude.

Or (6) A self-disposing object, or an object say ««, which in virtue
of its own phenomenality disposes itself amongst other objects of the
same gi'oup a^, a^, a^, etc., in a definite manner, so that it falls between
ar and at, and not between a^ and ay, and which in these relations
appears to be greater than a^ and less than a^ This kind of object
is known as a magnitude.

4. Is intensity a multitude or a magnitude ? [With regard to the
expression ' intensity differences ' in the title of this discussion, I take
it to mean, in the first place, intensities, and only in the second place,
if at all, differences of intensities, such as those between la and Ih, lb
and /c] On two points there seems to be agreement: (a) intensity
is at least a magnitude ; and (6) we cannot yet validly treat it as a
multitude. We can, therefore, proceed to discuss the possibilities that
are logically unaffected by these decisions. But before doing so it is
well to turn aside for a moment and ask another question.

5. What other objects besides intensity are at least magnitudes ?
It is agreed, I think, that felt distance and motion and other such
modes of experience or Gestalten are also at least magnitudes. We may,
therefore, infer that the world of experience is rich in objects of this kind.
Probably all forms of experience are, in some sense or to some degree,
self-disposing objects. But a number of them cannot be considered to

Henry J. Watt 179

be magnitudes, for example the above mentioned attributes of temporal
and systemic order, percepts, recognition, concepts, thoughts and the
like. The most obvious groups of experiential magnitudes are the
modes and figures (Gestalten) of space and time, their combination in
motions, and the various classes included under the term ' relations.'
Magnitudes seem to occur by preference on what is obviously a duple .
or multiple foundation, such as we find in distance, succession, and
change, or on what for various reasons may legitimately be held to be a
duple or multiple foundation, as in minimal distances, motions, changes,
etc. Feeling is one of the few cases in which a duple foundation seems
to elude our grasp, but even here there is some sort of positive evidenced
But there is at least no reason to doubt that differences of intensities
are magnitudes and that we find it comparatively easy to arrange them
and to observe and to indicate their apparent equality. In so far as we
consistently maintain their phenomenal equality, we have as much
reason to believe in the validity of our introspective judgments, as we
have to believe in them in other regions of introspective work. But if
a, b, and c are not multitudes, but experiential magnitudes, we cannot
suppose that judgments regarding the equality of the differences
between a and b and between b and c justify the statement that the
difference between a and c is twice that between either of the former
pairs. For the judgment regarding a and c has no bearing on the
other two judgments, and vice versa. All just noticeable differences
are equal in being just noticeable, but that does not make them equal
increments. Nor can equal differences be considered to be equal parts
of another difference, i.e. equal increments within the latter difference.
Is there any sense in calling the tone interval g — -f twice as great as
that bertween g and c', because the intervals g — c' and c' — f are equal
in being fourths ? Besides, a distance is not the difference between
two points, but these and the stretch between them in a unity.

It would carry me too far from the object of this discussion, were
I to enquire whether any non-mental, for example, material or ideal
objects, are at least or solely magnitudes. Nor do I think it would
throw any light upon the object of discussion.

6. Can intensity possibly be treated as a multitude ? The con-
clusion I wish to plead for in this discussion is that it cannot, so long
as the identity of the object under discussion, namely intensity, is
maintained. I would suggest that an object cannot at one and the

^ Cf . my discuasion of it in this Journal, rv. 184 ff.

180 The Intensity Differences of Sensation

same time be directly immeasurable and indirectly measurable, as
Meinong^ declares and as Professor Dawes Hicks- agrees. Such a
proposition can have an appearance of plausibility only by the sub-
stitution of a new measurable object for the one that is directly
immeasurable. This substitution may be occasioned by the close
connexion of the two objects in the world of reality, but it is none the
less a substitution. To speak of a surrogative form of measurement is
both misleading and wrong. What the medical thermometer measures
is not the patient's sensations of warmth or cold or how warm or cold
he feels. In this particular instance the departure from any sort of
regular correlation between magnitude of felt warmth and degree of
temperature is notorious. What the physician is usually concerned to
know is the temperature of his patient's body. And that is as little a
surrogative measurement of his patient's feelings as the sight or taste
of the physic he offers is a surrogative cure for his patient's felt discom-
fort, however much or little the material physic may be suited to restore
the patient's body to its normal condition. No one sets out to measure
the sensed distances evoked by a thermometer scale, but only the lines or
lengths of that scale. The latter are measurable, as are any multiple
objects, in so far as they produce regular changes upon lines or lengths.
In all cases it is only that aspect of the motion of matter which by an
obvious convenience has come to be called temperature that is measur-
able. And similarly in other such examples.

I would also submit that in every case in which the treatment of
single states of mind as multitudes is in any way made to be plausible,
we find a substitution of objects of the kind mentioned. So for example
in Fechner's formula, which is perfectly valid in so far as ^ in the
expression S = K Log / means ' the numerical value of S,' if it exists.
But unfortunately this value has no real object; the object and the
value are purely imaginary. The fault here does not lie in the applica-
tion of mathematical symbols and processes to the data of sense ; for
these are most certainly applicable to the data of sense whenever we
have an opportunity of dealing with multitudes of these data, e.g. in the
statistical manipulation of records of the frequency of visual and other
images, in the study of memory and so on. The error committed by
Fechner consists in applying mathematical symbols and processes to the
data of sense without any proper psychological or objective justification.

1 Ztschr.f. Psychol., 1896, xi. 239,

2 Cf. pp. 168 ff.

Henry J. Watt 181

There is no theoretical difficulty in discovering truths that are non-
truths. The difficulty is always a 'practical' one. The truths 'wanted'
are simply not there to be had.

The substitution of objects I speak of may also be illustrated from
Dr Myers's main thesis that the physiological correlate of intensity
differences is a sub-group of extensive changes. That may very well
be, but the thesis, as it stands, cannot be considered to afford any inter-
pretation or elucidation of intensity or its differences. If it is a valid
hypothesis, it certainly establishes a fact, it discovers a reality, a new
kind of extensive distribution of physiological processes; and it sets
this reality into relation with intensity. But that is all. We are not
thereby brought any nearer to a treatment of intensity as a multitude.
We merely know now a relation in which intensity stands that we did
not know before. It does not affect the case in the least that the object
with which intensity has been shown to stand in relation is itself a
multitude. Physiology can be said to throw light upon psychological
matters only in so far as a sufficient number of these relations between
experiences and physiological processes are discovered to warrant the
inductive assumption that certain known physiological units stand in
certain relations to known psychical units or that certain as yet unknown
psychical units exist and are related to these known physiological units
in certain ways. I do not by any means deny the possibility of this
inductive procedure. But I very much doubt whether the reverse does
not constitute the method of greater illuminative power.

In short, no single state of mind can be treated as a multitude, not
even the idea of 100 itself Only the object of the idea of 100 can be
so treated. But I do not mean hereby to imply that every object can
be treated as a multitude. We must, of course, discover and determine
whether any given object can be so treated or not. If we succeed, the
object is a multitude ; if we do not succeed, it may often still be a
multitude. We cannot tell a priori where we are to look for objects
that are multitudes and where not. Otherwise psychologists have made
a sorry waste of their time and energy. It is quite possible that some-
one may yet prove by new methods that behind intensity there lie
psychical objects now unknown to us which are to be considered as
multitudes and are responsible for the phenomenon of intensity (cf
Myers, i. § 2). But not even such a proof would enable us to look upon
intensities as themselves multitudes. Such a magnitude as intensity,
like the so much discussed and practically useful distance, must remain
a magnitude for ever and ever.

182 The Intensity Differences of Sensation

This may be enforced by another illustration. It is possible to
maintain that felt distance is realiter psychologically founded upon
repeated {i.e. a multitude of) sensational elements qualified by extensity
and order and that thus differences of multitudes are the real basis of
the differences of magnitude found in distances. But not even that
would make distance in any sense a multitude. Only its real psycho-
logical basis would be a multitude ^

If we had such as this imaginary knowledge of the real psychical
basis of intensity, we might formulate the laws of mind and predict the
psychical future better than we do now. But future mental states can
be predicted by the knowledge of the physical world we already possess.
We can, for example, arrange the illumination of a room so as to produce
various mental effects. Yet that fact does not imply that we can
measure intensity or its differences. Nor would the discovery I
imagined.

If, finally, it be suggested that intensity can be treated as a multi-
tude or measured by convention, I would submit that such ' measurement '
is only a means of naming what stands in a real relation to something
else that can properly be measured, as star brilliancies to the varying
intensity of physical light.

7. The source of the confusion in these matters is an epistemological
one — either a confusion of objects or a confusion of the immediate basis
of knowledge in sensory experience with the objects of knowledge. In
the latter case distance as sensed, for example, may be confused with
length, felt motion or its velocity with motion through real space or the
velocity of real motion. But it is surely absurd to suppose that any
sort of reality — called velocity — exists that is a unitary magnitude in
the sense in which colours and tones and felt velocity are such, and that
nevertheless is measurable in numbers. Such a unitary reality is a
myth, the hypostatization of a complex set of correlated relations in
which a real or ideal object stands. Whether these relations are
themselves real or ideal, actual or imaginary, makes, of course, no
difference to the case.

If I rejoice in the possession of a new book, neither the possession
nor the book thereby become feelings or emotions. If I know yonder
tree is budding, neither the tree nor the budding thereby become either
sensations, perceptions, or knowledge. They are only the objects of my
knowledge and as such come into relation to my knowledge. So if I can

1 Cf. the analogous theory given by E. R. Jaensch of the psychical representation of
empty space, Ztsch. f. Psychol., Erg.-bd. ti. 244 S.

Henry J. Watt 183

measure lengths, why should I worry about not measuring distances as
felt (Gestalten), when I have already ascertained that I cannot measure
them ? If lengths are in fact measurable, the equality or differences of
distances may be the sensory basis on which the cognitive processes of
conception and knowledge involved in the act of measurement build.
But that is no reason why I should require or expect to be able to
measure distances. If unitary distances are not to be converted into
multitudes, we must just enquire how our cognitive processes can
nevertheless' make measurement of lengths possible. It is futile to
think distances ought somehow to be measurable or to construe them
so as to imagine them measurable. A real object has certain definite
properties and it stands in certain definite relations to other objects; all
one can do is to find out these things by knowing. Knowing powers
will never by themselves alone change the properties of objects or set
them into new relations, unless these be relations to my knowing or
unless I somehow act upon the objects so as to change their real
relations.

It seems necessary to make these remarks as there is a consensus of
opinion that we actually do not succeed in measuring mental magnitudes
such as intensity ; and yet attempts are made to give the impression
that after all our intellect is not so ineffective and useless as it is (most
perversely) considered to be and that we really do measure these magni-
tudes ; only we do not do the measuring in these cases directly or
straightforwardly but indirectly or by substitution, or to put it bluntly
by make-believe.

ARE THE INTENSITY DIFFERENCES OF
SENSATION QUANTITATIVE?^ IV.

By WILLIAM BROWN.

§ 1. Qualitative and Quantitative.

§ 2. The relation of physical measurement to extensity and protensity.

§ 3. Conditions fulfilled in physical measurement.

§ 4. The theories of Fechner and Delhoeuf.

§ 5. The thermometer analogy.

§ 6. How far is a ' sense- distance ' quantitative ?

§ 7. Dr Myers's view.

§ 8. Professor Uicks's view.

§ 9. Dr Watt's view.

In view of the length and thoroughness of the foregoing discussion
on this subject, I trust I may be excused for making my remarks as
brief as possible. After setting out my own reasoned opinion on the
matter, in the fewest possible words, I will indicate how far I am in
agreement with my predecessors in the argument, and discuss those
points where I am in dissent.

§ 1. One of the most general statements that can be made about
the stream of consciousness is that every moment or pulse of this stream
is qualitatively different from every other. Any concrete datum of
experience is only identical with itself To say that it is equal, greater
or less, as such, than any other would be meaningless. Nevertheless
the qualitative similarities observable within experience justify the
distinction of aspects of consciousness each showing a homogeneity,
or a unity in difference, of a particular kind. Examples of such aspects
or attributes which are relevant to our problem are extensity, protensity,
clearness, penetratingness {Eindringlichkeit), saturation (of a colour),
brightness (of a colour), and intensity. They are not only homogeneous
but also show degrees and admit of the use of the words 'greater'
and ' less ' in their description. They may therefore be regarded as

1 A contribution to the Symposium presented at the Joint Meeting of the British
Psychological Society, the Aristotelian Society, and the Mind Association, in London,
7 June, 1913.

William Brown 185

magnitudes. Whether they are to be regarded as quantities, or measur-
able magnitudes, is another question. A consideration of the cases of
extensity and protensity makes this more than doubtful. These aspects
of consciousness must clearly have been the preconditions of the develop-
ment, in the course of mental evolution, of our knowledge of objective
and quantitative space and time, since if we think them away such
knowledge would be inconceivable. Without attempting the difficult
task of showing how this development took place, we may at least take
it as a fact that the use of the principle of superposition and the dis-
covery of fixed numerical relations, in terms of definite units, between
things in space and time gave birth to physical measurement. Measure-
ment was produced by measuring and not by philosophical analysis.
Quantitative relations are characteristics of the real world which are
proved to exist by the tentative process of experimenting. Whether
forms of measurement other than the physical are possible can only be
decided in the same way.

§ 2. But this very success of physical measurement might tend to
suggest that attempts at psychical measurement rest on a misconcep-
tion. The differences of degree of extensity and protensity have found
their real measurement in the measures of surface extent and time.
And although the existence of illusions shows that this correlation
of psychical feeling and physical size is not a complete one, the expla-
nations of these illusions which are asked for and are found rest on
belief in a real correspondence. All that follows, however, fi'om such
an argument is that the accurate measurement of extensity and pro-
tensity is not worth the trouble of carrying out, not that it is theoreti-
cally impossible.

§ 3. For measurement to be as complete as that m physical science,
we have seen that the following conditions must be fulfilled: (1) the
thing or attribute measured must be homogeneous and show degrees
of more and less, (2) a unit must be discoverable, of which the given
magnitude may be said to be a certain multiple or sub-multiple,

(3) there must be a zero from which the measurements are made,

(4) it must be possible, theoretically at least, to superpose one magni-
tude on another, and so get the measure of a difference or the proof
of an identity. It does not, however, necessarily follow that in cases
where some of these conditions are not fulfilled the thing or attribute
is not a quantity. There may be different kinds of measurement, of
different degrees of completeness. Experience alone will tell us how
complete it may be made.

J. of Psych. VI - 13

186 The Intensity Differences of Sensation

§ 4, Fechner tried, by building on Weber's Law and making
a psychological use of the concept of the limen, to get a scheme of
measurement of sensation-intensities themselves. He failed, because
no sensation-intensity can be regarded as a sum of smaller intensities.
Intensities, like the other general attributes which we have mentioned,
obey only the first, and, perhaps, the third of our four conditions.
Delboeuf quickly showed, however, that the sensible contrasts, or sense-
distances as they may be called, between sensation-intensities do admit
of division and summation. This conclusion was not based upon
a priori argument but upon experiment.

He showed that the problem of finding an intensity of grey which
should lie half-way between two more extreme shades of grey, or bisect
the sense-distance (contraste sensible) between them, was a real one
and admitted of a real solution. Despite a certain variability of the
judgment in the case of different persons or of the same person at
different times, which was only to be expected from the nature of the
experiment, there was an average constancy of the result which indicated
with a high degree of probability that the problem had been genuinely
solved. If it be objected, with Bergson, that the subject of the
experiment judges the intensities in terms of the corresponding stimulus-
values with which he has become acquainted in previous experience, we
may point to the fact that the stimulus-value of the middle grey is the
geometric, and not the arithmetic, mean of those of the extreme greys.
This experiment has been done many times in my laboratory by
numerous students during the last few years and the results have
always been definite and unequivocal in nature.

Whether it be light-intensities or sound-intensities or the intensities
of sensations aroused in lifting weights, the results show a remarkable
constancy. Stimulus-values so chosen as to form a geometrical pro-
gression give sensation-intensities forming a series of equal-appearing
intervals. Ebbinghaus expresses this relation thus :

where SSo is the sense-distance between ^o, the conventional zero-
sensation which may itself be any degree of intensity and not actually
zero, and a given sensation S, Ilo> ^ are the corresponding stimulus-
values, and A; is a constant. The correspondence of the equation with
that expressing the Weber- Fechner Law suggests that the just noticeable
differences of sensation-intensity involved in the latter may not only be
regarded as minimal sense-distances but also as equal to one another in

William Brown 187

different parts of the scale of intensities. The equality of just noticeable
differences (distances) of sensation, which Fechner assumed on the basis
of introspection, may be experimentally tested, by noting the number
of just noticeable differences necessary to take the observer from one
end to the other of a number of equal-appearing sense intervals. The
results of these experiments are somewhat conflicting, but the balance
of evidence seems in favour of the equality of the just noticeable
differences.

§ 5. In the choice of a conventional zero from which sense-distances
may be measured, the analogy of the thermometer is often referred to
as a justification. Just as two arbitrary points, corresponding to the
freezing point and boiling point of water respectively, are chosen and
the distance on the thermometer tube between these two points is
divided into a hundred equal divisions, so a sensation of low intensity
(not zero) may be taken as the conventional zero, another sensation-
intensity near the upper limit of intensities may be arbitrarily fixed as
100, and the sense-distance between them may then be divided into
100 equal sense-distances. In such a case as this the intensity bisecting
the two extremes would have the value 50, as being 50 units of sense-
distance removed from the conventional zero. It seems to me that
Dr Watt has mistaken Prof Dawes Hicks and Meinong in regard to
this analogy, by assuming that Meinong took it as more than an analogy
and considered the thermometer as an instrument for the indirect
measure of subjective intensities of heat-sensation. Meinong saw in it
a means of indirectly measuring temperature, which is a very different
thing, and for this reason Professor Hicks's use of the argument seems
to me to be perfectly sound.

§ 6. The system of psychical measurement which I have outlined
above is that accepted by Stumpf, Ebbinghaus, and Titchener among
modern psychologists. It fulfils the first three of the four conditions
found satisfied in physical measurement (see § 3) ; but although the
principle of superposition is inapplicable, this defect is not sufficiently
serious to rob it of the claim to be called measurement. Although it
cannot, for this reason, hope for so successful a career as that which
physical measurement has enjoyed, it nevertheless has many possibilities
before it which when realised may transform the science of introspective
psychology. To object that sense-distances can never be equal because
they start from different degrees of intensity is to overlook the element
of abstraction essential to all measurement. And what the system lacks
in the matter of superposition is in some degree made up to it in the

13—2

188 The Inte7isity Differences of Sensation

possibility of controlling and measuring the corresponding stimulus-
values. By means of the logarithmic law we can pass from the one to
the other at will.

§ 7. Turning to Dr Myers's treatment of the problem I find nothing
that is in conflict with the theory just mapped out, but much that is
a very valuable supplementation of it. His theory of the physiological
correlate of sensation-intensity is especially valuable as emphasizing
the differences in the laws of intensity-change obeyed by the different
classes of sensation and providing a physiological explanation of them.
There is some little difficulty in imagining a complexity of efferent
tendencies and motor responses sufficient to account for all the shades
of qualitative difference between the elements of sensory consciousness
and in some parts of Dr Myers's essay there are phrases which seem to
imply interactionism rather than the psycho-physiological parallelism
which is the basis of his general system {e.g. " At bottom, differences in
type of movement must be the cause of differentiation in the quality of
sensation ; it would be of no advantage for the organism to experience
different qualities of sensation, unless those differences were serviceable
in promoting different types of response ") ; but apart from these small
points, which are not real objections, I can accept his views. On the
psychological side, I should agree with him that sensation-intensity is
sui genet-is, but this does not prevent contrasts of intensity being
quantitative and therefore measurable.

§ 8. With Professor Dawes Hicks on the psychological aspect I am
equally in agreement. He has gone into the epistemological side of
the problem so thoroughly that it did not seem necessary for me to
devote any further attention to it. To my own mind, the question now
is one of practice; not "Is measurement of intensity differences theo-
retically possible ? " but " Has such measurement been practically
achieved ? " and the answer to this question seems to be in the affirma-
tive. It is important to have so convincing and detailed a refutation of
Bergson's negative as Professor Hicks gives, and also to know that
Meinong has made mental measurement so theoretically plausible ; but
as soon as we can get together a body of practical results of which a
beginning (a very small one, it is true) has already been made, these
questions will become of merely academic interest, although of course
their importance for metaphysics will remain.

I 9. I find Dr Watt's paper a little disappointing, since, although
he says much that is of interest about intensity in its relation to other
attributes of sensation and other forms of consciousness, he seems to

William Brown 189

think that by denying to intensity the characteristic of being a ' multi-
tude ' he has settled the question of ' quantity.' We have seen above
that ' sense-distances ' are quantities even although intensities them-
selves are not. Moreover his criticism of Meinong's surrogative form of
measurement seems to rest on a misunderstanding, as I have already
pointed out. Finally I find it impossible to agree with him when he
writes, " I do not think that extensity can legitimately be considered to
be a variable attribute. It is invariable. It is not really less present
in the sensation from a ' spot ' than in that from an area ; there is not
more of it in a square inch than in a square centimetre of colour." It
seems to me that such a view would deprive extensity of all its value in
enabling us to understand the development of the perception of surface
extent, besides being unsupported by introspection. In extensity dis-
tinctions of ' more ' and ' less ' seem to be as clearly present as in
intensity or any other of the attributes of sensation, and to this extent
extensity, like intensity, is a magnitude.

THE AESTHETIC APPRECIATION OF MUSICAL INTERVALS
AMONG SCHOOL CHILDREN AND ADULTS.

By C. W. VALENTINE,

Lecturer in Ecsperimental Psychology to the St Andrews Provincial
Committee for the Training of Teachers.

I. The purpose of the experiments.

II. The method of experiment with adults.

III. The order of popularity of the intervals.

IV. The aesthetic effect of the different intervals. Major and

minor intervals. The octave. Concords felt as discords.
V. The method of experiment with school children.

VI. Results of the experiments vnth Elementary School children.

VII. Results of the experiments with Preparatory School children.

VIII. Comparison of the results of the experiments in the

Elem,entary and Preparatory Schools.
IX. Tests for a ' musical ear.'

X. Introspection of school children.

XI. Sex differences in the Elementary School experiments.
XII. Summary of results and canclusions.

I. The purpose of the expenments.

In 1910 some experiments were begun in order to test the aesthetic
appreciation of musical intervals among school children. The object
was to discover, if possible, something as to the development with age
of a feeling for consonance, and to determine the differences in this
respect among children belonging to different cultural groups and
having had different degrees of musical training. It seemed desirable
also to obtain results from adults, for the sake of comparison.

Apart from this, I wished to ascertain the extent to which in-
dividuals could be divided into 'perceptive types' according to their
attitude towards musical elements, as Mr E. Bullough has classified

C. W. Valentine 191

them in reference to colours S and to note any marked difference be-
tween the sexes in reference to their appreciation of musical intervals^.

Further, a great deal of uncertainty still exists as to the order of
pleasingness of the twelve intervals playable upon the piano within
one octave, though there is of course a general agreement as to which
intervals are dissonant and which consonants Some indeed seem to
identify pleasingness with consonance. Others have doubted whether
we can isolate an interval so as really to hear it alone, their view being
that the pleasingness of an interval will depend not merely upon its
degree of consonance, but also on the more or less vague suggestions of
other notes. This is of course directly opposed to the assertion that
one can isolate an interval and that all consonances are then more
pleasing than any dissonances. As to the varying degrees of pleasing-
ness of the intervals one must expect great differences among individuals,
though there seems to be a general agreement that to the modern ear
the Third is the most pleasing interval, whilst during the middle ages
the Fifth was probably the most popular, and with the Greeks, the
Octave.

We shall describe the experiments with adults first. The method
of experiment with the children was substantially the same.

II. The method of experiment with adults.

The subjects of these experiments were university students at
St Andrews, or students in the Training College, Dundee. They
numbered 146, of whom 84 were women and 62 men. The great
majority of the women were Scots : of the men about one quarter
were English, one fifth Welsh and the rest Scots. Most of the tests
were done with a Chappell piano in excellent condition, the rest upon
a new Bechstein, both regularly tuned*. The subjects were taken in
groups — about 18 in each group on an average. The necessity of
perfect silence and absolute independence of judgment was emphasized.
All the subjects had undergone a course of experimental psychology
and, I think, must have appreciated the importance of guarding against

1 This Journal, ii. 406.

2 The discoveries as to perceptive types are not given in this paper. They will be
included in a subsequent joint paper by Dr C. S. Myers and the present writer.

^ As to the degree of consonance of the various intervals see C. S. Myers's Text Book
of Experimental Psychology, 2nd edition, i. 27.

* It should be recalled that, owing to the tuning of pianos by the method of 'equal
temperament,' the intervals (with the exception of the octave) have not their exact
theoretic value.

192 The Appreciation of Musical Intervals

the slightest amount of suggestion in such experiments. They were
provided with paper on which they were asked to record their judg-
ments upon the chords played, stating whether they found them very
pleasing, pleasing, slightly pleasing, indifferent, slightly displeasing,
displeasing, or very displeasing, adding the reason why, if possible.
Each interval was played twice, the notes first being held down for
three seconds. Then followed an interval of three seconds' silence,
whereupon the notes were again struck and held for three seconds,
the periods being timed by a stop watch kept going continuously.

I regret that I had no mechanical means of insuring that all the
intervals should be struck with uniform force. But I may add that,
though by no means an expert musician, I have been informed by
professional musical critics that my touch is accurate and sensitive,
having been trained from early boyhood, and judging from the intro-
spective remarks of all the subjects and the answers to specific questions
of mine addressed to some taken individually, I believe that I was able
to avoid giving any appreciable emphasis to any one note in any of
the intervals, and to preserve a fair uniformity in the loudness of the
intervals \

In some preliminary experiments I used the notes cd^, cd, ce, etc.
up to the octave c'c". But as certain observations of subjects revealed
preferences based on the pitch of the intervals, I decided that a better
arrangement of notes was possible. I was led to distrust the assertion
sometimes made that the pitch of a combination of tones is approxi-
mately that of the lower of its constituents. One feels diffident in
differing from so expert an observer as Stumpf^ but my own intro-
spection, and that of several other individuals specially tested, suggests
that the pitch of the higher note in any combination near the centre of
the piano is likely to be an influential determinant of the apparent
pitch of the combination. I gave to several subjects the test suggested
in this connexion by Stumpf. They were asked whether c'^c or c'c"
differed the more from c. Stumpf asserts that the lower octave differs
more, thus confirming, he says, his assertion that c" gives the pitch of
c'^c'. I put this test to twelve subjects, and nine of them asserted most
emphatically the opposite to what Stumpf says. Eleven of the twelve

1 Even if one note is slightly louder than another in an interval, it does not, according
to Stumpf, affect the consonance of the interval {Beitr. zur Akustik und Musikwissenschaft,
II. 10, quoted in Lalo's Esquisse iVune Aesthetique Musicale). But of course there might
be some difference in the feeling effect of the interval, even though the fusion were not
changed.

2 Tonpsychologie, 2te Aufl. ii. 384.

C. W. Valentine 193

also asserted that ^g' appears higher in pitch than ]fd%, though the
fundamental of the former is two tones lower than that of the latter.
This suggests that the pitch of an interval appears (for most people) to
be nearer that of the higher of its component notes rather than the
lower. It seems to me possible that owing to his exceptional musical
capacity and training there is a much stronger tendency for Stumpf to
refer every interval to its fundamental tone than is the case with the
average individual, thus bringing the fundamental into greater pro-
minence in the field of attention. Hence its greater influence in
determining the apparent pitch of the interval'.

Had I performed all the above tests before conducting my experiments
I might have felt inclined to keep the upper tones of all the intervals
always the same, or I might at least have let them vary in pitch less
than the lower tones. As it was, I resolved upon a compromise,
choosing the following series of notes to represent the various in-
tervals,

minor second, c'A^,

major second, c'd! ,

minor third, }fd ,

major third, }fd%,

fourth, 6b et?,

and so on, alternately raising the upper note or lowering the lower a
semitone until for the octave we get ^g . Thus the mean between the
two notes in any interval would always be c' or cjf, or some note between
these.

It was for the sake of the children in the main that this arrange-
ment was made, for it was surmised, rightly as it proved, that they
would be much more influenced by absolute pitch than would the
adults. That the new arrangement was justified is, I think, shown by
the following fact. As in the case of the adults, the children were asked
to give their reasons for liking or disliking an interval, and a fairly
frequent reason was that the notes were ' nice and high,' or ' too low.'
Now it appears that such votes given by the children for an interval
" because it is nice and high (or low) " are fairly equally scattered over
the various chords, and that judgments against intervals because they
are high (or low) are also similarly scattered. The evidence indeed (in

' Two of the three exceptions who agreed with Stumpf with regard to the first test were
highly trained musicians. With one of these subjects the nature of the intervals compared
seemed to have some influence upon the apparent pitch. Thus c'g' was judged higher than
d'/', but (fj' was judged lower than h^d'jjf.

194

The Appreciation of Musical Intervals

so far as the reasons given by such young children can be trusted) tends
to support the view I have put forward that the higher note is more
influential than the lower in determining the apparent pitch of an
interval. Thus the following table, giving the judgments of children
from eight years to thirteen inclusive, shows that the third hdt is
judged to be ' low ' more often than the octave ^g', and ' high ' less
often than the octave, presumably because the upper note of the third
is lower than the upper note of the octave.

Table I.

Octave, g^g'.

Number of times spoken of as

High Low

Age 13 0 3

„ 12 1 1

„ 11 3 2

„ 10 5 2

„ 9 2 1

„ 8 0 2

Totals 11 11

Major third, fc'^djf.

Number of times spoken of as

High Low

Age 13 1 1

„ 12 0 4

,, 11 2 6

„ 10 3 2

„ 9 0 1

„ 8 2 1

Totals 8 15

Before each sitting the intervals were arranged in a haphazard
order, care only being taken that an interval should not appear often
in the same position, and also that the same succession of two pairs
should not recur. The twelve intervals were then played in the order
arranged, time being given after each for the subjects to write their
introspective remarks at once. After the twelve had been given, some
easy tests of musical capacity such as have been used by Stumpf were
given S and then the twelve intervals were played over again in the
reverse order to the first, in order to distribute equally among the
intervals any effects due to familiarity, etc., and to equalise as far as
possible the effects of contrast due to the chord preceding the one
played. As they were thus arranged in about twenty different orders
for the adults, probably the various effects of contrast were fairly
scattered. The likelihood of such disturbing effects was further lessened

1 The subjects had to say whether one or two notes were being played on the piano,
and which of two successive notes, separated by a tone or semi-tone, was the higher. This
proved so easy for most of the adults that I did not attempt any division of the students
on the basis of the results. Only about half a dozen students could be reckoned aa
' unmusical ' upon the basis of the tests. The same tests were performed on the children ;
for the results see page 210.

C. W. Valentine

195

by the long interval allowed for the writing of introspective remarks.
But observations made by some subjects show that one cannot hope to
get rid of them entirely.

III. The order of popularity of the intervals.

On the basis of the judgments expressed by the subjects, the
intervals can be arranged in order of popularity. In reckoning the
scores of the various intervals the following values were assigned. For
'very pleasing' 2, 'pleasing' 1, 'slightly pleasing' |, 'indifferent' 0,
'slightly displeasing' —\, 'displeasing' —1, 'very displeasing' —2.
This scale gives the following results for all the 146 adult subjects.

Tabi

,E

II.

Major third

... 324

Tritone

153

Minor third

... 261

Fifth

139i

Octave ...

... 246i

Major second...

... - 99

Major sixth

... 243

Minor seventh

... -162

Minor Sixth

... 214

Major seventh

... -316

Fourth

... 157i

Minor second

... -368

Of course we cannot assume from this list that the major third was
the most pleasing and the minor second the least pleasing of the in-
tervals to all subjects. In some cases it was obviously not so. We can
only say that on the average the major third is the most pleasing^.

As will readily be seen, the order is very far from that of degree of
consonance. The major third scores much more highly than the octave,
both the thirds and both the sixths score higher than the more con-
sonant fourth and fifth, and even the tritone — which has been reckoned
on the border line between consonants and dissonants, is found more
pleasing on the average than the fifth — the most consonant of all the
intervals except the octave. The degree of consonance then is by no
means coincident with the degree of pleasingness.

If the votes of men and women are separated we get the following
results :

1 Theoretically it is of course possible that an interval — say the fourth — owes its
intermediate position merely to the fact that some subjects like it most and others dislike
it most. As a matter of fact no interval showed such results. A decreasing number of
' very pleasing' judgments goes with a decreasing number of ' pleasing' and an increasing
number of 'indifferent ' and of ' displeasing' judgments.

196

Tlie App7^eciation of Musical Litervals

Table III.

Men (62 subjects)

Women (84 subjects)

Major third

141^ (187)

Major third

183i

Octave

118^ (148)

Minor third

156^

Major sixth

105 (140)

Major sixth

138

Minor third

104^ (139^)

Octave ...

128

Minor sixth

103i (138)

Minor sixth

llOi

Tritone

67i (91)

Fourth...

93J

Fourth

64 (85)

Tritone

86

Fifth

61i (83)

Fifth ...

78

Major second...

- 41 ( - 55)

Major second

-58

Minor seventh

-67^ (-90)

Minor seventh

-95

Major seventh

-120i (-160)

Major seventh

-196

Minor second

-152 (-202)

Minor second

-216i

The numbers in brackets represent the votes of the men increased
proportionately to make them comparable with those of the women.

In view of the great variations in the preferences of different in-
dividuals, there is no very striking difference between these two orders
with the exception of the greater popularity of the octave and the
minor sixth among the men, and of the minor third among the women.

IV. The aesthetic effect of the different intervals.

The introspective remarks throw some light upon the nature of the
effects produced by the various intervals. We will consider them in
order of their popularity.

Major third. A great variety of reasons are given for liking the
major third. It is described as harmonious, melodious (frequent), some-
thing like the previous chord (which was an octave and was liked), well
balanced, blending, mellow, soothing (frequent), calm, sad (frequent),
solemn, minor, feeling of anticipation, melancholy, firmness tinged with
pleading, strong. Associations with the major third are solemn music,
church bell, Dead March in Saul (frequent), Amen in church. Reasons
for disliking the major third slightly, or finding it indifferent, are the
following : unfinished, feeling of lethargy, slightly too solemn. Dead
March in Saul suggested. The major third was disliked only four
times, in each case by a woman and either because it was too sad, or
meaningless.

Minor third. Very varied reasons are given for liking the interval.
Comments include the following : soothing, mournful, solemn, suggests
Dead March, refined, cheerful, dreamy.

C. W. Valentine 197

It is ouly disliked seven times, the great drop in the score (com-
pared with the major third) being due to the comparative frequency of
the judgment ' indifferent.'

Major and minor intervals. The question has been discussed as to
whether the effects of major and minor keys are ' inherent ' in the in-
tervals themselves, i.e. whether the major chord strikes some essentially
cheerful, responsive note in our nature, the minor rousing equally
'naturally' a sad feeling; or whether the different effects of the two
keys are due merely to association. If the latter we must suppose that
the custom of setting sad songs to minor keys originated without any
felt suitability of the key to the ideas, but that gradually, by repetition
of the association, we have come to connect the two, so that a piece of
music in a minor key now appears to us sad or plaintive. In favour of
the latter view we have the fact that in some civilised countries the
major key is frequently used for sad songs and the minor sometimes for
quite cheerful or even merry ones. Thus we find dance music and even
comic songs set to a minor key. Further, it is asserted that the music
in the minor key played by some primitive peoples, while sounding sad
and dirge-like to us, does not appear to be so to the natives'.

The results of the present series of experiments, as summarised in
the following table, certainly suggest that there is, inherently or through
association, no more sadness in the minor third or minor sixth than in
the major third or major sixth.

Table IV. Number of times interval is described as sad or plaintive.

Major thiid

Men 10

Women 16

Minor third
5
6

Major sixth
16
11

Minor sixth
7
7

Totals 26

11

27

14

The figures show that the major intervals are described as sad or
plaintive twice as often as the minor. Of course we must remember
that we are only testing the effect of one interval, and that, too, with
the notes played simultaneously, whereas in a piece of music in the
minor key we should also have the intervals given by consecutive notes.
Further, their relation to the scale as a whole is brought out more fully
in a piece of music and this is doubtless the most important point in
determining the impression made by the music. The recognition of the
key as minor however is not necessary for the effects of sadness and

1 Cf. Miiller-Freienfels, Psychologic der Kunst, ii. 70.

198 The Appreciation of Musical Intervals

plaintiveness, for these may be felt by persons who are quite ignorant
of the distinction between major and minor keys. It is noteworthy
that five times the major third was actually described as minor while
the minor was never called 'minor.' Probably this particular major
interval was felt to be ' sad ' and was termed ' minor ' because of the
familiar association of the two in music.

Even when a third note is added in these experiments and the
chords ceg and c^g are played, the major chord is still termed sad as
frequently as the minor, though judging from his own introspection, the
present writer is greatly surprised at this result. Thus among about
forty adults to whom these chords were played (among twenty other
chords), eight persons described the major chord and six persons the
minor chord as sad\

The evidence, then, of these experiments is that the minor intervals
(when the notes are played simultaneously) are not felt as sad even to
the same extent as the major intervals. And this is in favour of the
view that the general significance of the minor key for modern European
ears is not due to an effect inherent in the relation of the notes in a
minor interval, but is more probably the effect of association. Further,
for the average person it seems that more is necessary as a basis for this
association than isolated minor intervals or chords.

The octave. The octave is often termed indifferent and is disliked
more frequently than the minor third. It is found by some too tame,
dull, thin and skimpy, lacking in meaning, whilst others are attracted
by its clearness, smoothness, idea of no hindrances, or describe it as bold,
bright, strong and cheerful, giving the feeling of rest.

We saw that the octave was considerably more pleasing to the
men than to the women, and we find that only one man among 62
judged it positively displeasing, while 15 women among 84 do so. The
introspective remarks give us a clue to an interesting sex difference

1 In these trichord (Dreikldnge) experiments the chords had always c' as the tonic.
This fact precludes an explanation that might conceivably be put forward as to why the
major is described as sad, or even as minor, in the interval experiments, viz. that some-
times the influence of the preceding interval might determine that a major interval should
appear as part of a minor scale. For example, when aVgV is followed by cveP, if the tonic
aV were still held in mind when c/et7 was played (which would be after an interval of 2 or
3 minutes) and if cvev were then heard as in the key of a^, the cV might give, with the
retained impression of the oY, the impression of the minor third of the scale aP. This
seems to me extremely unlikely in view of the long time between the playing of the intervals
and also in view of the fact that very few of my subjects were even average musicians.
Further the major third usually followed intervals which could not produce such an effect.
In any case, the experiments with the trichords were free from even this remote possibility.

C. W. Valentine 199

here. Whereas the men frequently like it because it is strong, firm,
bold, suggesting majesty and force, this aspect of the octave does not
appeal in the same way to the women ; it is even displeasing to some
of them, who speak of it as too assertive, hard, or harsh.

The major and minor sixths. Coming to the next intervals in order
of preference, one is immediately struck with the enormous individual
differences which now show themselves. Thus of four persons listening
to the major sixth at the same time, one speaks of it as ' soothing,' the
second as * rousing,' the third as * sentimental,' and the fourth even as
'jarring to the ear.' This is generally characteristic also of the fourth,
tritone and fifth. A distinct tendency however is noticeable for the
sixths to be felt as sad or solemn. Here for the first time we find
suggestions of disharmony, half a dozen or more subjects finding that
the notes do not blend satisfactorily, a remark which applies also to the
fourth, fifth and tritone.

Fourth and fifth. The low position of these intervals is not
traceable, on the basis of the introspective remarks, to any definite
disagreeableness in them. The average value of the votes, it will be
seen, yields for each of them nearly 0*5, the equivalent of * slightly
pleasing.' But they rarely become 'very pleasing' and are often judged
'indifferent,' sometimes with the description 'ordinary,' 'no impression.'
It is very remarkable that these — the three most consonant intervals
after the octave — are sometimes spoken of as discords, or as lacking in
harmony, oftener indeed than the tritone'. Possibly we have here a
suggestion that the conventional concord may come to appear less con-
sonant by becoming for some reason very unpleasant (perhaps from
appearing, first commonplace, and then monotonous). On the other
hand there is ample introspective evidence that dissonant intervals,
where pleasing, are sometimes felt as consonant. This is probably the
case when they are introduced appropriately in musical compositions.
But the above statement is not confined to the cases where the discords
are heard as leading to a pleasant resolution. The impression occurs
too when they are heard alone.

As to the four discords we may further remark that any one of
them may appear pleasing through some definite association or symbolic
suggestion, and that sometimes discords are liked as a pleasant change
from a preceding harmonious interval.

^ It is possible that some of these judgments may be due to the fact that the fourth
and fifth are somewhat out of tune as sounded on the piano. But at least oue subject who
repeatedly gave such judgments was distinctly weak in detecting dissonances.

200 The Appreciation of Musical Intervals

V. The method of experiment with school children.

These experiments were performed on one hundred and ninety-five
boys and girls from two Elementary Schools in Dundee, between the
ages of six and fourteen, and upon seventy-six girls between the same
ages in a high class Preparatory School in St Andrews^.

The object, as already stated, was to discover if possible something
as to the development with age of a feeling for consonance, and the
difference in this respect between children belonging to different cultural
groups and differing with respect to musical training. There are, of
course, enormous individual differences even among children as regards
the sensitivity to music. By means of these tests one could only hope
to study the averages of a large number of children groups of which
differed in culture and age.

The method of procedure was the same as in the experiments with
adults, with the following exceptions. I played all the intervals over
to the children before the judgments on each interval were given, in
order to show them the kind of test they were to expect. During the
latter half of the experiments I also added a new interval (ninth) with
which I always began the list, though the judgments on it were ignored.

In the case of the Elementary School children I had the cooperation
of some of my own students, partly because many of the children were
too young to be trusted to make satisfactory written records, and partly
in order that the students might have some experience of research
work. The Elementary School children were taken in twelve groups
of about fifteen pupils, of various ages, at a time. Thus the intervals
would be played in twenty-four different orders in the course of the
experiments, a new arrangement being made for each group, and the
exact reverse of it also being used at each sitting.

The children were distributed over a large room, and a child was
allotted to each student, who recorded the child's judgments. All the
students had had some training in experimental psychology. The
method and purpose of the experiments were carefully explained to
them beforehand. I impressed upon them especially the supreme im-
portance of giving no sign of approval or disapproval of the child's
judgments, and of avoiding any possible suggestion. At each test
I explained the experiment to the children somewhat as follows. "I
am going to ask you to listen carefully while I play some notes on the

1 In a few cases the same pupil returned for the test again a year or so later. In these
cases they are reckoned twice in counting the number of children.

i

C. W. Valentine 201

piano, and then to say whether you like them or not. If you can, say
why you like them or don't like them. I want you to say exactly what
you think. No one will see your answers except myself. You are not
in school now, and no one will blame you for what you say." A little
friendly talk soon seemed to put the children at their ease. The
students were told to say nothing that was not absolutely necessary —
which practically reduced itself to asking ' why ' of those children who
seemed to need encouragement to state their reasons. As I had all the
students in full view I was able to see that this rule was remembered.
We know from our experiments on the adults the average order of
preference for the intervals with the students. Thus if there were any
influence of suggestion on the part of the students we know in Avhich
direction it is likely to work. Personally I believe that in these experi-
ments such influences were extremely small.

Such collective experiments have grave dangers and disadvantages
unless very carefully carried out ; but apart from the fact that they
enable one to examine far more subjects than would otherwise be
possible, they may, I think, have two advantages. In the first place
the children seem more at their ease when a large group of them is
examined simultaneously than when only one child is tested at a time.
Further, each group of children included children of various ages from
six to thirteen, and both boys and girls. Thus if any irregularity did
occur (as for example if an interval were played somewhat more softly
than usual), its effect, if any, would be distributed over children of all
ages.

A more serious difficulty was to prevent the children from being
influenced by what they heard other children saying. In view of this
they were separated as far as possible, and were instructed to whisper
and to say ' I like it ' or ' I don't like it ' instead of ' yes ' or ' no,' for
'yes' was likely to be heard by others near, even when whispered.
Generally this rule was obeyed well, but very occasionally a faint ' yes ^
was audible. Here again however we know something of the likelihood
of the effects of suggestion, if such there were. Seeing that the
youngest children are more subject to suggestion than the older
ones, they would be more likely to adopt the answers of their seniors
than vice versa, and thus would tend to raise the apparent degree of
development of the juniors. As will be seen, this possibility only makes
some of the later observations more significant. The question as to
whether the youngest children completely understood what they had to
do will be discussed later. I may say here that though one or two of the

J. of Psych. VI 14

202

The Ajjjfreciation of Musical Intervals

youngest children did not give any answer with respect to some of the
intervals, none of the answers given were foolish in the sense of being
inapplicable, i.e. they were always judgments as to whether they liked
the notes or not. In very few cases was the answer ' I don't know '
given — indeed it might have been more encouraging if this answer had
been given more often.

VI. Results of the experiments with the Elementary School children.

An interval scored + 1 when the judgment ' I like it ' was given,
— 1 for the judgment ' I don't like it.'

In Table V are given the votes of the children of various ages for
the different intervals. As the numbers of children of the various ages
were unequal, the votes are adjusted to make them represent pro-
portionately the judgments of thirty children of each age, for the sake
of easy comparison. The intervals are arranged in their order of
preference, as determined by adults. For the sake of comparison a
column is added showing the votes given to the intervals by the
146 adults, reduced to represent the votes of thirty individuals.

Table V. Shoiuing average votes for thirty individuals of all ages.

As each child judged each interval twice, the highest possible score for any interval is
+ 60 and the lowest possible - 60. But these figures do not apply to the adults owing to
greater variations of judgments which were permitted to them [e.g. very pleasing +2, very
displeasing - 2, slightly pleasing +i, and so on].

Age in years

6

7

8

9

10

11

12

13

Adults

Actual no. of children ...

15

24

25

27

27

24

22

31

Major third

36
38
26
20
32
34
28
22
32
24
28
16

37
35
29
22
36
35
40
41
42
30
35
30

44
35
37
35
36
26
28
35
20
40
26
20

36
42
42
42
29
36
33
22
16
28
21
4

33

38

31

31

33

36

23

29

0

18

9

4

41

36

42

40

25

37

17

35

9

0

-20

- 5

35

20

22

28

28

25

5

0

-11

3

-20

-35

37

25

31

32

37

-2

12

11

-5

2

-25

-32

36

Minor 'third

Octave

29
27

Major sixth

27

Minor sixth

24

Fourth

17

Tritone

17

Fifth

15

Major second

-11

Minor seventh

Major seventh

-18
-35

Minor second

-41

Av. score of 8 concords. . .
,, ,, 4 discords...

27-0
25-0

33*1
35-6

34-5
26-5

35-0
17-0

31-7
9-2

33-5
-40

20-4
-15-8

22-9
-150

240
-26-2

C. W. Valentine 203

An inspection of Table V shows that there is no consistent prefer-
ences for consonances over dissonances at the ages of six and seven.
With the six-year-olds two of the discords are liked more than the
octave, and three of them are liked more than the fifth or the major sixth.
At seven years all the discords are liked better than the octave or
minor sixth.

Of course one cannot infer that there is no capacity at this age to
discriminate between a discord and a concord. There may be a differ-
ence of sense experiences in the case of these children which corresponds
to the different experience we have in the consonances and dissonances.
It is more than likely that the well-known love of children for noise of
any kind keeps in the background any tendency to dislike a discord as
such^ Indeed, their love of sound for its own sake may act in favour of
discords. For as Stumpf has shown, dissonant intervals often appear to
children as containing three, four or even five notes, apparently giving
a greater body of sound, and even to adults they are generally more
stimulating, in a sensational way, than are consonant intervals.

We can, however, at least infer that at this age these children had
not such unpleasant sensations produced by discords as to diminish
appreciably their pleasure in the sound as sound.

There remains the difficulty as to whether the children really
understood what they had to do. But surely the question ' Do you
like that ? ' should be intelligible to the average child of six or seven
if there is any definite feeling of pleasure or displeasure produced.
Where there is precocious development of musical sensitivity, children
are able at an even earlier age than this to express very definite and
decided judgments upon intervals and chords, as in the case quoted by
Stumpf, of a five-year-old boy who could ' sing seconds ' to a melody
with ease, and who always gave an immediate judgment in favour of
the major trichord as compared with the minor, whichever was played
first-. The fact that occasionally a child would say ' I like it ' (or still
more rarely ' 1 don't like it ') to almost every interval, might appear to
indicate a lack of comprehension as to what was being done ; and this
may have been the case with a few of the least intelligent pupils. But
it seems quite possible that even in these cases the uniformity of
judgments signifies an inability to appreciate the contrast between

1 Cf. Dr Myers's remarks upon the dangers of inferring that some primitive tribes have
no feeling for consonance because they disregard it in their music. " The Ethnological
Study of Music," in Anthropological Essays presented to E. B. Tylor, Oxford, 1907, 239.

2 Tonpgyclwloyie , ii. 378.

14—2

I

204 The Ajjpreciation of Musical Intervals

consonances and dissonances; for series of judgments, all of them
'pleasing,' were also given by children of ten, eleven, twelve and
thirteen years, and such series can hardly be ascribed to incapacity
to understand at this age.

Nor do the more intelligent children show any greater antipathy to
the discords than do the less intelligent'. Unfortunately the numbers
are very small when one comes to divide the children of each year into
intelligent and unintelligent. But I may state that the average score
for the four discords among twenty-two 'intelligent' children of six or
seven years of age was 0"5, that of the concords being 0'6-. For eight
comparatively unintelligent children the scores were as follows : discords
0*65, concords 0*7. Among the eight-year-olds, indeed, it was the thirteen
unintelligent children who showed a preference for the concords, the
seven intelligent ones showing a slight preference for the discords.
Thus there is no evidence that, with these children, capacity to
understand the simple requests made of them was the main factor
in determining the trend of the scores, or, indeed, that it had any
such influence whatever^.

If we take the children individually, we find that none of the six-
and seven-year-olds give even one judgment of dislike to each of the
four discords without also giving such judgments freely among the
concords, including the octave. At eight years, however, we find two
children who only give three of eight possible votes of approval to the
discords, but give respectively fourteen and thirteen (of sixteen possible)
to the concords.

At nine years of age we find a great advance. At least seven children.

' My best thanks are due to Mr J. Williamson, Headmaster of South Tay Street School,
Dundee, for the care he took in personally classifying all the children in his school who
underwent these tests. In cooperation with the class teachers he divided them into three
groups, 'intelligent,' 'moderate,' and 'weak.' Owing to the smallness of the last group
they are reckoned with the ' moderates.' I did not obtain such a classification from the
other school from which the children were taken, chiefly because of the impossibility of
making sure that the same standard of intelligence was being used as a basis of division ;
moreover, the number from the second school was too small to make separate calculation
of any value.

'^ The proportions of intelligent and unintelligent children were approximately equal
for both these ages.

3 There is on the other hand no absolute proof to the contrary : it might still be the
case that the task was so much above the powers even of the most intelligent eight-year-old
children that even their superior intelligence would not be of any value to them or cause
any contrast between their results and those of the less intelligent. But this seems highly
improbable.

C. W. Valentine 205

out of twenty-seven, showed a marked and consistent preference for the
concords before the discords. Their totals are :

' Pleasing' judgments given upon four discords, 16
,, „ „ „ eight concords, 93.

This great advance at the age of nine is also reflected in the totals for
all the nine-year-olds, the average score of the discords being now only
one-half that of the concords.

Summing up, then, we may say that, on the evidence of 936 votes,
no preference for concords before discords is shown by the six- and
seven-year-old children ; that a slight preference for the concords begins
to appear among a few of the eight-year-old children, calculated on a
basis of 600 votes ; and that at nine years old the preference for concords
is decided. It is interesting to note that Mr J. A. Gilbert as a result of
his experiments on " The Musical Sensitiveness of School Children V' con-
cluded that in the discrimination of tones varying in pitch, the average
school child improves more than twice as fast from six to nine years, as
it does in the years from nine to nineteen.

At the age of ten the discords become still less pleasing, and at
eleven they have a negative score for the first time, though the score
of the concords is practically as high as ever. At twelve and thirteen
the children become much more critical in their attitude both towards
the discords and to those concords which were found to be least liked
by the adults, viz. the fourth, the tritone and the fifth. In regard to
intervals about which adults have been seen to differ so much in their
judgments, one cannot expect a perfect regularity of change in the votes
of the children with increasing age. But on the whole we may say,
judging from Table V, that no marked and continued preference is
shown for any of the concords, and none of them is conspicuously less
pleasing than the average, until we come to the ages of ten and eleven,
where the lesser consonance of the tritone seems to have its effect upon
the children. At twelve and thirteen, however, we suddenly find a
comparative indifference to the highly consonant fifth and fourth,
similar to what was found with adults. This is shown in Table VI,
where the scores of six- and seven-year-old children are added together,
and also those of eight and nine and so on, and then reduced to repre-
sent the votes of thirty children. The votes of the men and women are
also given, adjusted so as to make the votes given for the most pleasing
interval approximately the same as the children's vote.

1 Studies from the Yale Psychological Laboratory, 1892-3.

206 The Appreciation of Musical Intervals

Table VI. Showing average votes for thirty individuals.

Ages

6 and 7

8 and 9

10 and 11

12 and 13

Men

Women

Major third

36
36
27
-21
34
34
34
31
37
27
31
23

40
38
39
38
32
31
30
28
18
34
23
12

37

37

36

35

29

36

20

32

7

9

-5

C

36
22

2»;

30

32

11

8

5

-8

2

-22

-33

35

26

30

26

26

16

17

15

-10

-17

-30

-38

37

Minor third

31

Octave

26

Major sixth

28

Minor sixth

Fourth

22
19

Tritone

17

Fifth

16

Major secoud

Minor seventh

Major seventh

Minor second

-12
-19
-39
-43

By the age of twelve or thirteen then, these children have reached a
stage at which their preferences for the various intervals are remarkably
like those of the men and women. The major and minor thirds, the
octave, and the two sixths form the ' most pleasing ' group, both for
these children and for the adults, the major third leading in all three
columns. The fourth, fifth and tritone form a group of ' less pleasing '
(but still ' pleasing') intervals, while the discords all have minus scores
except the minor seventh which just secures + 2. It is curious that
this interval is the most pleasing (or least displeasing) of the discords
for children from the age at which any discrimination between concords
and discords takes place, while for the adults the major second is preferred
of the four discords.

VII. Residts of experiments on Preparatory School children.

Table VII shows the results of the experiments in the girls' Pre-
paratory School. These children wrote their own judgments on paper,
with the exception of one or two of the very youngest, for each of whom
one of the teachers did the writing. They were taken in four groups
of about eighteen each, including children of various ages. In a few
cases the children modified their judgments by ' very,' etc., saying ' very
pleasing ' (or ' displeasing ') or ' slightly pleasing ' (or ' displeasing '), for
which the scores of IJ (or — 1|) and | (or — ^) were awarded. These
children took a very keen interest in the tests, and most of them, even
the very youngest, seemed very decided in their judgments.

C. W. Valentine

207

Table VII, Preparatory School Results. Showing average votes
for thirty individuals of all ages.

Ages

6&7

8

9

10

11

12

13 & 14

Actual DO. of children...

7

10

11

10

9

15

14

Major third

45

57

52

43i
54

56

56

43

Minor third

45

40

30

53

51

44

Octave

4

45

14

49i

4H

50

52

Major sixth

17

45

3(5

33

51i

40

43

Minor sixth

0

42

18

39

41i

45

25

Fourth

40

42

3

18

7i

30

26

Tritone

-13
10

33
39

17

21

6

24^
8

26
38

11

Fifth

Ih

Major second

-34

-18

-47

-36

-50

-10

-43

Minor seventh

6

24

-44

-36

-23

-36

-36

Major seventh

-38

-24

-58

-42

-60

-58

-53

Minor second

-21

-45

-60

-54

-6U

-60

-58

N.B. As these children had permission to say 'very pleasing' and ' very displeasing"
(scoring + 1^ and - 1 J respectively) the maximum and minimum scores for each interval
are + 90 and - 90 respectively, instead of + 60 and - 60, as was the case with the
elementary children.

The scores are raised to represent proportionate numbers for thirty
children in each column for the sake of comparison between the different
ages and with Table V.

It will be seen at once that we have strikingly different results from
those given in Table V. At eight years of age, and even at six and
seven, the discords are already discriminated and all have minus scores,
with the exception again of the minor seventh, which, as with the
Elementary School children, retains its power to please longer than the
other discords. By the age of nine we have preferences very similar to
those of the adults, and such as are not given by the Elementary School
children before the ages of twelve and thirteen, viz. the major third
leading, the fourth and fifth low, and all the discords with negative
scores.

Of course, seeing that the intervals were only presented twice to
each subject, the numbers of children are very small on which to base
any conclusion, and some variations occur which we might expect under
the circumstances. From the first, however, a marked antipathy to the
discords is shown. Possibly for a thoroughly reliable average we require
a number given by the addition of not less than three of the columns of

208 The Ajypreciation of Musical Intervals

Table VII. But an order of preference remarkably like that of the
adults is given either by adding the eight- and nine-year columns or
those of the nine- and ton-year-olds, as is shown in Table VIII.

Children of 8 or 9 years.

1.

Major third.

2.

Major sixth.

3.

Minor third.

4.

I Minor sixth.

1 Octave.

6.

Tritone.

7.

Fourth.

8.

Fifth.

9.

Minor seventh

10.

Major second.

11.

Major seventh.

12.

Minor second.

Table VIII.

Adults.

1.

Major third.

2.

Minor third.

3.

Octave.

4.

Major sixth.

5.

Minor sixth.

6.

Fourth.

7.

Tritone.

8.

Fifth.

9.

Major second.

10.

Minor seventh

11.

Major seventh

12.

Minor second.

Children of 9 or 10 years.

1. Major third.

2. Minor third.

3. Octave.

4. Minor sixth.

5. Major sixth.

6. Tritone.

7. Fourth.

8. Fifth.

9. Minor seventh.

10. Major second.

11. Major seventh.

12. Minor second.

VIII. Comparison of the results of the experiments in the Elementary
and Preparatory Schools.

Table VIII shows that the Preparatory School children at about the
age of nine show a resemblance to the adults in their liking for the
different intervals, which is only shown at the age of thirteen by the
Elementary School children.

This great difference must be attributed presumably to one or more
of the following grounds : (i) greater intelligence, (ii) greater inherited
sensitivity to music, (iii) much earlier and more thorough instruction in
music and closer familiarity with good music. We have good grounds for
disbelieving that general intelligence has much to do with the results of
these music tests. The absence of any consistent difference between the
judgments of the more intelligent and the less intelligent of the six-,
seven- and eight-year-old children in the Elementary School has already
been mentioned (p. 204). Among the older children also there is absolutely
no regular tendency for the intelligent children to approximate to the
adult standard any more than the unintelligent do.

Doubtless, then, one or both of the last two causes mentioned are
responsible for the difference between the Elementary and Preparatory
School children. Unfortunately these experiments do not afford decisive
evidence as to which is the chief cause. After the age of seven
practically every girl in the Preparatory School learned some musical

C. W. Valentine 209

instrument, and they all often heard good music^ Many of the older
children had been learning music for live or six years; while among the
Elementary School children only four boys and eleven girls had had any
lessons on the piano, and two (girls) on the violin. Thus the difference
in musical training among the two sets of children is enormous. As
to how far this is accompanied by greater inherited musical capacity
it is difficult to say, but it seems to me probable that this latter has
comparatively little to do with the observed differences between the two
types of schools. The comparative ease with which the children of the
upper classes, as compared with those of the lower, take to music, even
if clearly demonstrated, would not help us in deciding the relative
importance of heredity and training, for the former have already been
more accustomed to hear good music than the latter.

Both Prof Spearman and Mr Burt found that children of more
cultured families had greater powers of pitch discrimination than those
of less cultured families, the average thresholds in two Preparatory
Schools being little more than half those attained from Elementary
School children, a difference which it was shown could not be attributed
to practiced

We are, I believe, ignorant of the extent to which sensitivity of
pitch discrimination affects the aesthetic appreciation of consonance.
But we shall see later (Table IX) that those Elementary School children
who did well in two of Stumpf's tests for a 'musical ear' (cf. p. 210)
reached a stage of development equivalent to that of the adults at an
earlier age than those who did badly in these tests. At least we can
assume that a fairly keen discrimination both of pitch and of con-
sonances from dissonances is necessary for a degi-ee of general musical
ability much above the average. Thus they would be selected together
in at least one of the ways in which we may suppose the average musical

' Of the seven children under eight years of age, three (two of seven years of age and
one of six years) had not studied music. They showed scarcely any preference for concords
over discords. The four who had studied music (three of seven years and one of six years)
showed much more discrimination, the scores giving an average of nearly 0*5 per vote for
the concords, and - 0*2 for the discords. But these numbers are of course far too small
to base any inference upon. Apart from this there is doubtless a tendency for the
precociously musical child to begin lessons earlier than the others. From the eight-year-
old son of a University Professor, well above the average in intelligence, and who had had
no instruction in music, I was able to obtain four judgments upon each of the intervals.
He showed practically no preference on the whole for concords (average vote 0-6) to discords
(average vote 0-5). There was however a marked preference for the major and minor
sevenths (average of each -f- 1-0, i.e. ' pleasing ') to the major and minor seconds (average
- -25 and 4- '25 respectively).
2 See this Journal, iii. 125.

210 The Appreciation of Musical Intervals

ability of the upper-middle classes to have been increased or maintained,
and that of the lower classes comparatively lessened, namely by the
selection and raising to a somewhat higher social standing of persons of
marked musical ability in the lower-middle or working classes. Further
speculation on this point is beyond the scope of this paper. But granted
some difference in the musical sensitivity of the two social groups, few
will doubt that their music.il training was an important factor in
determining the rapid advance in the Preparatory School children in
the capacity to appreciate the difference between consonances and
dissonances. One possible criticism must be dealt with. It may be
suggested that to the young children who had had musical instruction
the discords may have sounded ' wrong ' without having any of the
unpleasantness naturally associated with dissonance. But that implies
a capacity for remembering and identifying an interval which the music
mistress had frequently corrected, a capacity which is surely more
surprising and improbable in young children than the feeling of the
unpleasantness of a discord. Even a musical adult in these experiments
mistook a major second, which he found pleasing, for one of the thirds,
and was doubtful whether a pleasing minor seventh was not an octave.
Nor do the reasons given by the children give any support to this
suggestion. The only one who speaks of the notes 'sounding right'
is an Elementary School boy of seven years who had not had music
lessons.

IX. Tests for a 'musical ear.'

At each sitting, after the playing of all the intervals and before they
were repeated in the opposite order, other simple musical tests were
given, similar to those used by Stumpf, for discriminating musical from
unmusical individuals. In test A the children were asked to say whether
I was playing one or two notes on the piano. Three times two notes
were played (the octave, major third and tritone), three single notes
being interspersed. In test B, the children had to say whether the
second of two notes, played successively, was higher or lower than the
first. Three times the second note was higher, and three times lower
than the first. In four cases the notes differed by a semitone and in
two cases by a full tone. I first explained what higher and lower meant
by playing a succession of ascending notes on the piano and saying
that those were 'getting higher,' and similarly with a succession of
descending notes. As test A is appreciably harder than B, especially

C. W. Valentine 211

to children unfamiliar with the piano, mistakes in A were only reckoned
as half en-ors, those in B counting as full errors. In order to divide the
children roughly into two equal groups, and to allow fully for 'slips'
and misunderstandings, only those whose total errors were more than
two were classed as ' unmusical.' The number of boys taking part in
the experiments was 95, the girls numbering 100. Of these 37 boys
and 43 girls were ' musical,' numbers that do not allow us to infer that
either sex, at this age, is more ' musical,' as far as these tests can
indicate.

No marked correlation was observable between success in these
musical tests and general intelligence, a result which encourages the
belief that the task was explained clearly enough even for the duller
children to understand. The numbers were as follows :

Intelligent children. Unintelligent children.

31 musical, 42 unmusical. 26 musical, 38 unmusical.

Table IX shows the connexion between success in these tests, and
the votes for the various intervals. As there are only four musical
children of seven years we will ignore that column. An examination
of the others shows that among the younger children (ages eight and
nine) the musical ones are more critical throughout, but not more
averse, ^proportionately, to discords than are the unmusical children.
At ten years the musical children are slightly less critical both towards
concords and discords.

Thus up to the age of eleven or thereabouts, the greater sensitivity
of ear, as shown by the musical tests, does not result in greater dis-
crimination of concord and discord. There is, on the average, a slightly
more critical attitude towards all intervals, possibly the beginnings of an
attitude in which the notes are no longer pleasing merely because they
are sounds, an attitude perhaps more readily adopted by the child pos-
sessed of a more sensitive hearing (in a musical sense). But after eleven
years the more 'musical' children are emphatically more averse to discords
than are the ' unmusical,' though at least as appreciative of concords.
The total scores for the discords (for thirty children) are 15*3 for the
unmusical, but - 89 for the musical children over eleven years of age.
This marked correlation must mean either that a musical ear (as tried
by Stumpf's tests) does eventually reveal itself also in greater sensitivity
to the unpleasantness of discords, or that the two capacities, though not
directly connected, are generally found together. No doubt familiarity
with the piano would be one factor in determining success in Stumpf's

212

The Ajyj^reciation of Musical Intervals

Table IX. Votes of ' musical' and 'unmusical' children.

Ages

13

12

11

10

9

8

7

, — 1

, -,

,-^

1

..^

-3

.2

"3

"oa

03

1

.2

"cS

1

"oS

"S

m

o

CO

01

S3

a

S

a

s

3

-s

3

s

s

p

a

3

s

0

B

a

a

P

a

0

s

s ■

a

S

a

S

a

^

a

S

a

^

a

s

a

P

P

D

^

14

P

&

P

No. of children

22

9

10

14

14

10

10

16

13

7

18

4

20

Major third . . .

32

6

14

12

23

10

12

18

10

22

8

29

6

24

Minor third ...

22

4

10

4

16

18

8

26

16

22

7

22

4

24

Octave

19
27

13

7

4

8

12

12

24
16

10
16

16
12

16
16

18
16

20
22

7
3

24

26

2
2

21

Major sixth

16

Minor sixth

24

14

2

18

12

8

12

18

10

16

8

22

2

27

Fourth

-2

0

6

12

16

14

14

18

10

22

8

14

6

22

Tritone

11

4
-10

1

7
5

0

2

- 6

4

- 2

- 2

10

20

- 9

4

8

16

11
8
2

10

18

2

16
10

8

14

10

6

5

8

-1

18
21
18

6
6
8

26

Fifth

27

Major second...

26

Minor seventh..

0

2

- 8

10

- 6

6

8

8

6

19

7

26

6

18

Major seventh..

-22

-4

-10

- 4

-18

2

4

4

12

7

8

14

2

22

Minor second...

-28

-5

-10

-16

- 6

2

2

2

0

4

5

12

4

20

Total votes, adjusted to represent votes of 10 children —

8 concords
4 di.scords

61-81
- 27-2

57-7
-2-2

The musical

children are

much more

averse to

discords

46
34

51-3 I 97-81
-8-5 -27-8

83
26

93
16

87-5
10

75-7
18-5

113-8
27-6

The musical

children are

much more

averse to

discords

The musical

children are

much more

averse to

discords

No very

decided

difference

The musical

children are

more critical

towards all

intervals

77-1 I 103-3
27-1 I 38-8

The musical

children are

more critical

towards all

intervals

tests*. If SO, we should expect to find this correlation, on the assumption
which we have already made, that the aesthetic appreciation of conson-
ance is also developed by familiarity. But it is difficult to explain the
marked correlation among the older children by this familiarity with
the instrument alone, for nearly all these Elementary School children
gain their knowledge of pianoforte music largely from the school, where
they have equal opportunities. Very few would have a piano in the
home, and as we saw only 17 out of 195 were having music lessons.

It seems highly probable, then, that the musical sensitivity, as tried
by Stumpf's tests, is closely connected with the aesthetic discrimination
between consonance and dissonance, both doubtless being cultivated by

1 The same tests were given to the Preparatory School girls but only 7 of the 76 made
a total of more than two errors, and 3 of these were only 7 years old. All of them, except
one seven -year- old, were as averse to the discords as were the 'musical' children.

C. W. Valentine 213

familiarity with music. The average improvement in Stumpfs tests is
especially rapid up to, and including, the age of nine, as Gilbert found
was the case with tone discrimination. That it is not accompanied by
greater discrimination between concords and discords until a year or so
later may signify that further musical experience is necessary before the
improved capacity for discrimination (such as is involved in Stumpfs
tests) can be brought to bear on the higher aesthetic appreciation of
consonance or dissonance^.

X. Introspection of school children.

(«) Eleirientary School children.

Naturally the records were not rich in introspective remarks, although
the children were asked to give their reasons for liking or disliking the
intervals. Such terms as ' sweet,' ' good,' ' nice ' were common with the
youngest children. The first indications of the hearer's attention being
attracted to the effect of the notes upon himself occur at the age of nine
and are then concerned wholly with unpleasant effects, e.g. ' makes my
head ache,' * makes an awful sound in my ears,' ' makes my ears ring.'
At the ages of eleven, twelve and thirteen we find more frequent
judgments of this kind, and no longer confined to unpleasant effects,
thus : ' pleasant to the ear,' ' soothing,' ' grating,' ' nice for singing to.'

Children over the age of eleven also make the comments ' bold,'
' cheery,' ' strong,' ' awful sad ' and ' uncanny.' Frequent references are
made to the pitch of the notes as the following list shows.

Intervals said to be

A B C

Pleasing because high, I Pleasing because low, I Displeasing because too low,
by 15 boys and 15 girls | by 4 boys and 4 girls | by 20 boys and 24 girls

Six of the eight children who find intervals pleasing because ' low '
are over twelve years of age. In column A, twenty-three of the thirty
children were over ten years of age. In column C, thirty-nine of the
forty-four children were over ten, the numbers being fairly equally
distributed (in the case of both colunms A and C) among the various
ages from ten to thirteen. Thus the height of the pitch does not appear
to effect the younger children any more than the older ones.

1 I do not think that anything said Ijere is inconsistent with Mr T. H. Pear's criticism
of these tests as adequate tests of general musical capacity (this Journal, iv. 89), with
which I fully agree. Even the addition of a test for the appreciation of consonance and
dissonance would not make a satisfactory test of the capacity to enjoy music.

214 The Aj)2)reciatioii of Musical Intervals

Some increase after the age of nine in the number of children who
refer to height or lowness of pitch is only to be expected, as the children
become more able, with increasing age, to give reasons for their likes
and dislikes. The answers as a whole indicate that somewhat higher
intervals would have been more pleasing to the children. The pitch
used would be low for the children to sing, and some introspective
remarks suggest that this is of considerable weight in determining the
pleasingness of a note to a child.

The occurrence of associations is peculiar. They occur far more
frequently about the ages of ten and eleven than cither before or after
those ages. Thus 106 associations out of 165 are given by children
of ten or eleven years of age, i.e. nearly two-thirds of thp associations
are given by about one quarter of the children. Twenty-three boys
give associations, but only thirteen girls, though the average number
of associations given by these thirteen is higher than that of the
twenty-three boys^ Only two girls but ten boys under the age of
ten give associations. These facts may indicate a gi-eater interest on
the part of young boys (compared with girls) in the sounds of objects
suggested by the intervals, e.g. bells, clocks, motor horns, and one
delightful expression of discord on the part of a small boy, " like a
man smashing a tin can and he can't smash it any more."

(6) Preparatory School children.

Here the most striking fact is the absence of associations. In the
1824 judgments given by these children only one association occurred,
'like a bugle.' The principal cause undoubtedly lies in the high degree
of musical training of these children. They never think of associating
such sounds with anything but a piano, an instrument they know so
well.

XI. Sex differences in the Elementary School experiments.

We have already noted the more frequent occurrence of associations
among the boys. Another marked difference is revealed by the total
votes for and against the intervals. The boys are much more critical,
the girls much more disposed to say ' I like it.' Thus the eight concords
receive a balance of positive votes to the number of 879 from the 100 girls,
but only 688 from the 95 boys. This characteristic difference is shown
through all the ages. Table X shows the actual number of votes for the
various ages.

' Chiefly owing to the fact that one girl gave fourteen associations.

C. W. Valentine

•215

Table X.

Ages 6 & 7

Ages 8 (fe 9

Ages 10 .fe 11 Ages 12 & 13

All ages

17

boys

22

girls

28
boys

24

girls

24

boys

i

26 26

girls 1 boys

28
girls

95
boys

100
girls

Balance of votes)
for 8 concords 1

Balance of votes/
for 4 discords \

138
66

202
94

224
51

260
100

197
27

252 129

-8 -68

165
-50

688
76

879
136

In each column the boys and girls are distributed between the two ages (given at the
head of the column) in similar proportions, except in the first column where there are 14
girls of seven j ears to 8 of six years, and only 10 boys of seven years to 7 of six years.

The totals for all ages suggest that the boys as a whole discriminate
or at least dislike the discords more than do the girls. But the difference
is almost entirely traceable to the ages of eight and nine where it is very
marked.

XII. Summary of results and conclusions.

I. The apparent pitch of an interval is for most people determined
approximately by the pitch of its higher note, and not of its lower note
as has been previously asserted.

II. Of all the intervals used the major third was by far the most
liked by adults. Then comes a group of four, viz. minor third, octave,
major and minor sixth, which on the average are found ' pleasing.' Then
a third group, the fourth, tritone and fifth, each on the average ' slightly
pleasing.' This order, it will be observed, is by no means coincident
with the order of degree of consonance.

III. The major third and major sixth are described as sad by adults
twice as often as the minor third and minor sixth. This supports the
view that the usual effects of the minor key for modern European ears
are not due to any ' natural ' effect of the minor intervals, but are deter-
mined by association.

IV. Even the highly consonant intervals of the fifth and fourth are
sometimes described (by adults) as discordant.

V. Among the children in the Elementary Schools tested no ap-
preciable preference for concords before discords is discernible before the
average age of nine, at which age a considerable advance takes place.

216 The Ap2)reciation of Musical Intervals

VI. A group of" children ot" the age of twelve or thirteen gives an
order of preference for the twelve intervals within the octave, which is
remarkably like that given by adults.

VII. No appreciable difference is discernible between the prefer-
ences of the more intelligent and those of the less intelligent children.

VIII. No correlation appears to exist between general intelligence
and the capacity for such simple tests as comparing the pitch of two
notes, or detecting whether one or two notes are being played on the
piano at a time.

IX. Greater musical capacity, as measured by such tests, is cor-
related (but only after the age of eleven) with much greater aversion to
discords.

X. Associations occur with musical intervals most frequently by far
at the ages of ten and eleven.

XI. More boys than girls have such associations. On the average
boys are much more critical than girls towards musical intervals.

XII. The Preparatory School girls show an aversion to discords
(except to the minor seventh) even at the ages of seven and eight, and
about the age of nine they give, on the average, an order of preferences
for the twelve intervals very similar to that given by adults. Thus, by
the age of nine, they reach a stage of development only attained by the
Elementary School children by the age of twelve or thirteen^.

1 My thanks are due to Dr C. S. Myei's for a careful reading of this paper and for the
suggestion of several emendations; also to Miss Preston, Headmistress of St Catherine's
School, St Andrews; Mr J. Williamson, Headmaster of South Tay Street School, Dundee,
and Mr A. Swiuton, Headmaster of Balfour Street School, Dundee, for the facilities they
kindly afforded me for testing their pupils in these experiments.

(Manuscript received 20 September, 1912.)

NOTE ON THE PROBABLE ERROR OF URBAN'S FORMULA
FOR THE METHOD OF JUST PERCEPTIBLE DIFFERENCES.

By GODFREY H. THOMSON,

Lecturer in Education, Armstrong College, Newcastle.

The object of this note is to correct an error in Urban's application
of Bernoulli's Theorem for the calculation of the probable error of the
Method of Just Perceptible Differences ^ which was followed by the
writer in an article on "The Best Form of the Method of Serial
Groups^" Fortunately the conclusions drawn in both articles are not
seriously affected by the correction.

By the Method of Just Perceptible Differences Urban really
means a process of calculation (the Limiting Process^), which can be
applied to data collected by several psychologically different methods.
The full calculations as carried out by Urban can, however, only be
performed if a large number of experiments have been made with each
stimulus value : he himself in the article quoted applied the Limiting
Process to data collected by the Method of Right and Wrong Cases*
with lifted weights. The standard weight was 100 grams, and was lifted
before each of the seven comparison weights, which were so chosen that
the subject nearly always recognised the extreme weights as lighter or
heavier respectively than the standard. The experimenter presented
the comparison weights to the subject in an irregular sequence, and the
whole series was repeated 450 times. The judgments given were lighter,
eqvxd, or heavier than the standard. The answers were entered in a

^ F. M. Urban, " Die Psychophysischen Massmethodea als Grundlagen empirischer
Messnngen," Arch. f. d. ges. Psychol. 1903, xv. 261-415.

^ This Journal, 1913, v. 398-416. The writer will be obliged if readers, in referring to
this article, will note this correction.

^ G. H. Thomson, " Comparison of Psychophysical Methods," this Journal, 1912,
V. 210.

* See Thomson, op. cit. 204.
J. of Psych. VI 16

218 The Probable EiTor of Urban' s Formula

table containing 450 horizontal rows and seven vertical columns
corresponding to the seven comparison weights.

Provided with this table, the calculator applies the Limiting Process
as follows. Beginning at the left-hand end of each row with the weight
84 grams, he passes along until a judgment heavier is met. The weight
at which this occurs is noted as one reading of the just perceptible posi-
tive diiference. For some reason which I cannot discover, Urban only
uses 400 rows of the table obtaining that number of readings which in
the case of his Subject I. were distributed as follows^ :

Table T.

Comparison weights r

84

88

92

96
76

100

101

108

Frequencies N

0

7

30

106

169

12

The mean of all these readings is 10036 grams, which is taken as
the mean just perceptible positive difference, and it is the probable
error of this number which is required. Before examining Urban's
processes for finding this, we may see approximately what it must be.
The " quartiles " of the above distribution occur at 96 grams and 104
grams, the semi-interquartile range is 4 grams and the probable error of
the mean is therefore about

\/400

= 02 gram.

Urban' indicates three processes for finding this quantity more accu-
rately. One of these he does not recommend or use. It is the ordinary
algorithm of Least Squares, and gives about 016 gram. The two
formulae which Urban does use are based on an inverse use of Bernoulli's
Theorem, and are both incorrect*. They give values 2-832 grams and
2*373 grams respectively, more than ten times too large.

In his first formula Urban assumes that with each of the comparison
weights r« there is associated a probability P« that it will be recorded
as a reading of the just perceptible positive difference. The frequencies
N in Table 1, when divided by 400, are experimental determinations of

1 Urban, op. cit. 322, Table 18. 2 Urban, op. cit. 313-317.

•■' I am inijebted to Professor Karl Pearson for confirmation of this statement.

Godfrey H. Thomson 219

these probabilities. The inverse use of Bernoulli's Theorem gives as
the probable error of P^

^,^.,-,^JP.O^iI.)

•(1),

where s is the total number of experiments (400). The mean just
perceptible difference T is

T=' =^P.r, (2).

Therefore (and this is where the mistake occurs) the probable error F of
T will be given by

F^=ico,'r,' (3).

But this last step would only be correct if the P's were independent
of one another. By the nature of their formation, however, their sum is
necessarily unity. The largest possible value for l.Pr is therefore
108 grams and the smallest possible value 84 grams, a range of
24 grams. Were the P's independently measured, their sum would not
necessarily be unity. Independent measurement would mean that in
one set of experiments, using all the weights, we would ascertain how
often the weight 84 grams was noted as a just perceptible positive
difference and how often it was not so noted, and nothing else. Then
in another set of experiments we would do the same for 88 grams, and
so on for each weight. Now the chances might possibly be against
each weight in turn just as we were doing the set of experiments which
were focussed upon it ; or on the other hand the chances might be in
favour of each weight in turn just at the right time. In the first case
each P might even be zero, which would give a value zero for the
threshold X Pr. In the second case each P might even be unity
(except the P at 84 grams, which must in any case be nearly zero, for
otherwise we would simply take a still lower weight as the beginning of
our set of comparison weights). This would then give a value of 588
grams for the threshold, namely

X Pr= 88 + 92 + 96 + 100 + 104 + 108.

The possible range assumed by equation (3) for the threshold is
therefore from zero to 588 grams. Of course no experimenter would
accept such results, but Urban's formula assumes their possibility.
Were experiments really made independently, they would be continued
until 2P approximated to unity and then the values would be adjusted

15—2

220 The Probable Error of Urban's Formula

as are the angles of a closed polygon in a survey. We may expect
therefore that Urban's probable error 2832 will be too large in some-
thing like the proportion of these two ranges and that the correct
value will be approximately of the order

108 — 84 -, „„£^ rv 1 o

ggg _ Q- X 2-832 = 0-12 gram,

a value much more in accordance with what we found by Least
Squares \

Urban's other process is based upon the actual number of answers
heavier recorded for each comparison weight. These were as follows in
450 trials-^ :

Table II.

Comparison weights r

84

88

92

96

100

104

108

Answers heavier

1

9

40

100

186

403

423

Urban now assumes that associated with each comparison weight r^
there is a probability p^ that the subject will answer heavier. The
numbers in Table 2, when divided by 450, are experimental determina-
tions of these probabilities. Each is therefore subject to a probable
error

o). = -6745

y^i

.(4),

where

s = 450 and q=l—p.
The former probabilities P can be calculated from the p's and then
the threshold T can be found from them. Since therefore T is ulti-
mately compounded of the values p, it ought to be possible to calculate
its probable error from the probable errors of the p's ; and since these
latter probabilities are quite independently measured, the objection
previously raised does not hold here. Unfortunately Urban performs
the algebra in two steps, calculating first the probable errors of the P's

1 This is not suggested as an exact or practical way of finding the probable error.
The alternative to Least Squares, if only Table I is known, is indicated towards the end
of this Note, after Urban's second formula has been discussed.

2 Multiply the numbers in Urban, op. cit. 287, Table 11, column Vp. I grosser, by 4-50 ;
or read direct from Urban, Application of Statistical Methods to the Problems of Psycho-
physics, Philadelphia, 1908, 174, Table 3,

Godfrey H. Thomson 221

and from these that of T, thus reintroducing the same mistake. His
second equation on page 316 op. cit. is incorrect. It assumes that

dpr'' ^®^'

which is not true. The correct formula is

^-^"KIJ^- • <«>-

where V2/3-' = -6745.

After performing the differentiation, and remembering that qn and jo,
must nearly equal zero, we get

2 ^(2 P,r,-r. 2 PA (7),

a much simpler formula to calculate than Urban's. It gives in the
present case the value 0'133 gram approximately. This is the formula
for just perceptible positive differences, that is for ascents. For descents
interchange p and q, and the suffixes 1 and n. Similar formulae hold
for the negative differences.

Urban's conclusion that the Limiting Process of calculation is, for
what it attempts, more accurate than calculation by the <I> (7) hypo-
thesis is of course not altered by this correction : the accuracy is even
greater than he supposed, and is hardly distinguishable from that of the
Lagrange interpolation formula, — as might be expected, since both pro-
cesses are alike in accepting the data as given and in finding the fifty
per cent, point without any attempt at smoothing the curve.

The writer, in the article cited above on the Method of Serial Groups,
fell into the same error. The Limiting Process is an extreme form of
the Group Process, and the writer checked his equation IV^ by seeing
that for certain values it reduced to Urban's equation. The correct form
of equation IV is

i^= •67450a/ S P' f 1% Wr^-r. S ^A ...(8),

where the letters have the same meaning as in the article quoted. For
the Limiting Process, ^=0, g=l, l=p and C=l; and the equation
reduces to equation (7) for descents.

The equation on page 409 of the same article is also incorrect. This
is Urban's case (la) where the P's but not the p's are known ^ to which

> This Journal, 1913, v. 415, Appendix II. 2 Urban, op. cit. 313.

I

222 The Probable Error of Urban' s Formula

he applies his first formula discussed above. If it is desired to avoid
making the assumption of a certain distribution which underlies the
Least Squares process, then the only way of handling this case would be
to calculate the jp's from the P's and use the observed P's and calculated
^'s in equation (7) or (8) as the case might be.

The incorrect probable errors in the writer's article are, like Urban 's,
too large; but the difference is not so great, because an absolute not
a difference threshold is being calculated, so that the values of r are
small. The correct probable errors are from one-third to a quarter of
those given ; and fortunately this proportion is sufficiently constant to
keep the various forms of the process in the same order of merit. The
two general conclusions on page 412 are therefore still correct although
the advantage of small groups is weakened.

{Manuscript received 30 August, 1913.)

THE EFFECTS OF OBSERVATIONAL ERRORS' AND
OTHER FACTORS UPON CORRELATION COEFFICIENTS
IN PSYCHOLOGY.

By WILLIAM BROWN.

{From the Psychological Lahoratm-y, King's College,
University of London.)

§ L The need of more careful determinations of individual corre-
lation coefficients.

§ 2. A means of testing empirically the validity of Spearman's
rnodified ' correction formula.'

§ 3. Preliminary results of an experimental research into the

coTrelation of errors of measurement.
[§ 4. Description of the application of two tests to a group of
school-hoys. By Mr W. H. Winch.]

§ 5. Detailed correlation, and other, results of these tests.

§ 6. The inapplicability of Spearman's formula in the case of these
tests. Suggestion of the best m,ethod of obtaining a reliable
measure of correlation.

§ 7. I'/ie causes of correlation between mental abilities.

§ 1. Despite the very considerable amount of careful work that
has been done during the last two or three years, both in England
and America, upon the correlation of mental abilities, and a greatly
improved insight into the significance and requirements of the mathe-
matical technique devised by Professor Karl Pearson and his school for
the manipulation of statistical material, much divergence still exists in
the amount of correlation found between identical abilities by different
investigators and even by the same investigator at different times and
with different groups of subjects. The cause of this divergence is to be
found in the great complexity of factors involved in the correlation
of any two mental abilities, and until these have been adequately

224 The Effects of ' Observational Errors '

investigated and allowed for, speculations as to the general cause or
causes of psychical correlation must be little more than futile. Instead
of attempting to support or disprove anticipatory hypotheses on the basis
of palpably inadequate material, investigators will now, it would seem,
be better advised to adopt a more ' intensive ' method of work and
obtain the fullest possible insight into the conditions of application of
the various mental tests and the factoi-s influencing their performance.
The time for general surveys of the entire field, with the valuable
training in method which they have brought, is now almost over, and a
change of tactics is needed if the present deadlock of assertion and
counter-assertion is to be surmounted.

§ 2, The object of the present short paper is to deal, somewhat
inadequately it is true, with one or two of the more salient difficulties
of method, and so to clear the ground for the new line of investigation
into mental variation which has already been commenced at King's
College with the aid of a grant from the Royal Society. The main
difficulty is that of the assumptions involved in the use of mathematical
formulae. It might have been expected that such assumptions would,
wherever possible, have been tested on actual psychological data. Yet
those who suggest these formulae in psychology seem singularly oblivious
of the necessity of this precaution. Instead of facts we are given
theoretical discussion. It was, for example, the merit of Prof C.
Spearman to suggest a mathematical method of eliminating ' observa-
tional errors ' in the correlation of series of mental measurements, but
I was able to show^ by empirical tests, that the assumptions upon
which his formula was based were incorrect. In response to this criti-
cism he has modified his formula ^ but unfortunately the new form still
involves assumptions that are not empirically justified. The essence of
the new method is to separate (in thought) the 'regular' deviations,
which are due to such factors as practice, fatigue, habituation, etc.,
from the so-called ' accidental ' deviations. The former, which may
have correlations of their own with one another and with the two
abilities whose correlation is sought, are to be eliminated by an appro-
priate grouping of the series of measures of the two abilities. Thus
Spearman recommends that if three measures of each ability be made,
the first and third may be taken together as one group, and the second
as the other ; if four, the first and fourth may form one group, and the
second and third the other. If a larger number of measurements be

1 Biometrika, April, 1910, vii. This Journal, Deo. 1910, in. 320.

2 This Journal, Dec. 1910, iii. 275.

William Brown 225

made, the odd members of the series should form one group and the
even members the other. In this way the effect of the 'regular'
change of observed ability from measure to measure may be neutralised.
The remaining deviations are called by him ' accidental ' deviations,
and on the assumption that they will be uncorrelated with one another
or with the abilities to be measured, he devises a new formula for their
elimination. His proof of this formula may be summarised and
simplified as follows :

Let Xa, ya = average measures of one particular individual in group
a for performances x and y respectively.
,. xi, yit = average measures of same particular individual in

group b for performances x and y respectively,
„ Xab, ydb = average measures of same particular individual in the
combined groups a and b.
Let d, e represent the corresponding 'accidental' as distinct from
the 'regular' deviations from these averages.

Then, owing to the way in which the groups are chosen, we have

Xa = X + da, Xij = X + db, Xab = X + dab,

ya = y + ea, yb = y + eb, yab = y ^^ab,

all the measurements being deviations from their mean values.
Assume d and e uncorrelated with each other or with x or y.
Then S{xaXj,) = S{a?), S(yayb) = S(y% and S(xabyab) = S(xy).
Assume, too, that o-j. = o-_^^ and a-y^ = o-^^.

Then r = '^(^^>_ = ^'(^^^^"^^

"' ^^x-)S{f) s/S{xaX,)8{yayb)

y

S{x„^yab) / S{a?ab)S{y-'„^)

^S{x^ab)S(y'ai)^ S{XaXi,)S{yayb)'

which, on reduction, = r.^^y^ 'l.\/U+ -^) ( 1 + -^) •

The validity of these assumptions may be tested by the same criteria
which I employed in my examination of the earlier formula, viz.

(1) S (xaya) = 'S (xy) = S(xbyb) within the limits of their p.e.'s,

(2) r^ _^ = 0 (a merely negative criterion, since it can also hold

Va - Vb when the ' accidental ' deviations are correlated).

§ 3. But before applying these criteria to some actual observations,
I should like to quote here a few preliminary results of a detailed
research into the correlation of errors of measurement which I carried

1 Op. cit. 288, 289.

226 The Effects of ' Observational Errors '

out three years ago at the suggestion of Professor Karl Pearson. Owing
to the very large number of observations made, the final results are not
even yet completely ready for publication. Stated very briefly, the
research consisted of the estimation, by two independent observers, of
the lengths of two series, X and Y, of lines ranging from 30 mm. to
170 mm. There were 50 lines in each series, and the length of every
line in series Y was five-thirds of the length of the corresponding line
in series X. The lines were observed from a fixed convenient distance,
and the estimated lengths marked off on ruled paper and then measured.
That is, the method of production was used. The 100 lines were
presented in irregular order and worked through five times by each
subject.

Clearly the correlation between the actual series X and the actual
series Y, when the measures in each are arranged in ascending order of
magnitude, is 1, and it is interesting to note how closely the observed
values are correlated with one another. Taking one observer only, we
have five separate estimations of series X and five of Y. We may test
Spearman's formula by grouping the second and fifth together as group
a, and the third and fourth as group 6. This gives

'^^abyab~

•99,

r.

• ' • fxy =

•99.

1
'2'

=

10001.

/+4)(^+4)

Although this result is slightly above 1, and therefore apparently
impossible, it is so close to the true value that, allowing for the prob-
able errors of the observed coefficients, we might regard it as evidence
in support of the applicability of the formula. But the following
product-moments refute this view, and show that the presuppositions
of the formula are not obeyed :

Siar") =19129, but SiwaX^,) =18689,
S(f) =53156, „ S{i/ayb) =40052,
S{ccy) = S18S4>, „ S(x„^yaA) = 280SS.

These divergences might be partly explained as due to a tendency
to underestimate the lines, but are far from being entirely explained
thus, since

Sixaya) = 24>206 and S(a;i,yb) = 30812,

whereas according to theory they should be approximately equal to one
another.

William Brown 227

Taking series 1 + 3 as group a, and 2 as group h, the divergences are

greater :

;S(a^) =19129, but SixaXb) =15538,

S{f) =53156, „ S{yayt) =32770,

^(^?/) = 31884, „ ^(^„53/„6) = 22210.

This case is of course not identical with that of measuring abilities
and their correlations, but is sufficiently analogous to throw some light
on the problem of ' observational errors ' in correlation. Here errors
of observation, which might be loosely called 'accidental' errors, are
correlated with one another and with the objects measured.

I have worked out the correlation between the 500 errors of observa-
tion and the corresponding true lengths of lines for the same observer",
and obtain the following results :

r = -033 ± -030,

V. = ^^^^^§^^- = -S2S±-02l,

V^= ^^''y^^^-^y'^^ =-182 + -029.

Here it will be observed that whereas r is negligible, r)^; is appreci-
able and 7}y is very considerable. A glance at the correlation table and
regression curves (see Appendix) shows the reason of this. The correla-
tion is ' skew ' and ' hetero-skedastic^' and whereas the 17a; curve is
almost vertical and shows very little correlation, the rjy curve approxi-
mates to a Gauss curve. It is probable that much of the correlation
between mental ability and errors of observation of that ability is skew
or non-linear. Where such is the case, it would be very difficult to deal
with, since the method of elimination by partial correlation assumes
linear regression. The error-correlation in the present instance is of
course at least partly due to the working of Weber's Law and could be
reduced by estimating the errors as fractions of the actual lengths, but
Weber's Law, or one equally non-linear, might well hold for the estima-
tion of different forms of mental ability, and so produce equally disturbing
results.

§ 4. With a view to obtaining more direct evidence upon the
question of Spearman's formula, I chose two tests which I had found

^ The scatter-diagram of a correlation table is said to be heteroskedastic when the
standard deviations of successive rows or columns ('arrays') increase or decrease instead
of remaining constant. A normal correlation table is homoskedastic, since here the S. D.'s
of the arrays are all equal to one another.

228 The Effects of ' Observatiofial Errors '

very reliable in previous correlation research, viz. (1) marking through
every letter in a page of print (Simple Motor), and (2) marking through
the as, n's, o's and s's in a similar page^ (Complex Motor), and con-
sulted with Mr W. H. Winch, who very kindly agreed to get these tests,
together with others of his own devising, applied six times each to
a group of 40 school-boys of the same age and school. The following is
a detailed description of the mode of application of the tests, kindly
contributed by Mr Winch, under whose direction the tests were done.

[Description of Simple Motor and Complex Motor Tests. By Mr
W. H. Winch.

"The Tests were done by all the eleven-year-old children in a
municipal elementary school for boys, situated in a rather poor district
in the south-east of London. The school was a strongly-disciplined and
hard-working one, and the boys might, on the whole, be relied on to
give their full attention to the work, even after the influence of novelty
had passed away. The Motor Tests formed part of a series consisting
also of Tests in Rote and Substance Memory, in Productive Imagina-
tion and in Reasoning. The results of the whole research are not yet
ready for publication, but those of the Motor Tests are now partially
presented. The exercises were worked according to a definite time-
table commencing on Wednesday, June 12th, 1912, at 9.45 in the
morning, and on the same day at 2.15 in the afternoon, six tests being
done on each day. Two days later the second series of tests was
worked, then on the next Wednesday a further series, in the mornings
and afternoons as before. The work continued thus on Wednesdays and
Fridays until each boy had worked six tests of each kind, the concluding
exercises occurring on Friday, June 28th.

Each morning's work began with a Simple Motor Test and each
afternoon's work with a Complex Motor Test, each Motor Test occupy-
ing exactly five minutes. The Simple Motor Test consisted in marking
out every letter from a page of print consisting of casually arranged
French words; the Complex Motor Test consisted in marking out the
a's, the n's the o's and the s's from similar pages. In the simple
exercises one mark was given for every letter clearly marked out with the
deduction of a mark for every letter omitted or doubtfully marked. In
the complex exercises one mark was given for every specified letter
correctly marked out, with the deduction of a mark for every omission
and for every wrong letter marked. Preliminary exercises were given
1 See Mental Measurement, 102; also This Journal, Dec. 1910, iii. 300.

William Brown 229

before the Test Series commenced, so that every boy might start with
a full understanding of what he had to do. The printed papers were
supplied from King's College by Dr W. Brown, and the exercises were
administered by three men on the staff of the school, all of whom had
had much experience of experimental work with boys.

The following is a brief summary of the results of the six Simple
Motor Tests :

Simple Motor Tests. Average Marks per Boy per Test.

School
Standard

m

No. of

Boys

1

Wed.
a.m.
427-0

Friday
a.m.
407-0

Wed.
a.m.
405-0

Friday
a.m.
415-0

Wed.
a.m.
374-0

Friday
a.m.
372-0

Average of

Six Tests

4000

IV

3

444-0

456-0

519-0

556-0

573-0

573-3

520-2

V

11

470-0

506-6

561-8

620-8

626-0

651-8

572-8

VI

19

482-8

517-8

582-5

626-3

637-4

678-3

587-5

vn

7

643-3

574-9

677-8

816-8

853-0

881-1

724-5

Averages per boy per test calculated from the individual figures taken to the nearest ten :

48-5 51-7 58-5 62-4 63-6 69-5

Standard deviation 6-8 8-1 9*9 14-0 15-4 15-2

It is interesting to note that even in this simple exercise the boys
in the upper classes draw further and further away from the lower ones
as the tests proceed.

The coeflScient of variability (found by dividing the standard devia-
tions by the averages), which commences approximately at ^, rises
to ^ and then to ^ as the exercises continue. It will be remembered
that all these boys are eleven-year-old children and are in no way
selected, since every boy in the school who was eleven at the date of
the tests worked the exercises. I called attention to this feature in the
work of school-children^ when discussing the educational question of the
same curriculum for all. Such a curriculum may not produce equality ;
often, on the contrary, an increased divergence may appear between
children of varying abilities.

The correlations between the results of the successive exercises were
worked out from the individual results by the product-moment formula
of Pearson. The coefficients were found to be as follow : between the
first and second, + -819 ; between the second and third, + "825 ; between
the third and fourth, 4- "842 ; between the fourth and fifth, + -951 ; and
between the fifth and sixth, + "909. The probable errors were negli-
gible. This Test, carried out under the conditions indicated, is evidently
of very high reliability.

1 Mind, XVIII. No. 69 : "A Modern Basis for Educational Theory," by W. H. Winch.

230 The Effects of ' Observational Errors *

The following is a brief summary of the results of the six Complex
Motor Tests :

Complex Motor Tests. Average Marks per Boy per Test.

School

Standard

III

No. of

Boys

1

Wed.
p.m.
65-0

Friday
p.m.
64-0

Wed.
p.m.
97-0

Friday

p.m.

91-0

Wed.
p.m.

102-0

Friday
p.m.

1360

Average of
Six Tests

92-5

IV

3

107-7

127-7

144-7

169-7

170-3

194-0

152-3

V

11

116-5

146-5

181-5

198-1

231-5

246-2

186-7

VI

19

124-1

149-8

176-9

196-2

218-1

239-3

184-1

VII

7

149-4

174-4

218-0

239-0

257-1

283-3

220-0

Averages per boy per test calculated from the individual figures taken to the nearest ten :

12-4 15-0 18-1 20-0 22-2 24-2

Standard deviations 36 3-5 4-2 4-3 5-1 5-0

Contrary to what might have been expected a priori, the abler boys
(as estimated by school progress) do not, in this more complex exercise,
draw away from the boys in the lower classes as they did in the simple
motor exercises ; the coefficient of variability falls rather than rises as
the exercises proceed. The reliability of the Test is very high. The
coefficients of correlation were as follow : between the first and second,
+ '874 ; between the second and third, + "QSl ; between the third and
fourth, + '951 ; between the fourth and fifth, + "952 ; and between the
fifth and sixth, + "930. The probable errors were negligible. The
coefficients were, as before, worked out from the individual results by
means of the Pearson product-moment formula. It would appear,
therefore, that this test, as well as the Simple Motor Test, if carried out
under proper experimental conditions, is very reliable and gives most
steady results. It will, I think, scarcely be doubted that, in so far
as these tests do yield a measure of motor fimctions, the number, the
succession, and the steadiness of the tests are such that we may be
fairly satisfied that we are not dealing with ' chance ' results, but that
the boys were working on their 'true form'."]

§ 5. I have independently calculated the values given in the pre-
ceding section, together with many others, and get results in some
cases slightly different owing to the fact that I omitted from all my
lists the single boy in standard III because he was so exceptionally
inferior in every respect. The differences are hardly appreciable, but I
refer to them here in order to avoid possible misunderstanding later on.
Calling the two abilities tested x and y, I get the following sequence of
correlations :

William Brown

231

^•^32/3 ~

29
44
59

48
47

Thus the general tendency is for the correlation coefficients to
increase at first and then remain fairly steady. This is in conflict with
the results of Binet^ and Burt^ and negatives the view that correlation
is due simply to voluntary attention or, what is almost the same thing,
to lack of practice.

Other correlations that throw light upon the factors at work are the
following :

V4

= •32
= •27
= •63

= •48

= •51

i

= •51

i

.= •60
= •53

V4='51

r. _. = -28

The first two of these coefficients show that y^ is less reliable than
the other series. The last shows that improvement in the one test is
not very highly correlated with improvement in the other.

The correlations between improvement and average ability in each
test is fairly high,

^x(.r6-.T,) = '^8, r-y(.^^_^^) = -50.

§ 6. The correlation between the totals of the two tests is

r^y = -56 + -07.

We have now to consider whether it is possible to improve on this
by means of Spearman's formula.

Taking measurements 1, 3 and 5 together for group a, and 2, 4 and
6 for group h, we should have

S {ccaya) = S {xy) = 8 {x^yi,).

On calculation (totals divided by 10),

S {xaya) = 6099, 8 (x,y,) = 8079.

^ L'Annee Psychologique, 1899, vi. 395.

2 This Journal, 1909, in. 168. It should, however, be noted that Burt is considering
the successive correlations of different tests with intelligence, so that our results are not
necessarily incompatible with his, especially if our correlations are to be considered as
mainly due to a community of specific rather than general factors.

232 The Effects of ' Observational Errors '

Now P.E. of S(.y)= ~^-- ^ ^- "Hr^.

.-. P.E. of S {XaVa) = 97, P.E. of 8 {Xbyi) = 131,
. ■. P.E. of S {wf,y,) - S {xaVa) = V97^+13P = 163.
Since this value is less than \ (8079 - 6099), Spearman's formula is
inapplicable.

(Had we applied the formula, the corrected value would have
been "61.)

Again, taking 2 and 5 as group a, and 3 and 4 as group h, we have
(totals divided by 10)

8 {xaVa) = 3494 ± 56, 8 {x^y^) = 3834 ± 62.
P.E. of difference = 83, which < ^ (3834 - 3494), thus again contra-
dicting the assumptions of the formula.
In this case, also, we find

r^ _ = "13 + '10 instead of 0.

Va-Vb

Finally, taking 1 and 3 for- group a, and 2 for group b, we have
8{xaya) = 48554 ± 821, 8{xbyb) = 45740 ± 774.
Difference = 2814. P.E. of difference = 1 128.

Here the difference is nearly three times its P.E., thus excluding the
use of the formula.

The conclusion to be drawn from these empirical tests is that either
the various methods of grouping (which Spearman himself recommends)
have not effected the elimination of the 'regular' deviations, or the
residual 'accidental' deviations are not uncorrelated with one another
or with the abilities themselves, or that both alternatives are realised.
One cannot help feeling that the term ' accidental ' is here used simply
as a cloak for ignorance and that no assumptions as to the correlation
or non-correlation of such deviations are in the least justified. Our
attitude in this as in so many problems of statistical psychology should
be one of extreme empiricism. There is still the big problem of indi-
vidual variability and its correlations to be reckoned with, as I pointed
out three years ago. Unfortunately six tests are not sufficient to give
any satisfactory measure of this, so that the problem must stand over
for a later research \

^ For the sake of completeness, however, I have worked out the correlations of
variability and ability on the basis of these six tests, and get the values :

r- =-74; r„ = '47 ; r„ „ .^ = -16.

kv.otx ky.oiy M.V.ofx

William Brown

2B3

The following table of means, standard deviations, and coefficients

of variation ( 1 throws additional light upon the matter of the

Vmean/ ° '^

' regular ' deviations.

487
70-6

14-5

X2

X,i

^4

•■^5

Xg

Mean

<r

520
79-2

15-2

581
82-4

14-2

653
136 0

20-8

667
146-5

220

701
145-4

^jTj (coeflQcieut of variation)

20-7

Ih

ya

Vz

Vi

y-o

J/6

Mean

(T

125
351

28-0

152
32-4

21-4

183
40-0

21-8

202
40-9

20-2

225
47-2

210

246
48-0

— (coefficient of variation)

Mean '

19-6

The shifting of the mean is slightly non-linear in both cases, though
more definitely so in the case of the afs. The coefficient of variation of
the x's increases from the first half to the second half of the series while
that of the ys remains fairly constant after the first. There is a marked
increase of steadiness of the figures towards the end of the series in
both cases. The conclusion to be drawn is, I think, that for the
accurate determination of a correlation coefficient a large number of
measures should be made at fixed intervals throughout an extended
period of observation, the means and as being determined in each case
in order to serve as controls, and then the later measurements showing
a sufficient degree of constancy of mean and a- should be averaged and the
coefficient calculated from them alone. This method would eliminate
'observational errors' as completely and safely as possible, since the
' accidental,' as distinct from the ' residual ' deviations would be
approximately neutralised by the law of averages, and their correlations,
even if an actuality, would be eliminated with them, while the 'residual'
deviations themselves would be reduced to a minimum. It is of course

J. of Psych. VI 16

234 The Effects of ' Observational Errors '

free from the criticism that the ' uncorrected ' coefficient contains the
same disturbances that render the use of the formula invalids

In the present case we must content ourselves with the mean of all
six measurements, since the number is too small to allow of the neglect
of the earlier ones.

§ 7. The question as to what exactly causes correlation between
measures of different forms of mental ability is a very difficult one, and
can only be adequately considered on the basis of results from that
' intensive ' method of research to which I have already referred as the
research of the future. The causes may — some of them certainly must
— come from without rather than from within. Mechanical memory
correlates with power of logical reasoning in all probability because
in the struggle for existence and attempted adaptation to one's environ-
ment the former ability was a precondition of the development of the
latter. The school environment, viz. the methods of teaching, may in
like manner contribute to the production of a correlation between
mathematical ability and scientific ability (in chemistry and physics).
The inner factors are partly general, partly specific, and the interesting
article of Hart and Spearman'* is an important contribution to the
study of those that are general. The material they work with certainly
indicates the probability of the existence of a general source of correla-
tion, although the absence of probable errors in the correlations of such
short series of correlations as those with which they had to deal deprives
their conclusions of the quality of absolute mathematical demonstration.
So far as my own results were employed in this investigation, I still
think that my own more cautious conclusions were all that could be
safely drawn from my data. With the exception of the tests that
patently measure intellectual ability and cognate aptitudes, the correla-
tion coefficients are so low as compared with their probable errors that
they contribute very small weight to the ' correlation of parallel series of
correlation coefficients ' which is in this work taken as the fundamental

' Cf. C. Spearman, "Der Beobachtungsfehler in der Korrelationslehre," ZUch. f. ang.
Psychol. 1912, vi. S. 74. Spearman writes here : " Brown und Betz Laben jedoch einen
groben Fehlschluss begangen, indem sie daraus eiitnahmen, dass die Storungen dem
Erganzungsprozess znr Last fielen. Denn diese Storungen sind schon im gewohnlichen
unerganzten Koeffizienten enthalten " This is of course quite untrue of the 'accidental'
disturbances, if, as I recommended, a sufficient number of independent measurements are
made of which the mean is taken. The 'regular' deviations do remain in both and can
best be reduced to a minimum by the procedure recommended above.

* B. Hart and C. Spearman, " General Ability, its Existence and Nature," this Journal,
1912, V. 51-79.

William Brown 235

criterion of the existence of a 'central factor.' And in the more
recent research of Mr Stanley Wyatt^, which supports the theory of
a central factor, the tests employed are in most cases chosen with special
reference to their involving what we in ordinary parlance know as
'general intelligence.' There is also the danger, which I pointed out
in one of my own researches some years ago, that community of external
influence, heterogeneity of material, and other ' irrelevant ' factors may
superimpose ' spurious ' correlation upon the results, thus emphasizing
the general causes of correlation as compared with the specific.

If the sole general cause of correlation is a certain ' common fund of
energy,' the order of closeness of correlation of different tests should
be constant in different groups of subjects, as I pointed out in 1910.
Results hitherto obtained do not show this agreement, and we can only
hope for conclusive evidence on the subject from a more accurate deter-
mination of the individual correlation coefficients.

1 Stanley Wyatt, " The Quantitative Investigation of Higher Mental Processes," this
Journal, 1913, vi. 109-133. The chief reason why he obtains larger coefficients than I did
for the same mental abilities is that his groups of subjects were not so stringently selected
as mine were. Selection always reduces correlation. I now think that, for this reason,
my own correlations were too low, although freer from the danger of superposed spurious
correlation than they would otherwise have been. As soon as one mixes classes, even of
children of the same age and school, spurious correlation, due to difference of discipline,
etc., comes into play.

[Manuscript received 20 August, 1913.]

16—2

■ «0

rO

s

o

•■?»

c>,

e

K

g

-1

o

p

o

1^

hH

i^

XI

o

•1

r«

^

M

<y-i

'TS

:z;

o

Plh

lO

p

PCS

o

O

^-

US

>C l(J

o

+ 1

f- I-- ip o >■> >p >p >o

ip

lO

w

i-(c-«'i;ao-^t-cb't:i«Tti-4i(XrH

C<1

+

■*

CO

lo «5 lo US in »o

o

+ i

C>1 (N ip Cd S'l (M t-

»p

cq

I-H

iM?OOii5<l'^t>-»l<50-*CI5':)<^'^

50

+

1 !

i«

•H

U5 1.5 >0 »C »C >0

o

"7 +

»0 (N (N >p >0 t^ C<1 ■?• t-

CO

ci5r-lb-ifft-(MOb-cbn5(MaDTt<>0

«o

',

I-l I-H I-H

o>

o

rH

1-H

1 1

>p ip »p >p ip »p «p *» >«

Ui

00

«

i«O5NCi|>OU5(N0<!l?-t^«5«i-<!»<

§

+

CO

I-H

! 1

>o ip in ip lo ip 1 »p

ip

o

»o

r-HC<jTHTjH-^ioicn5cbi-Hi<i ;o'

en

+

lO

CO

1 1

■-■^ >o 1 1 »o o 'fs lo >re

ip

1^

t~

(n'ot^ coco coe^t-wAi'

CO

+

1-H

tr-

1 1

I ip I 1 1 1 'P "P

ip

OS

o

1 1

iH' 1 1 1 l.-H^,-HOJM!Mr-(

io

CO

05

+

Ui

05
1 1

1 1 1 1 III 1 1

' Ji

1 1 1 1^1 1 I^l5<,„^l

00

CO

r-l

+

1— t

eo

1 1 1 1 I 1 1 1 1 1 1 '•"'•*

*

1 1

1 1 1 1 1 1 1 1 1 1 1 ^ ^

eo

lo

T-i

+

1 1 1 1 1 1 1 1 "^ "i^ 1 1

CO

1 j

IIIMIII^ iliHlH

-*

4«

1-1

+

a

I-H

a

1 1

1 1 1 1 11 1 1 '-' 1 11 1 1

1-H

CM

_c

' 1

.-1

1 1 1 1 1 1 1 1 1 1 1 1 1

+

C~

1

o8

1-1

a

1 1
05

1 1 1 1 1 1 1 1 1 r 1 1 1

■-H

+

£

iH

o

u

!i

05

1

f^

I-H

1 1

1 1 1 i 1 1 1 1 1 1 1 1 1 -'

I-H

t-

' 1

1 1 1 1 1 1 1 1 1 1 1 1 1

+

^

l?J

-H

*

iM

1 1 1 1 1 1 1 1 1 i 1 1 1 1

,

1

CO

1 1 1 1 1 1 1 1 1 1 1 1 1 1

1

1

(M

o

U5C«50»CO>OOiOOiOO»«0

W

00

so-^-^>^>ccc?Dt~t»aoxa535i-H

1 1 1 1 1 1 1 1 1 1 M 1 1

o ^

(5J IS^

oiooiooicoiocidoiooic

s^

eocc-*Tj<ioio«c^c-t~ooQ005o

s

•rata ai 'tsauii jo miindrj

09

0>OTt<«0t»(Nt~n5lCiO>ClQ0Q0O
•OUSSO-^OWi-Oi-li-liOCvlC^IWO

I + I++I+ + + +I+-I- +

(N tr-
io 05
00 iJ-l

II II

b

o 3-
H

(3SO«i<f?Ot-t-Q04|t(iiO»0iO»0

►i! illi

b

+ i

1 1 1 1 M 1 1 1 1 1 1 1^

+

i-H

+ 1
Oi

1 1 1 1 1 1 1 1 r 1 1 1 1

I-H CO
+

+

05

+ 1

fH

1111111111111-^

iH

+ .1

1 1 1 1 1 11^ 1 1^ 1^ 1

to

CO CO

+

i-l

+ 1
CO
i-H

1 1 1 M 1 1 1 1 11^ 1 1

I-l lO

+

CO

+7

1-1

1 1 1 ' 1' '(Mi-|l l'i-|l

o
lo c-i

+

rH

+7

«5 lO

1 1 1 1 >C (?J o» f •?> 1
1 1 1 li^'fliOqiMi-liMINiH!

00

rH
'P

+

.T

t-

II »0 1 >p lO lO 1(5

CO
CO

U3 op
I'o C<l

+ 1

|(jq<N ip lOiftipuS US

i-H I.H<JQ(NiN(N«biilOJ>0>0(M<N

o
ta OS

r—

05

(N

<M

o

o

-«

-H

CO

<N

:n

00

CO

rH

I

IH

238 The Effects of ' Ohservational Errors

APPENDIX 11.

Regression Cai-ves of Correlation Table.

Errors made, in mm.

THE MAIN PRINCIPLES OF SENSORY
INTEGRATION\

Bv HENRY J. WATT.

A. 1, The systematization of the sensations. ,

a. Note on extensity.
2. The systematization of the integrative sensory modes,
a. Note on the word ' mode.'
h. Note on the word ' integration.'

B. The main principles of integration.

1. "7%e mode which restdts from the integration of an attribute

must bear an immediate introspective resemblance to it."

2. " The results of tJie integration of the same generic attribute

in the different senses must be introspectively and func-
tionally similar."
a. The sub-principle of the explanation of apparent exceptions
and limitations to this rule : " If a mode of experience
does not occur where we might for any reason expect it,
this can be explained only by the absence of the variant
experiences upon which it is integratively dependent
and for the latter the natural limitations of physical and
physiological processes must be ultimately responsible."

3. " Every typical mode of experience must to some extent at

least arise spontaneously and automatically and inde-
pendently of such processes as will, attention, inference,
proof, or the like."
C Conclusion.

The formulation of principles is an important stage in the advance
of any science. Its beneficial effects far outweigh its disadvantages and

^ An abstract of this paper was read before the Sub-section (to Section I) of Psychology
at the Meeting of the Britisli Association at Birmingham, 1913.

240 The Main Principles of Sensory Integration

dangers. It is the sign of an increasing unanimity, a concentration of
criticism in various fields round one or two points of view, a growing
sense of the inherent connexions of the subject-matter. It means the
abandonment of extraneous principles of explanation most successful, it
may be, in objectively neighbouring provinces of science, but really
inapplicable to the one under consideration. It serves, moreover, as a
guide to research and to theory, thus supplementing mere exhaustive-
ness by some degree of enlightenment. And it is perfectly safe, unless
it is the outcome of a movement towards prejudice and bias.

The formulation of principles is highly necessary in psychology, for
it is recognised by many to be a sphere in which the effects of the
interaction of all the main forms of being — physical, physiological
biological, psychical, and social — are made patent. The introduction of
extraneous principles of explanation is highly probable in this case,
unless sufficient attention be given to the nature and applicability of
the principles to be admitted. The principles of the natural and
biological world do, of course, make themselves felt in the sphere of
experience. But they do not provide a sufficient basis for the proper
systematization of that sphere. The peculiar nature of the psychical
itself must be emphasized and principles must be devised for its eluci-
dation which are drawn from its own sources and may therefore be
expected to do the only full justice to its particular difficulties. This
claim is, in fact, an assertion of the priority of the psychical in the
psychical realm. It is also an assertion of the possibility and necessity
of a purely psychological systematization of the psychical.

A. A systematic psychology of sensory experience is perhaps the
greatest need of our science at the present time. It has been very
much neglected. That, no doubt, is due to the fact that the chief
motive of the study of the senses has been physiological. There
seemed to be so much to be gained by this physiological study and so
little air to breathe in a purely psychological atmosphere. But surely
there is no use in talking of a science of psychology at all, unless
the realm of sensory experience can be properly systematized. The
simplest and most fundamental problems involved in this task fall
into two main groups.

1. The systematization of the sensations is the first of these. Some
sort of a ' periodic table ' of the sensations must be formed, which will
serve as a framework and basis for any theory regarding the qualities of
sensation ; and the attributes of sensation must be reduced to a type.
This psychological task is a necessary preliminary to any pure psychology

Henry J. Watt

241

of the senses. I have attempted to fulfil it elsewhere ^ Only a short
summary and revision of the outcome of that attempt need here be
given. Of the six attributes of sensation, quality and intensity stand
somewhat apart from the others. Quality may be considered to occur
only in single and discrete forms in all cases, except in the senses
of vision and smell whose purely psychological treatment is still
problematical. Hardly in any case is there any dispute or difficulty
concerning intensity. The four other attributes — of extensity, order,
duration, and position in time — may be arranged usefully in the
following scheme :

Generic names of the

various dimensions

of sensation

Generic names of the attributes : j

Extensity

Order

1

These are

inherent

WITHOUT

WITH

variation, ii

1 all sensory

(Intra-) Systemic

extents,
masses,
volumes

1
localisations, j
positions,
pitches

Temporal

durations

positions-in-time

It is tempting to bring quality and intensity into parallel with this
scheme, so as to reduce the six attributes of sensation to a triad of
pairs, each pair being extensive and ordinal in its own peculiar dimen-
sion. But this is impossible for the following reasons ^c

(1) Qualities cannot be treated as orders, for they give no distances
or motions; even if that fact be ignored, it is introspectively evident
that they do not bear the stamp of an ordinal attribute. Even the
different colours we do not think of as points in a system ; how much
less then do we consider the qualities of the different senses in this

1 "The Elements of Experience and their Integration; or Modalism," this Journal,
1911, IV. 135 ff., esp. 148 ff. Psychology, London, T. C. and E. C. Jack, 1913, 21 f. Cf.
" The Psychology of Visual Motion," this Journal, 1913, vi. 26 f.

2 Cf. my paper "Are the intensity differences of sensations quantitative?" This
Journal, 1913, vi. 176 f.

242 The Main Principles of Sensory Integration

way. And if quality is not an ordinal attribute, it is certainly not a
merely extensive attribute.

(2) Intensity, likewise, can be treated, neither as an extensive
attribute, for it is essentially variable and is not introspectively identifi-
able with the extensive form of attribute ; nor as an ordinal attribute,
for it is neither phenomenally nor functionally like one of these.

(3) If quality and intensity formed such a pair of attributes, they
should prove readily adaptable to quantitative purposes, as do the
attributes of extensity and order in combination with one another
in connexion with the measurement of space and time. But this is
not the case.

a. Note on Extensity. The critical point of any discussion of this
attribute lies in the problem of its relation to the attribute of order.
When extensity is present in a pure form, according to Stout, as in the
case of the voluminousness of sounds, "it has no distinctively spatial
character, no internal order of positions and distances'." It seems as
if the quantitative aspect of space could exist without a spatial order ^.
Such statements suggest the following question, which may be expressed
in various forms : Is extensity as an attribute really variable ? Has it
for example, a minimum, say the sensory ' spot ' ? Or we might ask : Is
the extensity of the minimum different in variety or amount from that
of a postage stamp ? Is the voluminosity of a high tone different in
variety or amount from that of a low tone ? Surely it must seem
absurd to suggest assent to these questions.

What, then, are we to understand by the differences referred to,
e.g. the ' vast discomfort of a colic or lumbago,' the peculiarities of high
tones and of low tones, the differences of the areas felt from the
contact of a pencil point and of a postage stamp ? If extensity and
massiveness and voluminosity do not differ, extents and masses and
volumes surely do ; these are the things we distinguish in these cases.
But obviously no part is played in their composition by quality or
by intensity, not to mention the temporal attributes. The only other
attribute besides these and extensity is order, which does vary.

We may, therefore, suppose that extents and masses and volumes of
sensation differ in virtue of the varying number of orders included
within them (or by the varying number of sense-organs of neighbouring

1 G. F. Stout, Manual of Psychology, 1899, 337.

'^ Op. cit. p. 334. Cf. also p. 336: "We have all kinds of gradations between pure
extensity and fully definite extension." "Typical cases of extensive diflfuseness or
massiveness are afforded by organic sensations" (p. 337).

Henry J. Watt 243

position that have been excited). This conclusion is quite consistent
with the psychology and the physiology of the cutaneous, gustatory,
and visual sensations. Hesitation can only arise in connexion with the
massive sensations, articular, muscular, organic, and auditory. But it
must yield to a reiteration of the priority of psychological systemati-
zation and of the probable conformity of the results of physiological
study thereto. If muscular sensations from muscles of different size>
and articular sensations from joints of different size, differ in massive-
ness, surely there need be no hesitation in supposing that this difference
is correlated with a difference in the number of receptors excited. The
same remark applies to the sensations of colic, lumbago, hunger, thirst,
and the like. The varying voluminosity of sounds suggests that each
sound is really a mass or extent of sounds ; high tones are thin and
short, low tones are longer and perhaps bulkier, and, it may be, more
tenuous as well. Such a view would explain why the pitch and the
voluminosity of tones are fixedly correlated with one another. It is the
psychological statement to which Ewald's theory of hearing^ in many
respects forms a most suitable physiological counterpart.

But although extensity is not variable, it is a true attribute of
sensation, readily distinguishable from order. Without it we should
have neither areas nor voluminosities. That is evident if we remember
that a cognitive form of order ^ exists to which there is no accompanying
extensity, so that it is impossible to make a series of concepts, such
as those of number, adequately represent the real continuity of an
objective line or area. It might be supposed to be a sort of sensory
stuff, which is repeated and multiplied by the repetition of orders.
But the same notion would apply equally to any of the other attributes.
The quantitative treatment of extents and durations is possible, only in
virtue of the close, psychical kinship between sensory orders and con-
ceptual orders ; in a certain respect the latter grow immediately out of
the former, although they are extended very much beyond the range of
the variations of sensory order. Measured extents are not measured
extensities at all; for, as we have seen, extensity is not varied. But
extensity can be involved indifferently in a statement of what is
measured, because it is itself unvaried and can, therefore, introduce no
confusion or complication into the comprehension of that statement.
Extensity, for the same reason, seems to have a minimum only in
relation to order. A distinction of orders within the 'spot' is, of

1 J. R. Ewald, Arch. f. d. ges. Physiol. 1899, lxxvi. 147 ft.

2 Cf. K. Buhler, Arch. f. d. ges. Psychol. 1907, ix. 357 f.

244 The Main Principles of Sensory Integration

course, thinkable, but it does not exist in sensation. So extensity
seems to be variable only in conjunction with orders, especially when
the latter are all continuously adjacent and are given along with
uniform quality and intensity. Then the fusional function of extensity
comes into action and we get continuous extent or area. But the
differing orders involved in this extent, though no longer separately
distinguishable, are effectively present. It is just they which deter-
mine the extent of the sensational area or mass.

If orders are to be separately distinguishable under areal or massive
conditions, they must evidently be accompanied by variation in some
other attribute. The only other variable attributes are quality, inten-
sity, and position in time ; but there may be variation in more than one
of these at the same time, of course. This consideration seems to be of
some importance for the theory of orders and their complications \

It must be obvious that the above statements apply equally to the
attribute of duration. It is essentially an unvaried attribute, which
gives variable durations or stretches of time only in conjunction with
the variable attribute of position in time.

2. The systematization of the integrative modes of sensory experi-
ence is the task that for a scientific psychology inevitably follows upon
the systematization of the simplest sensations. In so far as these
modes occur under different circumstances, they must be identified and
reduced to types of graded complexity and referred to their t3^ical con-
ditions, so as to come within the purview of a general, systematic theory
of the constitution and interconnexions of experiences. Of these modes
there are two main groups — those which take place between sensations
which belong essentially to the same sensory system and those which
take place between sensory experiences which, like those of the two
eyes or the two ears, belong to different systems. Of the former,
intrasystemic integrations, distance, and interval of time are the
simplest. In many cases they involve a difference in the sensations
which make up the distance or the interval of time only in respect
of the attribute of order or of position in time ; and in those cases
in which a variation in extent or in duration is noticeable without
any accompanying discreteness or separateness of sensations in respect
of order or of position in time, we are justified by consideration of
the circumstances of stimulation in extending our statement and in
assuming that, in these cases also, distance and interval of time are
based upon sensations which differ only in respect of the attribute

^ Cf. my discussion of "The Psychology of Visual Motion," this Journal, 1913, vi. 26 ff.

Henry J. Watt 245

of order or of position in time. Moreover, distance occurs only in
those senses whose sensations differ readily and obviously in the
attribute of order. We are never called upon to distinguish hunger
or thirst distances, or distances of muscular sensation, or smell distances.
In these senses the variation that we notice is at most one of extent
or of massiveness. In so far as distance occurs in different senses,
however, we must expect and do find that it is phenomenally and
functionally the same.

All experiences are qualified by position in time of some form ;
consequently we can experience an interval of time between any two
experiences. But the interval is distincter when it is constituted by
experiences belonging to the same sense, and still more so when it is
given in those senses which are specially rhythmical, namely sound,
vision, and the motor group of senses — the articular, the muscular,
and the tactual. In these senses the stimulus can be readily manipu-
lated so as to cause an experience to begin and to cease at any desired
moment.

Distance and interval of time are, as modes of sensory experience,
peculiarly simple, in that they are the only modes which necessarily
involve a variation in only one of the attributes of the sensations upon
which they are, or may legitimately be supposed to be, dependent. On
the other hand, order and position in time are themselves the only two
attributes of sensation which can vary apart from variation of any of
the other attributes of sensation. Thus analysis confirms the intro-
spective simplicity of these modes.

The sensory mode that stands next to these two in point of
simplicity is motion. For many reasons it may be considered to be
a combination of the modes of distance and of interval of time. It is
therefore found in those senses which present the mode of distance.
Its phenomenal and functional identity in these senses, especially in
that of sound where it forms a part of what is collectively called
melody, is a problem of great interest at the present time. But the
study of motion presents peculiar difficulties'. For the present it may
suffice to say that motion is a combination of the two modes of distance
and of interval of time, involving simultaneous and continuous, though
not necessarily concomitant, variations in the attributes of order and of
position in time of the sensations which integrate to form it I

1 Cf. "The Psychology of Visual Motion," this Journal, 1913, vi. 26 ff.

2 For preliminary work towards the systematization of the modes of distance and of
motion, see my paper in this Journal, iv. 172 ff. and 1.57 ff.

246 The Mam Principles of Sensory Integration

a. Note on the word 'mode.' I find the use of this word very
convenient^. It serves, of course, in the first place to distinguish those
experiences which we may legitimately suppose to be integrated out of
simpler experiences, from experiences such as the simplest sensations
which show no sign of such derivation. But if we may presuppose
the systematic classification of these modes, we can then with the
help of this word and of adjectives signifying the name of each class
of modes indicate without any ambiguity or confusion exactly the kind
or complexity of experience involved in any particular state of mind.
That cannot be done with the commonly used word 'perception.'
When we speak of the perception of distance, it is not clear what
exactly is meant. Do we mean the perception of distance as an object
for the mind or as an experience, or do we merely mean the presence
and effectiveness of distance in our sensory experience ? If we wish to
study perception as distinct from any sensation or sensory mode, we
can indicate that by speaking of the study of the perceptual modes of
experience.

The word ' mode ' will also translate the German word Vorstellung in
many of its uses, for example in its application to the term Oestalt,
which has been used to indicate distance and motion and many other
experiences which differ from sensation in the same way as these do.
But it can only be misleading to talk of the 'quality' of a mode or
Gestalt. Every mode has its own introspective nature and affinities,
but these have only seldom anything to do with quality. Although
the unqualified use of the word * mode ' well translates the unqualified
use of the word Vorstellung, the use of the latter word is apt to be
as misleading as the English word perception, e.g. when we read in
one sentence of the Vorstellung der Zahl, Vorstellung der Distanz,
Vorstellung der Aehnlichkeit, and Vorstellung der Verschiedenheit^.
There are such things as sensory number and diff'erence, but they
are surely not modes, the same things as are distance and motion ;
there is a sensory mode of distance and a conceptual mode of distance,
but there is a great difference between them. We proceed unscientifi-
cally if we lose sight of these differences.

b. Note on the word 'integration.' This word indicates that the
resulting mode unifies the sensations to which it refers and is attached
and upon which it is psychically, if not also psycho-physically, dependent.
The word may therefore be used generally to express the known relations

^ Cf. this Journal, iv. 203 ; Psychology, 1913, chaps, ii. iv.
2 E.g. Witasek, Grundlinien der Psychologie, 1908, 222 B.

Henry J. Watt 247

between modes of experience and the simpler experiences upon which
they rest. And an inductive study of these relations in various
cases may be expected to lead us on to knowledge we could not gather
from any one particular case. So the word integration may imply
the general theory of the relation of a mode to its basis in experience,
which psychology may hope some day to attain. If this is borne in mind,
the use of the word can make neither for obscurity nor for confusion,
but can only be the means of scientific concentration and inquiry.

B. After these preliminary statements we may now consider the
main principles of sensory integration.

1. The first principle is as follows : The mode which results from the
integration of an attribute must hear an immediate introspective resem-
blance to it^. Or: Among the attributes or features of the simpler
experiences upon which a mode of experience is, or may legitimately be
supposed to be, psychically dependent, there must be one to which it
bears a much greater introspective resemblance or affinity than to any
other. The latter statement is more inductive in outlook, while the
former is more deductive. Only on the basis of such a principle as
this can a theory of psychical derivation or causality be built up which
will reveal in the world of mind that rationality and intelligibility which
we naturally expect to find in all things. The position involved in this
principle has been reached by psychology in three distinct steps.

a. For each variation in the derived or integrated state analysis
and experiment must show an unambiguous complex of stimulatory or
sensory data. This is an obvious and uncontestable truth. Only about
the relation of the derived state to the experiences with which it is
objectively correlated can there be any dispute.

b. Either : we talk in all cases only of stimulatory data, no matter
what the experiences we are investigating may be, mere aggregations
or unique modes. This position is taken by very many psychologists of
the present time. It leaves, of course, no room for the principle stated
above ; but neither does it leave any room for a science of pure
psychology. All we can then expect is a mere distinction of mental
states from one another and a correlation of them with physical or
physiological data, that is to say, psycho-physics or psycho-physiology. An
inquirer of a logical turn of mind might well ask how we can have
mere distinction without some trace of interconnexion by resemblance,
and, thereafter, without some theory in explanation of this resemblance ;
but if this thought arises in the minds of those who remain at the

* Cf. my Psychology, p. 26.

248 The Main Principles of Sensory Integration

position of this paragraph, it is rendered ineffective by some indefinite
belief which makes any hope of constructing a reasonable explanation of
the merely similar, or generally of the psychical, untenable. It must,
of course, be obvious that if there can be no pure psychology of sensory
experience, there can be no pure psychology of any kind of experience
at all.

Or: we allow a resultance of certain experiences from others by
association or by ' experience,' while denying the principle under
discussion. This position is closely associated with the theory of local
signs, but it is also in vogue with many in the treatment of cognitive
and other experiences. But it must be clear that the effect of
experience is unintelligible and association is impossible unless each of
the associating elements already differs from every other, whether it be
by its locality or order, or by its place in experience, or what not. A
series of identicals cannot be differentiated by any association with a
series of variants, if that association operates from the identical
elements towards the variants. To allow this would be to deny the
truth of the rule stated under (a) above. This alternative position,
then, allows of a pure psychology, in the sense of a system of correla-
tions of an objective kind between single experiences or between groups
of experiences. But it blocks the prospect of an intelligible and
reasonable science of experience. We must look for a corrective to its
negative attitude in further insight into the origin and nature of
association.

c. Association cannot be mere blind mechanism, a sort of bond
that arises when experiences impinge upon one another in the mind
and that requires no sort of counterpart or basis of origin in the
experiences that become associated. The purely mechanical view of
association prevails at the present time in the treatment of memory ;
for association can be treated systematically from a mechanical point of
view. But this abstract theoretical procedure may be only a part of
the whole truth. Purely mechanical memory involves the assumption
that experiences associate when they come into contact in the mind in
complete indifference to the affinity or dissimilarity of their ' contents.'
The most reasonable constellation of ideas, then, has a greater coher-
ence than any other grouping only because there are in it a greater
number of frequently repeated and therefore strong associations.
Meaning is just a general convergence of associations. But this is
surely not confirmed by the facts. What is associated must surely
cohere as conscious experience before the association arises. Of course

Henry J. Watt 249

there must first be contiguity of a certain degree between the associat-
ing parts; they must occur within a certain stretch of time. But
must we not suppose that having thus occurred they cohere because of
their psychical affinity, and that having cohered and integrated they can
then become associated to one another so that the one can revive the
other ? Mere mechanical memory means mental chaos and irrationality.
Fortuitous contiguity would as easily produce a coherent mind, as
fortuitous grouping of elements and natural selection would produce
the biological world without the coherent basis of law given in the
physical and chemical world. " A unitary mode of experience in which
the associating experiences are integrated is always presupposed,
although it is usually ignored^"

This principle is the outcome of all unsuccessful attempts to derive
special experiences from the grouping of other kinds of experience with
the help of association alone. Neither local sign, nor stereoscopic
vision, nor perception, nor the concept, nor recognition, nor thought,
nor any other unique and special kind of experience, can be satisfactorily
explained in this way. And if we must return to a direct consideration
of the basis of coherence or of integration in the introspective nature of
the experiences that form the basis of integration in all these cases,
must we not also look for an integrative basis in experience even in the
case of the seemingly most mechanical of associations ? We may be in
doubt about thus generalising the result, but there can be no hesitation
about accepting the principle in the case of all unique modes of
experience. If the objective dependence of one experience upon others
compels us to classify it as a special mode of experience, and if we may
therefore hope for a theory of its derivation or integration out of some
one or more features of the experiences it is psychically dependent
upon, then it is clear that we can look for its integrative basis only
among those features of the experiences upon which it is dependent
which bear an introspective resemblance to the mode in question. The
true basis of integration will bear a greater resemblance to the mode in
question than any other feature of the integrating experiences. It is
evident that such a principle will serve as a guide both to experimental
research and to theory. Moreover, if a mode is variable, the components
of its integrative basis must be variable, as in the cases of distance and
feeling ; but if it is invariable, as in the case of recognition, the com-
ponents of its integrative basis cannot be variable.

Whatever is, is rational. In reference to the present position
1 Psychology, p. 60. Cf. this Journal, iv. 130, 139, and esp. 149 f.

J. of Psych. VI 17

250 The Main Prhiciples of Sensory Integration

of integrative psychological theory, this means that if we are to
suppose that dependent mental states are derived from the integra-
tion of those upon which they are dependent, it would seem to us
more satisfactory and intelligible that there should be some degree,
or the highest possible degree, of resemblance between the dependent
state and the feature or attribute of the conditioning experiences
upon which the former in the case of variable modes is known to
be dependent and in the case of invariable modes may be supposed
to be dependent. More than this we cannot expect. If unique
types of experience do not bear quantitative relations to one another,
the relations that exist between them cannot in all cases be those
of the type of reasoning. For that would be a denial of their
specific nature. A standard for the discovery of these relations can
then be found only in some other general appeal which the typical form
of these relations in known cases may make to our minds. One
element in that appeal at least must be degree of resemblance between
integrative basis and derived mode. What other elements it may
contain inductive research will show. Only on these lines can we hope
for a science of pure psychology.

2. The second principle of integi-ation is as follows : The results of
the integration of the same generic attribute in the different senses must
be introspectively and functionally similar^. Stated more generally it
reads : the introspective and functional nature of an integrated mode of
experience is essentially independent of the attributive or other accom-
paniments of its integrative basis. Wherever the requisite integrative
basis occurs, the same generic mode will result. This principle is a
necessary step in the systematization which is to constitute a pure
psychological science. I have attempted to establish it in detail in the
case of the simplest sensoiy modes of distance and motion^ But it
must also hold in such cases as feeling, recognition, thought, and the
like, for these can be occasioned by the most varied sensory and other
experiences. The integrative bases of any mode must be considered to
be the same in all cases, no matter what the accompanying differences
may be. Experimental research will undoubtedly lead to the confirma-
tion of this principle in all accessible cases. Very often the similarities
of modes are passed by as mere analogies. That may serve as a good
maxim where there is no insight into the systematic nature of experi-
ence to act as a guide. But it would be wrong to block the outlook

' Cf. my Psycliologxj , p. 27. ^ gee this Journal, iv. 1-57 fif.

Henry J. Watt 251

and progress of systematization by an ascetic cult of this idea of
analogy.

If this principle be granted, we can hope to establish general rules
for the relation of generic modes to the generic attributes or features of
the experiences from which they are integrated. For example, " motion
is found developed upon every group of sensations which show distinct
variations from one another in order ^ " ; and, " we find distance in all
those senses which show order and are capable of the modification of
motion I" Rules may also be expected to hgld for the limits of time
within which alone the integration of those modes that are based upon
successive experiences can take place. For we have reason to believe
that in so far as all experiences are qualified by the attribute of
temporal order, all integrative processes which involve successive experi-
ences are subject to certain limits of difference of temporal order.

This principle would also lead us to expect that if a certain mode of
experience can be integrated from simultaneous components it should
also result from the integration of components which follow one another
within the time limits just mentioned. Conversely we should be able
to transfer our expectation in a similar manner from successive to
simultaneous integration of the same mode, unless, of course, differences
in either of the temporal attributes be an essential part of the founda-
tion of its integration, as is the case in the integration of motion. If
the temporal attributes are not the essential basis of an integration, it
is clear that any differences in them that fall within the time limits of
integration, should be as irrelevant to the integration as is the presence
of identical or unvaried attributes.

It cannot, of course, be evident in detail how far this irrelevance of
accompanying differences, such, for example, as those of quality in the
case of distance and motion, extends. But it is assured by a broad
consideration of the conditions of occurrence of the various experiences
hitherto distinguished by psychology. We must therefore be on the
look-out for it ; and if it is not forthcoming as we should expect, we
must find good objective reasons for its absence. It is fortunate that in
the finding of these good reasons we can accept the guidance of a minor
principle of explanation.

a. The sub-principle of the explanation of apparent exceptions to
this law. If a mode of experience does not occur where we might for
any reason expect it, that can be explained only by the absence of the
variant experiences upon which it is integratively dependent and for

1 This Journal, iv. 157. ^ Ibid. 173.

17—2

252 The Main Principles of Sensory Integration

this the natural limitations of physical and physiological processes must
be ultimately responsible.

It is the task of science to expound with the utmost detail the
nature of the coherence that binds events into unitary systems of
greater and greater extent. Each particular science is concerned with a
part of the whole that more or less obviously forms a unitary system.
If it discovers in its sphere that kind of coherence that characterizes
another sphere of science, it thereby joins with that other to form a
system of greater extent than either. But it does not therefore identify
its subject-matter wholly with that of the cognate science. The two
remain distinct in so far as the forms of coherence that characterize
them differ. Now no one would deny that the forms of coherence that
characterize the psychical world differ very much from those that
characterize the physical and the biological worlds. But they are not
wholly independent ; something is common to them all. For on any
view whatsoever it is clear that our knowledge of the physical world is
dependent, not only upon the actual occurrence of physical processes,
but also upon the transmission of these in some form or other through
the sense-organs to the central nervous system. We can know of a
physical process only if the dififerences of the parts and the manner of
the arrangement it involves can be brought into correlation with those
involved in a unitary psychical process. This holds, not only for
cognition, but also for any kind of adaptation that may exist between
the physical and the psychi&il realms. Such adaptation can occur only
in so far as by some means or other a correlation of process can be
carried through the three kingdoms of the physical, the physiological,
and the psychical. In so far as physical processes occur at a slower rate
of change than the minimum required for psychical integration, we
cannot become aware of them, unless we can secure some means of
bringing their rate of change within the narrow compass of the mind.
If a physical change cannot be made to affect a physiological organ
appropriately, we must remain ignorant of it, unless we transfer it
through some medium which we understand so as to obtain the
appropriate effect. And so on.

The mode of distance, for example, cannot be produced apart from
variation of the attribute of order; it is therefore practically absent
from the organic, muscular, and olfactory senses. In the organic senses
there may be a certain variation in massiveness, involving difference of
orders, but we do not have a hunger distance or a thirst distance in
any proper sense of the word. Similarly we notice that the muscular

Henry J. Watt 253

sensations from different muscles differ in massiveness and are localised
at different parts of the body, but the sensations that come from one
and the same muscle do not seem to vary in massiveness or in localisa-
tion. Thus a muscular distance, which might be constituted by the
simultaneous occurrence of sensations from different muscles can hardly
occur without the simultaneous excitation of such tactual sensations as
would form a tactual distance. The latter for various reasons, such as
variation, frequency, and correlation with other senses and modes, have
a cognitive value that the former can never acquire for want of
variability. Muscular distance will therefore be so obscure or so
blended with tactual distance as to be hardly noticeable. In the sense
of smell, distance seems to be quite lacking. If there is any olfactory
order or localisation it seems to be so unvaried as to be useless. And
even if smell has its order in some other form than localisation, in us at
least the sense is so sluggish that the variations of order necessary for
distance cannot occur within the time limits of integration. The
same reasons as prevent the occurrence of distance prevent pari passu
the integration of motion.

Interval of time is found under all possible circumstances, in all
regions of experience. Only in the form of rhythm is there any
restriction to its occurrence. The reason for that fact has been already
mentioned : only certain experiences can be made to begin and to cease
at any desired moment or periodically. So we cannot have rhythms of
taste, temperature, smell, organic sensation, feelings, ideas or thoughts.

The peculiar correlation which is found in the sense of sound
between pitch and voluminosity is responsible for all the limitations
of integration which specially characterize this sense. Pitch is an
aspect of sound which represents the individuality of the sounding
object much better than does its spatial localisation. Besides, it seems
clear that if the latter had been maintained at all costs on the basis of
simple sensation as a sort of local sign, the former would never have
been developed. The greater advantage lay in the attainment of a
discrimination of pitch even at the temporary or permanent sacrifice of
a direct auditory form of localisation. But two more or less efficient
methods of localisation have been secured — the mobile-ear-funnel method
of many animals and the binaural method of man. As a consequence,
however, of the preferential development of pitch we have no true
experience of auditory solidity and the smaller variations of tonal
interval are rendered highly unclear or even impossible by the presence
of beats and intertones.

254 The Main Principles of Sensory Integration

It is of interest in this connexion to recall a remark made by Ewald,
He wrote^ : " Man begeht immer gewisse Fehler wenn man die
Funktionsweise eines Sinnesorganes init der eines anderen vergleicht."
" Wenn der physikalische Anlass fur eine bestimmte Empfindung sich
in irgend welcher Weise andert imd dadurch eine Veranderung der
Empfindung bewirkt, so scheint mir keine Uebereinstimmung ira
Wesen der beiden Veranderungen bestehen zu miissen." But this is
a principle of apology which cannot be accepted from Ewald. For
the merit of his theory, apart from its experimental foundation, — a
merit that is brought forward into the light by his own sixth argument
against Helmholtz's theory — is the facility with which the phylogenetic
development of hearing can be traced with its help. For it is just
because and in so far as the physical variants of sound have always
been the same and the physiological apparatus they play upon has
gradually changed in the course of the development of the race, that the
psychical results have gradually developed. The peculiar nature of the
physiological apparatus has secured for it, not a fragmentarily specialised
development, but an equalised development. The system of sounds
which results is just as equalised and balanced in its nature. Besides,
Ewald does assume that there must be some agreement between
physiological and psychical changes ; for he postulates a special
physiological means of getting round the necessity for this agree-
ment in the case of the ear: — his coupled-buttons theory^. This,
however, is a forced and artificial way of overcoming his chief difficulty,
which is to explain why, on his theory, we do not hear a series of
identical tones for each component of a tone picture, instead of only one
tone. In the light of his criticism of Helmholtz's theory, this part of
Ewald's theory is just as fantastic as is Helmholtz's. For what deus
ex machina is to make all these coupled-buttons-connexions for the
organism ? How are they to begin and to be progressively developed ?

If we can once decide in what manner any mode of experience
varies, we thereby obtain an index to the integrative basis of that
mode. This guidance is of great importance in those cases in which
the integrative basis of a mode stands in a complex psychical environ-
ment from which it is not easily distinguished or isolated. If the
variation of a mode is restricted or if there is none at all, its integrative
basis should consist of only one pair of unchangingly different
experiences. Such a case may perhaps be exemplified by the mode
of recognition.

• Op. cit. 181 f. 2 Op. cit. 183 f.

Henry J. Watt 255

A problem of considerable magnitude is presented in the case of the
absence from certain minds of experiences known to other minds.
Animals, for example, do not reason. Probably they also lack the
general concept and all those cognitive experiences which involve it ;
they can hardly be supposed to localise their memorial experiences
in their past. With all other simpler experiences they may well be
presumed to be equipped. But if they can see and hear and smell and
feel as well as we can, perhaps in varying ways better, why does their
experience not develop upon this sensory basis to the heights it reaches
in the human mind i The answer to be deduced from the principle
here stated denies that the animal possesses the full integrative
basis of the experiences it lacks. It would be presumptuous in the
present state of knowledge regarding the higher cognitive states to
attempt to indicate what is lacking or why it is lacking. An alterna-
tive view refers the limitation to res^trictions set by the level of
development that the brain of the animal has reached. But that
explanation is either psychically blank and valueless, or it implies
that a further development would add some experiences to those the
animal already has and so make the appearance of the higher cognitive
states possible. Thus either the view stated above is conceded, or it is
assumed that the higher modes of experience come into being by direct
dependence on the development of the brain, not through the medium
of the simpler experiences of whose integration the modes in question
may legitimately be supposed to be the result. On the alternative view
a pure science of psychology is, of course, impossible. Such a con-
clusion can hardly be entertained seriously for long, whatever divergence
of views there may be regarding the kind of elementary experiences
that are lacking in the animal.

3. The third principle of integration is as follows : Every typical
mode of expeHence must to some extent at least a,nse spontaneously and
automatically and independently of such processes as reason, thought,
determining purpose, and the like, unless these processes themselves are the
modes in question.

If it be borne in mind that a mere aggregation of experiences
presents no problem and that every mode of experience worthy of that
name must make some new addition to experience, it might hardly
seem necessary to state this principle explicitly. It might seem so
obvious as to be trite. But much of the past and current theory of
the growth and development of the mind so thoroughly ignores the
problem of the unique modes of experience that the principle may seem

256 The Main Principles of Seiisory Integration

to contain a new and startling truth. There can be no universal guide
to the development of the mind, be it called reason or thinking or
self-realisation or teleology, or what not. The mind must develop
when it can, when the conditions for that development have been
given ; and what then happens is really development, a step forwards,
something new, no mere unmasking of the obscure. The only guide to
mental development, if it can be properly called by that name, is the
illumination each step of integi'ation brings with itself. It is itself
its own coherence and justification. It reveals its own necessity, in
part at least, when it comes ; but it cannot be foreseen. In the light of
the preceding two principles, integrative processes are most reasonable
and intelligible, and with increasing knowledge they will appear still
more so. But they are not themselves the product of reasoning ; they
must arise spontaneously. It is important to emphasize this in view of
the fact that thought and purposive determination and such other
processes are not only the instruments of science, but are themselves
also modes of experience which must arise spontaneously. As the
instruments of science, reason and thought provide us with standards
of coherence in the form of identity and repetition, approximation and
similarity, and these are our favourite tests for the manifold forms
of coherence we find in the various spheres of being, including the
relations of modes to their integrative basis. But while retaining these
tests even in these last cases, we must not lose sight of the fact that
each unique integrative process is and remains unique, and therefore
contains a justification of its own, which we can never hope to extract
from it by any inductive or other cognitive procedure. That justifica-
tion is simply the coherence and insight the integrative process itself is.
The higher cognitive and the conative processes bear another
important relation to the integrative processes in that they may
serve to extend the conditions under which they take place, to support
them by making these conditions more enduring, more compatible with
the limitations of integration, and therefore virtually wider in scope.
Once an integrative process has occurred, its signs or criteria can be
established for indirect use. " But unless our minds recognised, or
thought, or felt spontaneously, we could never even begin to collect
tests for the recurrence of experiences, or for the truth or falsehood
of asserted relations or for the justification of beauty. Nothing but
the direct insight of experience can set the mind the larger task of
extending that insight to the uttermost bounds of reason \"
^ See my Psychology , p. 27.

Henry J. Watt 257

In the case of certain nova of experience it is relatively easy to show
that they are integrative modes, but it may be very difficult to show
from what features of the experiences upon which they are, or may
legitimately be supposed to be, dependent in an objective psychical
sense they are integrated. This difficulty may be supposed to be
due partly to the complexity of the experiences which regularly
accompany the essential integrative basis, partly to the fact that the
nova are nova and can draw the attention and be compared and generally
be the basis of new integrative processes, as if they were independent
elements. From another point of view, however, this peculiarity is
of great advantage; for it maintains the same freedom of mind for all
stages of development. As integrative processes are originally sponta-
neous, the mind can accept their product without making special
reference by attention or otherwise even to those experiences that form
the essential basis of the integration. We can compare distances,
tonal intervals, motions and melodies, without troubling to compare
the orders and times that constitute them. We are immediately aware
of the identity or difference of the mode itself in the various instances
given. Thus the subjective efforts of the mind can be applied to any
level or to any one of all the integrative processes which arise sponta-
neously upon any given occasion. This statement is absolutely
thorough-going, as we have already noticed that every integrative
process, no matter what its nature, must, to some extent at least, be
spontaneous and automatic. Effort and attention may have to be
applied indirectly to procure its appearance, as when we adjust our
sense-organs, our body, our actions, our memories, our thoughts, in
order to maintain a certain stream of experiences. But that stream of
experience must, to some extent at least, flow spontaneously. The
attention may then be applied to any point of it, usually its highest, in
order to aid the spontaneous integration which is taking place at that
point. The aid given may consist in rendering the integrative basis
stabler, or in reducing the differences which present themselves to
within the limits of spontaneous integration by means of special
manipulation of the corresponding stimuli, or in repeating the series
of integrating experiences so that the binding power of associations
derived from simpler forms of integration may extend the integration
in question over a longer stretch of time than that natural to the
integration. What cannot be brought simultaneously within the
compass of the mind, so as to integrate spontaneously there, may be
taken in successive series and made to pass through the mind so

258 The Main Principles of Sensori/ Integration

rapidly that it will then spontaneously reveal all its integrative
secrets.

C. Conclusion.

The first principle of integi-ation is, by growing consent, almost
agreed to already. In one form or another, sensory or motor, it is
the only acceptable conclusion of the long-drawn-out discussion of the
origin of local signs. They cannot be thought to originate out of the
association or combination of anything that is not already local sign.
What is derived is therefore not primitive local sign, but only the
complications and modifications of local sign that arise under varying
circumstances, on the basis of a correlation of the local signs of
experiences of different systems, such as eyes, ears, vision and touch,
touch and sound, vision and sound, etc. The same conclusion appears
to be inevitable in the discussion of other important problems. The
outcome of Jaensch's extensive investigation of depth is : " Die Tiefen-
wahrnehmung hangt aufs engste zusammen mit Wanderungen der
optischen Aufmerksamkeit und den mit ihnen verkniipften Impulsen,
also mit einer dem Gesichtssinn eigentiimlichen Funktion. Hieraus
erklart sich, dass Tiefenwahrnehmung des Gesichtssinnes in keiner
Weise mit Empfindungen und Vorstellungen, welche einem anderen
Sinnesgebiet entstammen, identifiziert werden kann, sondern einem
eben nur dem Gesichtssinn eigentiimlichen Inhalt darstellt^." A
similar remark may be quoted from a discussion of the various theories
that have been given for the state of recognition. In criticising
Rabier's theory, Katzaroff says : " Pourquoi ces divers sentiments
qu'invoque Rabier, sentiment d'absence d'effort et de necessity qui
caract^rise le souvenir par opposition a la fiction, sont-ils permutes dans
la conscience en vm sentiment de ddja vu, au lieu de rester ce qu'ils
sont originairement^ ? " So also Titchener : " Wundt's theory is open
to the objection urged against his theory of space. The blending of
affective process with sensation means, elsewhere in the mental life, not
time but feeling; and we cannot understand how, in this particular
case, the new product should arisen" Every criticism of the insufficiency

* E. R. Jaensch, " Ueber die Wahrnehmung des Baumes," Ztsch. f. Psychol. Erg.-bd.
6, 1911, 357.

2 D. Katzarofif, *' Contribution k I'^tude de la Recognition," Arch, de Psychol. 1911, xi.
15, of. also p. 19 and elsewhere.

^ E. B. Titchener, Textbook of Psychology, 1910, 347.

Henry J. Watt 259

of mere association and the hopelessness of all attempts to come through
with its aid alone are founded on this first principle of integration.
Reid's answer to Hume's scepticism is the first step towards recovery
from failure to do justice to the facts. The facts must be recognised.
But this acceptance cannot now be framed so as to exclude further
inquiry. For if some plausibility of derivation, some sort of resemblance,
is what we desire, on finding it we necessarily accept the task of making
an inductive study of these resemblances and of furnishing as adequate
a theory of derivation as possible.

The second principle of integration is not by any means generally
conceded. In fact it is usually implicitly denied. But whatever beliefs
or prejudices may oppose it, it is the inevitable consequence of a
systematization of the sensations and an essential part of any scientific
psychology. It calls, of course, for the fullest experimental study of
each mode of experience, both in respect of phenomenology and of
function. The greater the disinterested devotion applied to its study,
the more likely is it to be confirmed. For it promises the coincidence
of broad rational demands with the facts, if only we treat the facts
exhaustively enough. The psychology of the day presents many cases
of difficulty and of opposition between reason and fact which call
urgently for resolution.

The insight into the third principle is clouded by all sorts of
philosophical generalities regarding continuity which do not attempt to
define or to delimit precisely the mode of operation of the principle of
continuity or to reconcile the demand for continuity with other legiti-
mate demands. But the continuity and coherence are there. We do
not need to create them ; we have only to recognise them as they are,
and to explain them. Recognising them for what they are cannot, how-
ever, mean attempting to maintain that experience brings no progress,
no enrichment, nothing new, nothing more than was already within
its compass. It is equally futile to barter the facts for a notion of self-
development, or of the realisation of an end, as if that were a form of
process in which all that is finally attained were already there from
the lowliest form of consciousness, and so satisfied a craze for barren
continuity. For purposive process in experience is itself undoubtedly a
unique form of process, which therefore no more offers a standard for all
other forms of integration than does any other unique process. If the
continuity is there, we must just study it as we can and by inductive
procedure extract from it what secrets it has to yield. Similarity is
surely a kind of continuity. Whether it will suffice to cover the facts.

260 The Main Pinncijyles of Sensory Integration

only detailed study can tell. But that it plays an important part in
them, cannot be denied.

This third principle is indispensable in the formation of any theory
that exceeds the bounds of sensationalism or its analogues. But it
would be one-sided without the balance of the other two principles.
Mere nova are inexplicable, whereas nova within a matrix of similarity
offer the hope of an approximation towards completeness of theory.
Even if distance were procured by a sort of sensory presentation of the
orders intervening between those which bound it^, it could not be
thought, as distance, to be a mere aggregate of orders, for it is more
than that. It integrates these orders in a special way, which can only
progi'essively be exhausted by knowledge.

This principle has another important aspect. It offers a basis for
the separation of the objective mind and its processes from the subjec-
tive mind of effort, assent, attention, and the like. If we know that we
have the objective mind before us at any point, we can hope to
determine its scope progressively by following out the various steps of
its integrative development. There is evidence that the processes of
integi'ation can be influenced in various ways more or less extensively
by the attention, but it must be just as erroneous to suggest that they
originate in the processes of attention^, as it would be to adopt the
view that the mind involves only processes of integration of the kind
found in the senses or in the cognitive states. If attention is involved
in integration, it can only be supposed to support or to oppose the
process of integration. It is not likely that the objective mind is a sort
of image or parallel of the subjective mind of attention. Such a thing
would not only be hardly intelligible, but it would refer or transfer all
the problems of the objective mind to a shadowy world of subjective
attention without any prospect of ultimate solution.

^ Cf. Jaensch, op. cit. chap. 6. ^ cf. Jaensch, op. cit., especially chap. 5.

{Manuscript received 20 July, 1913.)

PUBLICATIONS RECENTLY RECEIVED

Mental and Social Measurements. By Professor Edward L. Thorxdike.
Second Edition — revised and enlarged. New York : Teachers' College,
Columbia University. 1913. pp. xii + 277. $2.50.

The second edition of this well-known work aims at presenting the student with
a clearer, though perhaps more elementary, treatment of the subject than before ;
greater care being expended in describing the methods employed in solving
statistical problems. As he writes in the preface to the former edition, " the
author has had in mind the needs of the students of economics, sociology and
education, possibly even more than those of students of psychology, pure and
simple.... The book may, with certiiin limitations, be used as an introduction to
the theory of measurement of all variable [)henomena." It is intended for those
who find the mathematical treatment, given in such books as Brown's or Yule's, too
difficult.

Variations in the Grades of High School Pupils. By Clarence Truman
Gray. Educational Psychology Monograph No. 8. Baltimore: Warwick
and York. 1913. pp.120. $1.25.

By 'grades' are meant the percentage marks allotted by a teacher to his
students in any subject. The special problems with which the writer here deals
are (i) the variations in grading of the same pupils in different years and in different
subjects of the high-school curriculum, (ii) the distribution-curves of the grades in
different schools and in different subjects, (iii) the various methods of grading
adopted by different teachers and by different schools, and (iv) the influence of
home conditions on the variability of a pupil's grading from year to year. The
chief value of the book lies in its attempt to provide "a relatively simple method by
means of which any high-school principal can study the condition of the grading
in his own school and take due steps to remedy the faults that he may find."

The Conservation of the Child : a manual of clinical psychology presenting
the exam,ination and treatment of backward children. By Dr Arthur
Holmes. Philadelphia and London : J. B. Lippincott Company. 1912.
pp. 345. 4«. 6c?. net.

The writer is assistant-director of the Psychological Clinic at the University of
Philadelphia, which was established as long ago as 1896 by Professor Lighfcner
Witmer, having for its objects the " collection and filing of data [obtained from
mentally abnormal children] ; the development of the best clinical tests for
measuring the mentality of children ; the training of teachers and social workers
for service among mental defectives ; the diagnosis of mental diseases [in the
child] ; and the most expeditious and satisfactory methods of cormecting backward
children with the proper sources of aid for relieving or ameliorating their con-
dition." The book describes in detail the work of the clinic and the classification
of cases brought there for investigation. But its frequent crudities make it better
suited for the public, the general practitioner, and the school teacher, than for
the trained psychologist, neurologist, and psychiatrist. A more serious work,
written by a psychologist who has had fuller experience and received a medical
education, would be of great value in this country. Here such a book as that
of Dr Holmes is capable of doing considerable harm as well as good.

262 Puhlicatio7is Recently Received

The Mental and Physical Life of School Children. By Dr Peter Sandiford.
London : Longmans, Green & Co. 1913. pp. xii + 346. 4s. 6c?.

This book, says the author, " is intended to serve as a text-book, i.e. it aims at
giving, in as brief a space as possible, a large number of facts which may be
utilised in class discussion " by students in Training Colleges for Teachers. The
first two sections are devoted to the physical life of the child and to the physio-
logical basis of mental life. The next two sections are psychological, the one being
headed 'dynamic or functional psychology,' the other 'descriptive psychology';
the author elects to include instincts and memory in the former, and emotions
and perceptions in the latter section ! The three remaining sections discuss
the psychology of babyhood and adolescence, exceptional school children, and the
develo2)ment of language in children. Considering the wide ground covered, the
author has compiled a useful and interesting little work. The defects are chiefly
those inseparable from its ' scissors and paste ' character.

Mental Fatigue. By Dr TsuRU Aral New York City : Teachers' College,
Columbia University. 1912. pp.115. $1.00.

In many cases " the subject of the exi)eriment was the writer herself. But the
danger that her presuppositions affected the results was precluded by the fact that the
writer's knowledge of mental fatigue at that time was not enough to enable her to
form any expectation of what form the fatigue curve in mental work would take."
As might be expected, the results are very meagre and indefinite in comparison
with the number and length of the experiments.

Vorlesungen zur Einfiihrung in die experimentelle Fddagogik und ihre
])sychologiiichen Grundlagen. Yon Prof. Ernst Meumann. Zweiter
Band. Zweite umgearbeitete und vermehrte Auflage. Leipzig : W.
Engelmann. 1913. S. xiv + 800. M. 11 ; geb. M. 12.25.

The second of the three volumes which constitute the second edition mainly
treats of the investigation of the individual mental differences in children, and the
application and results of tests of mental and physical efficiency. No writer
has yet attempted so complete a review of the now copious literature of these
subjects.

The Interpretation of Dreams. By Prof. Sigmund Freud. Authorized
translation of third edition with introduction by Dr A. A. Brill.
London : George Allen & Co., Ltd. pp. xiii + 510.

A very readable translation of Freud's best-known work.

Psychanalysis : its theories and practical application. By Dr A. A, Brill.
Philadelphia: W. B. Saunders Co. 1913. pp. 337 13s. net.

This book presents a useful resume of Freud's views, which are all uncritically
accepted by the writer.

Memory : a contribution to experimental psychology. By Hermann Ebbing-
haus. Translated by Prof. H. A. Ruger and C. E. Bussenius. New
York City : Teachers'College, Columbia University. 1913. pp. viii+r23.
$1.00.

This is a translation of the well-known pioneer work by the late Professor
Ebbinghaus on the experimental psychology of memory, published in 1885. But
any one with only a moderate knowledge of German will find the book more
attractive and easier to read in the original than in its 'English' dress.

Publications Recently Received 263

Prestiyp. : a pui/chological study of social estivuUeti. By Lewis Leopold.
London: T. F'isher Unwin. 1913. pp. 352.

The interest of this book is ethnological and sociological ; it contains nothing of
psychological value.

I'lie I'sycholoyy of Revolution. By Gustave Le Bon. Translated by
B. MiALL. London: T. Fisher Unwin. 1913. pp. 33G. lOs. 6c?. net.

The main argument of the author, who deals chiefly with the French Revo-
lution, is that all revolutionists have " obeyed invisible forces of which they were
not the masters. Believing that they acted in the name of pure reason, they
were really subject to mystic, affective, and collective influences, incomprehensible
to them, and which we are only to-day beginning to understand." Nowhere does
the book treat more intimately with the ' psychology ' of revolution.

Modern Classical Philosophers : selections illustrating modem philosophy
from Bruno to Spencer. Compiled by Dr Benjamin Rand. London :
Constable «fc Co., Ltd. 1911. pp. xiv + 740. 10s. 6(/. net.

This book "is virtually a history of modern philosophy based not upon the
customary description of systems, but upon selections from original texts, and ui>on
translations of the authors themselve.s."

The Classical Moralists : selections illustrating ethics from Socrates to
Martineau. Compiled bv Dr Benjamin Rand. London : Constable ik Co.,
Ltd. 1910. pp. XX + 790. lOs. 6d net.

" A companion volume in the field of ethics, to the authoi-'s ' Modern Cla.ssical
Philosophers ' in the domain of philosophy."

Die Praxis der Konstanzmethode. Von Prof. F. M. Urban. Leipzig :
W. Engelraann. 1912. S. 26. M. 1.

The writer here developes a technique of the method which he hopes, by the help
of the tables he publishes and other means, will render it as generally used as the
easier, though theoretically less satisfactory, methods of least perceptible diflercuces
and of mean error.

On the Relation of the Methods of Just Perceptible Differences and Constant
Stimuli. By Dr S. W. Fernberger. Psychological Monograph, Vol. xiv,
No. 4. Princeton : Psychological Review Company. 1913. pp. 81.

The basis of this monograph is a series of experiments on two subjects,
Dr Urban and the author, in lifting weights. Six variable weights were successively
lifted, each with a standard weight of 100 grams, the standard weight being lifted
first, and space errors being eliminated. The six comparisons furnished by this
series provided data in the usual way for the constant method. But into this series
was introduced a seventh weight, which, unlike the others, was changed at every
lift in successive series in such a way as to aSbrd data for the method of just
jxirceptible differences by complete ascent and descent. By this procedure the two
methods were worked simultaneously. The threshold values obtained by the two
methods turn out to be extremely close, and, as might be expected, "the more
nearly the experimental arrangement of the method of just perceptible differences
approaches that of the method of constant stimuli, the closer do the values
under discussion coincide." But this is regarded by the author as "a curious
fact."

264 Publications Recently Received

Zur GrumUegung der Tonpsijchologie. Von Dr Geza Revesz. Lei|>zig :
Veit & Comp. 1913. S. viii+ 148. M. 4; geb. M. 5.

The writer's exi^riments on the abnormal hearing of his friend Dr Paul
V. Liebermann form the basis of these novel and important views on tone-
psychology. He is led to distinguish the ' quality ' of a tone sensation from its
'l)itch'; all tones of the same name, e.g. c^, c^, c^, are qualitatively equal, but they
diflFer widely in pitch. In the case of Dr v. Liebermann he believes that the pitch
attribute was preserved although the quality attribute had become abnormal. To
this subject a whole range of tones, from g'^ — o^J* appeared of identical quality,
viz. as ^If. AVhich particular ^ J he heard depended on the ' pitch character ' of the
tone ; thus tones between g^ and b^ were judged as ^J^, whereas tones between
c^ and b'^ were judged as g^^. Thus the writer explains this subject's occasional
answers that two successive tones in the affected region were ' not quite ' a prime.
As qualities they were identical, but their pitches were different. With the same
subject the tone Ci had a g quality. Consequently the notes Cq, Cj, given
successively, were heard as a fourth. At the same time the subject admitted that
the distance appeared much greater than a fourth — jierhaps a major seventh or an
octave. So too he declared that Gi, Cj, might be either a prime or an octave,
explaining that the distance was too great for a prime and too small for an octave.
Hence, says the author, what the subject heard was a g quality, preserving the
pitch of C, but identical neither with Gi nor with G,^. In a case when d^ gave the
quality e, if c^ and d^ were sounded successively, the subject would at once judge
the interval as a major third (c — e) ; if, however, the same tones were simultaneously
sounded, the subject recognised by its degree of fusion that the inten'al was a major
second. Thus the author separates the former judgment — which he calls one of
interval and ascribes to diffierences in quality and pitch — from the latter to which
he gives the name ' orthosymphouy,' and which he ascribes to differences in pitch
only. Differences in quality and their interrelation determine what he calls ' size of
segment ' ; differences in pitch determine ' tone distances.'

Die Beziehungen der Psychologie zur Medizin und die Vorbildung der
Mediziner. Von Dr W. Peters. Wiirzburg : Curt Kabitsch. 1913.
S. iv + 33. M. 1.20.

The author, a graduate in philosophy, follows several recent writers, e.g. Kiilpe
and Marbe in Germany, and S. I. Franz, Adolf Meyer, Southard, Watson and
Morton Prince in America, in urging that a prominent place should be foimd for
psychology in the curriculum of the ordinary medical student. He traces the
history of the development of psychology as an independent science, and treats
of its connexion with physiology, pharmacology and psychiatry (with special refer-
ence to cerebral localization, Pawiow's conditioned reflexes, Korsakoff^'s psychosis,
mnemasthenia, and hysteria), and its special interest for the school and prison
doctor.

The Distinction between Mind and its Objects : the A damson Lecture for 1913
with an Appendix. By Bernard Bosanquet. Manchester : University
Press, pp. 73. Is. net.

The author's main conclusions in this interesting lecture may be thus sum-
marised. " Objects of finite mind, in short, and finite minds in themselves, are
bound, after our discussion of physical realism, to strike us as details of reality
essentially continuous with each other and reciprocally indispensable. But yet
any object picked out and isolated within the whole is eo ipso not-mental, for you
have taken it ajwirt from the life of the whole, and have, by abstraction, killed and
stuff"ed it for examination."

Volume VI FEBRUARY, 1914 Parts 3 and 4

FREUD'S THEORY OF THE UNCONSCIOUS^

By WILLIAM BROWN.

From the Psychological Laboratory, King's College,
University of London.

I /. The general laws of mental process, as illuslrated in dreams

and hallucinations as well as in normal waking conscious-
ness.
II. Repression and ivish-fulfiiment.
III. Psycho-analysis and hypnotism.

Freud sunimarises his fundamental views as to the nature and laws
of working of the human mind, which he has formed on the basis of a
detailed study of dreams and functional diseases, in the final chapter
of the Traumdeutung. In fact, no one who has failed to master this most
difficult chapter can justly claim any real insight into the theoretical
and psychological aspects of Freud's work. It is only here that the
exact meanings of such conceptions as 'wish-fulfilment' {Wunscher-
filllung), 'repression' (Verdrdngung), and the 'censor' (Zensur) are to
be found, and the popular and figurative nature of much of the Freudian
terminology is corrected. I shall therefore make this chapter the basis
of my discussion.

I.

It is well known that dreams, like hysterical symptoms, are regarded
by Freud as being the disguised fulfilments of repressed wishes I The
manifest content of a dream is made up of a collection of memories
from the waking life joined together by the most superficial forms of

1 Bead before Section I (Subsection of Psychology), British Association for the
Advancement of Science, Birmingham, 1913.

2 For a detailed description of the dream-theory, see my two articles on "Freud's
Theory of Dreams," Lancet, April 19th and April 26th, 1913.

J. of Psych. VI 18

266 Freud's Theory of the Unconscious

association. Organic sensations and other sensory disturbances occur-
ring during sleep, if not sufficiently intense to produce awakening, are
either ignored or woven into the texture of the dream by arousing
corresponding memories after the manner of an illusion. The method
of psycho-analysis, to which we shall refer again later, enables us to
find a meaning for this dream-formation in a set of latent dream
thoughts which are invariably of the nature of wish-fulfilments. Since,
according to Freud, the repressed wishes to which hysterical symptoms
likewise point are always derived from an infantile source, he is strongly
inclined to the view that the dream wishes are also either themselves
infantile wishes or else wishes analogous to and sustained by wishes
dating from the period of early childhood. He admits that this view
has not yet been conclusively proved, but contends that it cannot be
disproved. A large proportion of the dream-interpretations hitherto
made do, as a fact, point to infantile wishes as the underlying motive
power.

The discrepancy between the manifest dream content and the latent
dream thoughts is due primarily to the resistance of the endopsychic
censor. In order to evade this resistance and reach consciousness, the
latent wishes undergo certain changes which may be summed up in
the words ' condensation,' ' displacement,' ' dramatization,' and ' second-
ary elaboration.' In ' condensation ' the numerous dream thoughts are
replaced by a much smaller number of ideas selected because they act
as nodal points in many intersecting trains of ideas and allude to these
rather than directly represent them. ' Displacement ' refers to the
shifting of psychic accent from one part to another of the manifest
content whereby the direct correspondence between it and the latent
content is masked. Affects may also be displaced to produce the same
result. ' Dramatization,' or regard for dramatic presentability (Ruck-
sicht auf Darstellbarkeit), is provided for by the important process of
' regression,' in which the dream thoughts are reduced to their raw
material, viz. sensory (chiefly visual) presentations, of hallucinatory
vividness. These three changes constitute what is known as the
'dream-work,' and are characteristic of a form of mental activity
neglected by normal psychology and almost unknown to normal waking
consciousness, but identical with that responsible for the symptoms of
hysterical patients. The fourth change, 'secondary elaboration,' is a
process akin to that of waking consciousness, being an attempt to
rationalise these strange and perplexing dream-formations and knit
them up into a story or event with some degree of coherence. This

William Brown

267

process continues after the dreamer awakes, and is one of the causes of
the falsification of his memory of the dream during the following day.

Freud, in his attempt to form a general explanatory system within
which these various psychical processes may be co-ordinated and rendered
intelligible, finds it convenient to approach the subject with a number
of ' auxiliary ideas ' which, like auxiliary equations in mathematics, act
as a sort of scaffolding in the discussion and are to be abandoned or
drastically modified later on, according to the needs of the argument.
Borrowing from Fechner the idea of a difference of ' psychic locality ' in
dream-formation, he conceives the mind as a system made up of a
number of subsidiary systems placed in a definite order one behind
another, so that mental activity will involve the excitation of these
systems in a definite sequence. The spatial idea is here used meta-
phorically and has no necessary relation to the neural changes under-
lying mental activity. The following diagram ^ sums up this first
tentative hypothesis:

P. Mi

m.

Me

m.'

Unc.

Foi

P. represents the perceptual system, devoid of memory. Mem. is
the system conserving the lasting traces of individual perceptions, in
firont of which are situated a series of other memory-systems, Mem,'
corresponding to the various forms of association between these
memories — simultaneity, similarity, etc. — and also, presumably, to
higher thought-relations. Normal functioning of the psychical ap-
paratus involves a transmission of excitation from the perceptual
system in a progredient direction through the various memory systems
to discharge itself eventually in motor innervation. But Freud draws
an important distinction between two classes of memories or uncon-
scious processes, one of which (Forec.) is in more immediate relation to
movement (M.) than the other (Unc). It is called the ' preconscious' (das
Vorbewusste) or foreconscious, and the other is the ' unconscious ' (das
Unhewusste). Excitations in the preconscious can reach consciousness
and pass over to movement so soon as they attain a certain degree of
intensity and thus attract sufficient attention to themselves. Those in

^ Truumdeutung (Brill's translation), 429.

18-2

268 Fj-end's Theory of the Unconscions

the unconscious can only reach consciousness and control of the motor
system by passing through the preconscious. In so doing they undergo
certain changes. If, however, one inferred from this that, according to
Freud, the preconscious is 'the censor' {die Zensur) of the dream-
theory, one would probably be wrong, for he distinctly refers to the
latter as "the resistance watching on the boundary between the un-
conscious and the preconscious\" and in several other passages seems
to make it clear that his conception of the censor is that of a ' non-
conscious resistance' situated between the two systems of the uncon-
scious. There is also a second censor between the preconscious and
consciousness I For consciousness, in Freud's view, is to be regarded as
simply a " sense organ for the perception of psychic qualities^," and even
ideas in the preconscious may, if objectionable, be denied entrance to
consciousness. The various processes we have been hitherto describing,
and in fact all those of which the mind is capable, are to be regarded as
running their course independently of consciousness. The true function
of consciousness will be revealed later, when we come to closer grip with
the central problem of psychological explanation. We may, however,
conveniently quote at this stage of our discussion the significant words
of Freud on ' unconscious psychical process ' : " Everything conscious
has its preliminary step in the unconscious, whereas the unconscious
may stop with this step and still claim full value as a psychic activity.
Properly speaking, the unconscious is the real psychic ; its inner nature
is just as unknoivn to us as the reality of the external world, and it is
just as imperfectly reported to us through the data of consciousness as is
the external world through the indications of our sensory organs*." In
this passage Freud is using the term ' unconscious ' in the wider sense
subscribed to by many modern psychologists, but in his own conception
of the unconscious, as distinguished from the preconscious, we have an
entirely new contribution to psychological theory. Freud's unconscious
comprises the memories and mental processes of very early childhood,
which have been repressed or abandoned in later life but which still
retain their power of indirectly influencing consciousness by transferring
the energy at their disposal to analogous ideas repressed from the
preconscious, thus making these also unconscious.

Before leaving this first approximation to an explanation of the
working of the mind, we may use it to illustrate what is meant by

1 Op. cit. 430. 2 Op. cit. 490. » Op. cit. 121, 453, 488.

* Op. cit. 486. For a similar view of psychology as the science of the Mwconscious, see
my " Epistemological Difficulties in Psychology," Proc. Aristot. Soc. 1909-10, x. 63-76.

William Brown 269

* regression ' in the Freudian system. Regression occurs when the
excitation within the psychical apparatus takes a regressive instead of
a progressive direction. This is, in Freud's view, the cause of the
hallucinatory nature of dreams, and indeed of all hallucinations. Re-
pelled by the censor and attracted by infantile wishes in the uncon-
scious which transfer to them their energy, the latent dream thoughts
abandon the progressive path through the preconscious towards move-
ment and consciousness, and pass backwards through the various memory
systems until they reach the perceptual system. The intensification
necessary for this penetration to the perceptual system is mainly
accounted for by the processes of condensation and displacement,
although in the case of dreams the cessation of the progressive stream
of excitation present in waking life is a contributory factor. In this
way consciousness is aroused at the sensory end of the apparatus, and
the dream has succeeded in evading the censor rather than surmounting
it. The lowered activity of the censor during sleep, which Freud also
assumes, only explains the formation of those few dreams which lack
the dramatic character, and come to consciousness as thoughts, not as
images. These pursue the progressive course throughout. In regres-
sion, on the other hand, " the structure of the dream thoughts is broken
up into its raw material^" and the thoughts are transformed into
images.

As an instance of a hysterical hallucination produced by the same
mechanism we may mention the case, given by Freud, of a twelve-year-
old boy who was prevented from sleeping by a terrifying vision of green
faces with red eyes. This hallucination corresponded to a suppressed
memory, dating four years back, of a boy companion who had taught
him many bad habits, including onanism. The patient's mother had
remarked at the time that this boy had an unhealthy greenish counten-
ance and red-rimmed eyes, and warned her little son that such wicked
boys became backward at school and die young.

The explanations of hallucinations given in psychological text-books
are for the most part physiological in nature and tend to slur over, if
indeed they do not ignore, the problem of the ' meaning ' of the hallu-
cination. Thus James^ explains these phenomena in the following
way : The sensory vividness of an actual percept is due to (or, rather,
correlated with) the passage of afferent nerve currents at high potential
across the synapses of the sensory centre in the cerebral cortex. A

1 Op. cit. 431.

2 Principles of Psychology, 1890, ii. 123, 124.

270 Fre^tfVs Theorij of the Un con scions

mental image, on the other hand, lacks sensory vividness because it is
due to the excitation of the sensory centre by nerve currents of low
potential flowing along association fibres from other parts of the cortex.
While falling asleep, however, or under abnormal conditions in waking
life, the synaptic resistances of the centre increase, so that the nerve-
currents of low potential which are continually flowing to it along
association paths can no longer pass through it and drain away into
efferent fibres. The result is that nervous energy accumulates, the
potential of the nerve-currents rises until it once more overcomes the
synaptic resistance and produces an ' explosive discharge ' of the nerve-
cells corresponding in intensity to that accompanying perception.
Hence the subjective hallucinatory experience. This theory assumes
an identity of physiological site for the percept and the corresponding
mental image, whereas for Freud the P system and the Mem. system
are quite distinct, since he considers that the former must be quite
devoid of memory if it is to perform its functions adequately. More-
over Freud does at least attempt to explain why certain mental con-
tents are chosen to form an hallucination and not others. Those are
selected which are recent and in themselves unimportant, since they
have not had time or opportunity to enter into far-reaching associative
connexions in the preconscious, and therefore are suitable material to
receive the ' transference ' ( Uebertragung) of energy from desires in the
unconscious. Their unimportance and superficial connexions with one
another also protect them from the censorship.

The idea of regression is also to be found in McDougall's explanation
of hallucination. McDougall writes : " It is known that in many cases
of hallucination there is chronic irritation of a sense-organ ; in cases of
auditory hallucination, for example, it has sometimes been found that
there is disease of the ear leading to continual irritation of the sensory
neurones. We may suppose that disease induces an irritable weakness
of a certain system of paths in one of the sensory areas of the cortex, so
rendering them paths of abnormally low resistance, and that any im-
pulses passing up from the corresponding sense-organ, and possibly also
from other sense-organs, are therefore liable to be diverted to them
from their normal paths, so re-exciting the chains of cortical neurones in
their whole length, and producing a representation of sensory vividness'."
In Freud's theory, however, such an "irritable weakness of a certain
system of paths " would not in itself sufiice to produce the hallucination

^ W. McDougall, Physiological Psijcholotjy, 86 (italics mine).

William Brown 271

without the aid of energy from the powerful wishes of the unconscious
which also determines the exact form which it shall take.

II.

In attempting a more accurate statement of his theory, Freud
attributes to his unconscious and preconscious systems two different
kinds of psychical process, viz. a 'primary process' and a 'secondary
process' respectively. The one fundamental difference between these,
which accounts for all the others, is that the secondary process is
capable of ' inhibition ' while the primary is not. The primary is the
primitive and infantile, although even in earliest childhood it is probably
not entirely unaccompanied by at least the germs of the secondary
process. Its activity is limited to that of 'wishing,' and it strives to
satisfy desire solely by reviving the memories of previous satisfactions
and by intensifying them to hallucinatory vividness. Since permanent
satisfaction is not to be obtained in this way, the mind has had to
develop a secondary process which treats the memory of a previous
satisfaction not as an end in itself but merely as a means to a more
round about process of reinstating the actual satisfying object, or one
like it. The primary process strives after a ' perception identity,' the
secondary after a ' thought identity.' In both cases the motive power
is a wish, since, as Freud says, "nothing but a wish can impel our
psychic apparatus to activity \"

But Freud's fundamental explanatory principle is that of a Besetz-
ungsenergie, or ' occupation energy,' which is subjected to different
distributions within the psychic apparatus under different circumstances.
Within the system of the unconscious this occupation energy is capable
of a complete displacement from one presentation to another, so that
ultimately one or a few presentations, which may be regarded as repre-
senting the rest, become sufficiently intense to penetrate to the perceptual
system of the psychic apparatus. This is, of course, the primary process
of wish-fulfilment ; and the processes which we have hitherto classified
under the heading of the ' dream-work ' are nothing but aspects of the
primary process. It is the same process which is responsible for the
symptoms of hysteria, where the effects of condensation (' identification '
or ' composition ') and regression are clearly visible.

^ Op. cit. 447. This sentence indicates one fundamental weakness of Freud's system,
since conations below the ideational level are, of course, motive forces of the mind.

272. Freud's Theory of the Unconscious

The distribution of ' occupation energy ' under the influence of the
secondary process is quite a different one. Freud writes: — "The manifold
activity of the second system, tentatively sending forth and retracting
energy, must on the one hand have full command over all memory
material, but on the other hand it would be a superfluous expenditure
for it to send to the individual mental paths large quantities of energy
which would thus flow off to no purpose, diminishing the quantity avail-
able for the transformation of the outer world. In the interests of
expediency I therefore postulate that the second system succeeds in
maintaining the greater part of the occupation energy in a dormant
state and in using but a small portion for the purposes of displacements"
This is what he calls regulation by the 'principle of the smallest
expenditure of innervation ' {Prinzip des kleinsten Innervationsauf-
wandes).

Another principle which is obeyed by both systems is the ' principle
of pain ' ( Unlustprinzip). This is simply the deviation of the psychic
process from any memory involving pain. By virtue of it, " the first
system is altogether incapable of introducing anything unpleasant into
the mental associations. The system cannot do anything but wish^."
Such a mere turning away from a painful memory is " the model and
first example oi psychic repression (Verdrangung)."

The second system retains control over painful memories in the face
of this principle by so ' occupying ' them that the pain attaching to
them — which, like pleasure, is an efferent process analogous to a motor
or secretory innervation — is almost completely inhibited. Now, owing
to the insufficient development of the secondary process in the first
two or three years of childhood, the memories and wishes of this period
temain beyond control and inaccessible to the consciousness of later life.
Some of these unconscious wishes are in conflict with the later wishes
of the preconscious, so that their fulfilment would now produce pain
instead of pleasure; "and it is just this transformation of effect" says
Freud, " that constitutes the nature of what we designate as ' repression,'
in which we recognise the infantile first step of passing adverse sentence
or of rejecting through reason^" An example of this transformation
of affect is the appearance of * disgust ' at a certain point in infantile
development while previously absent.

These unconscious inftxntile memories are the precondition of all
later repression. They are able to transfer their energy to any neglected

1 Op. cit. 475. 2 Op. cit. 476. » Oj,. cit. 479.

William Brown 273

or suppressed thoughts of the preconscious whose content may happen
to stand in some relation with their own. The preconscious then turns
away from these thoughts of transference in accordance with the principle
of pain and thus they are, as it were, drawn into the unconscious. This
deviation from thoughts 'capitalised' by wishes in the unconscious is
what is known as ' repression.' We thus see that ' repression ' ( Verdrdng-
ung) is not quite the same thing as 'suppression' (Unterdrilckung), and
has a definite technical meaning of its own in the Freudian system of
psychology.

The repressed thoughts originating from the preconscious are now
strong enough to persist in an independent and unconscious existence
of their own, but can only attain to consciousness, if at all, by pursuing
a regressive course and reaching the perceptual system. This is the
way in which hysterical symptoms — paralyses, anaesthesias, aphonias,
tics, contractures, convulsions, obsessions, phobias, etc. — originate,
although it appears that another universal condition of their production
is that a counter-wish from the preconscious, generally of the nature
of a self-punishment, should also be present and fulfilled by the same
symptoms. Hysterical symptoms are thus to be regarded as ' compromise-
formations,' satisfying as well as may be a wish from the preconscious
and one or more wishes from the unconscious. The dreams of normal
persons are exactly analogous to such symptoms, being a compromise
between the wish to sleep of the preconscious, and unconscious wishes
aroused during the previous day or in the course of the night.

An essential part of Freud's theory of the psycho-neuroses is the view
that "only sexual wish-feelings from the infantile life experience repression
(emotional transformation) during the developmental period of child-
hoods" These are directed towards the parents, or their substitutes,
and constitute the well-known ' Oedipus complex ' or ' Electra complex,'
according to the sex. It is because they are capable of an organic
re-inforcement in later life, especially at the time of puberty, that they
endanger the mental equilibrium as no other tendencies can do. Space
does not permit me to make more than this very inadequate reference to
Freud's sexual theory in the present paper, although its importance for
a true appreciation of his entire system can hardly be overestimated.
As regards 'the dreams of normal persons, Freud prefers to leave it
undecided whether these are ultimately based upon sexual wishes of the
unconscious I Indeed in some passages of the Traumdeutung, he definitely

1 Op. cit. 480. " Of. p. 481.

274 FreucVs Theory of the Unconscious

leans towards the view that tendencies like hunger, thiret, and the desire
for power are fully competent to produce dreams without further aid
from the unconscious. Anxiety dreams are certainly sexual in signi-
ficance, if not always so in origin. The feeling of anxiety is due to an
overpowering of the second system by the first, and indicates a failure
in that ' compromise ' to which we have already referred. Thus the
function of compromise-formations, such as dreams and hysterical
symptoms, is to guard against the outbreak of anxiety. Freud illustrates
this by reference to the case of agoraphobia. " Suppose a neurotic
incapable of crossing the street alone, which we would justly call a
'symptom.' We attempt to remove this symptom by urging him to
the action which he deems himself incapable of The result will be an
attack of anxiety, just as an attack of anxiety in the street has often
been the cause of establishing an agoraphobia. We thus learn that
the symptom has been constituted in order to guard against the out-
break of the anxiety. The phobia is thrown before the anxiety like a
fortress on the frontier^" But in some cases the originating cause is
the intense pain of certain organic sensations aroused during sleep,
especially with people who sufier fi-om disease of the heart or lungs.
The anxiety thus somatically aroused gains a psychical interpretation
in the dream by liberating unconscious wishes, the fulfilment of which
in face of the censorship would be accompanied by a similar feeling of
anguish.

m.

With regard to the use of ' symbols ' in dreams, it is only necessary
for us, in the interests of theory, to point out that these are not products
of dream activity. The symbolizing tendency is already present in the
latent dream thoughts, and is identical with that responsible for our
myths and legends. The predominant use which the dream makes of
such symbols is due to their dramatic fitness and their freedom from the
censorship. Although certain of these symbols tend to have the same
meaning among a whole class of individuals, it must never be forgotten
that their significance in any single case can only be accurately deter-
mined by means of psycho-analysis. It is because Pierre Janet has failed
to realise this that so much of his recent criticism of the Freudian school
is unconvincing'^.

1 Op. cit. 459.

2 P. Janet, "Psycho-analysis," XVIIth Internal. Congr. of Med., London, 1913,
Section xii. 13-64. See especially p. 26 for the point here raised.

William Brown 275

Psycho-analysis is something more than a mere catechizing of the
patient. Experience has shown that certain memories which are in-
accessible under ordinary circumstances will rise to the surface of the
mind if the patient adopts an attitude of uncritical meditation and follows
the sequence of associated ideas as they appear, rejecting none of them
however painful, objectionable, or absurd they may seem to be. In the
case of dream-interpretation, the separate sections of the manifest content
are taken as the independent starting-points for these chains of ' free '
associations; in the case of a psycho-neurosis the symptoms serve the
same purpose. It is important to realise that these chains of ideas are
not truly free or aimless associations. When, by adopting the attitude
of uncritical reverie, the patient succeeds in freeing himself from ' con-
sciously purposeful mental activity' (bekannte Zielvorstellimgen), his mind
does not cease to be purposive but is now dominated by ' unconscious
trends of activity' (unbewusste Zielvorstellungen) which determine what
ideas shall rise to consciousness. The ideas which in this way are
eventually reached are found to allude to, if not to form an integral part
of, a system of preconscious thoughts which had by transference been
dragged into the unconscious and which constitute the interpretation
of the dream or the psycho-neurotic symptom, as the case may be. The
process of psycho-analysis, by bringing these thoughts once more under
the control of the preconscious, ipso facto biings about the resolution of
the hysterical symptoms and the cure of the patient. It is in this sense
that we are to take the dictum of Breuer and Freud that " solution and
treatment go hand in hand*." The course of treatment is as a rule a
lengthy one and makes considerable demands upon the tact and energy
of the physician, since the trains of associations are being continually
interrupted by ' resistances ' which the patient is unable to cope with
single-handed, despite his best intentions, and it is only with the aid of
persistent urging on the part of the physician that the hindrances are
overcome and the ideas again continue to flow. In order that the cure
may be complete the patient must be able to live again through the
intense emotions attached to the repressed ideas and direct them upon
the personality of the physician. This indispensable cathartic process
is known as ' abreaction ' (Abreagierung).

The well-known ' word-association method ' of C. G. Jung is very
useful as an adjunct of psycho-analytic procedure, and in the case of
some of the psychoses is the only suitable method. It serves to indicate

i Op. cit. 83.

276 Freud's Theory of the Unconscious

the principal unconscious ' complexes,' i.e. systems of repressed and
emotionally-tinged ideas, from which the patient is suffering. The
clearest and most frequent sign of the existence of such a complex
is : (1) a prolonged reaction time, but it should not be forgotten that
there are other ' complex-indicators ' of equal importance. These are :
(2) a failure to react; (8) an over-reaction, giving more than is asked
for, many words, with supplementary explanations, instead of one;
(4) a repetition of the stimulus word; (5) an identical word-reaction
to the most varied stimulus words; (6) a superficial association, especially
if combined with a prolonged association time; (7) a meaningless re-
action; (8) an assimilation of the stimulus ivord, where it is misread,
misunderstood, or taken in an unusual sense under the influence of the
complex, being thus ' assimilated ' to the complex ; (9) a failure in
reproduction, the patient giving a different reaction- word on a second
presentation of the stimulus-word, although asked to reply if possible
with the same word as before'. The chief theoretical interest of Jung's
work on association is that he has succeeded in giving an experimental
proof of the validity of the main assumptions upon which Freud's
psycho-analytic technique is based.

The relation of psycho-analysis to hypnotism is a problem of great
interest, which I am inclined to think is still awaiting solution,
despite the claim to a satisfactory understanding of it made by the
Freudians. Ferenczi^ has carried out psycho-analyses of patients whom
he had previously treated by hypnotism, and considers that the results
confirm Freud's view that in hypnotism unconscious sexual tendencies
of the patient are ' transferred ' from their original object, the parent,
to the person of the hypnotist. " Hypnosis is a special form of arti-
ficially increased suggestibility," and suggestibility is nothing more than
the survival in the unconscious of the child's readiness to believe blindly
and obey uncritically those whom it loves. Now, since the symptoms
in hysteria are likewise perverted satisfactions of psycho-sexual wishes
emanating from the infantile unconscious, it follows that the removal
of such symptoms by hypnosis or by the milder forms of suggestion
is merely a case of replacing them by another symptom, viz. "psycho-
sexual dependence upon the physician." For this reason hypnotic cures

^ Some results obtained by this method will be found in a short paper by me, "A Case
of Extensive Amnesia of Remote Date cured by Psycho-Analysis and Hypnosis," Brit.
Med. J., Nov. 8th, 1913.

2 Ferenczi, " Introjektion und Uebertragnng," Jhrb. f. psychoanal. u. psychopath.
Forsch. 1909, i. See also Ernest Jones, " The Action of Suggestion in Psychotherapy,"
J. of Abnorm. Psychol. Dec. 1910, v.

William Brown 277

are so seldom permanent. Psycho-analysis, on the other hand, avoids
this unsatisfactory result by dragging up the psycho-sexual tendencies
into consciousness, so enabling the patient to understand their true
nature and to 'sublimate' them, i.e. direct them to useful social
activities.

Janet considers that the attachment of the patient to his physician^
upon which this theory is based, is not to be so simply explained. He
writes: "Get attachement se pr^sente de bien des manieres differentes
et semble dependre de phenomenes psychologiques tres divers dans
lesquels inter viennent suivant les cas des suggestions, des aboulies,
I'incapacite a conclure par soi-meme, le besoin d'etre compris, le besoin
d'etre dirige et surtout le besoin d'etre excite si important chez les
d^sprimes\" Only on the assumption that every form of docility is
sexual in origin can Ferenczi's theory lay claim to truth. The question
is largely one of fact, and although Janet's extended and world-famed
experience as a hypnotist lends great weight to his words, we cannot
overlook the empirical results of psycho-analysis ; and if Ferenczi's
comparative investigations are confirmed by independent and unbiassed
observers, his theory must be accepted. Even then a number of out-
lying questions of great importance remain to be answered. For
example, what is the cause of the remarkable broadening, of the field of
consciousness and improvement of memory that occur in the hypnotic
state, prior to any suggestions made by the hypnotist ? In the case
refeiTcd to on page 276 of an extensive amnesia of thirteen years
standing — the loss of memory covered the period fi-om September, 1897,
to February, 1900 — almost all the essential memories reappeared directly
the first hypnotic slumber had been induced, without any special
prompting from myself. I had, during the previous fortnight, plied
the patient repeatedly with word-association tests without much
apparent success, but am inclined to think that this treatment acted
as a very powerful predisposing influence towards hypnosis, since the
patient, who had never been hypnotized before and had repeatedly
expressed great scepticism as to anyone, myself included, being able to
hypnotize him, went into the hypnotic trance with complete loss of
consciousness in less than ten minutes. Moreover, in the course of the
word-association tests he frequently forgot the stimulus- word, and some-
times also the reaction-word, immediately after replying. This suggests
a close relation between the state of hypnosis (before any suggestions

1 P. Janet, op. cit. 38.

278 Freud's Theory of the Unconscious

have been given) and the state of mind during psycho-analysis, — a
relation which has not escaped Freud's notice, for he writes (of
psycho-analysis): "As may be seen, the point is to bring about a
psychic state to some extent analogous as regards the apportionment
of psychic energy (transferable attention) to the state prior to falling
asleep {and indeed also to the hypnotic stateY" This resemblance is
worthy of further investigation.

The remarkable physiological manifestations often observed in
hypnotized subjects also still await an explanation that will be com-
pletely satisfactory to the scientific mind. The Freudians may retort
that these are identical in nature with the symptoms of conversion-
hysteria, thus agreeing with the dictum of Charcot that "hypnosis
is an artificial hysteria." But this does not help us much, for the
wish-fulfilment theory merely indicates the psychical significance of
these symptoms; the psycho-physiological or purely physiological
changes which occur in the nervous system must form an integral
part of any complete causal explanation. Freud is fully alive to
this lacuna in his theory, though his disciples tend to push it into
the background and often ignore it completely. He writes in reference
to the inhibitory functions of the secondary process: "The mechanism
of these -processes is entirely unknown to me ; anyone who wishes to
follow up these ideas must try to find physical analogies and prepare
the way for the visualising of the dynamic process ( Veranschaulichung
des Bewegungsvorganges) in (the theory of) the stimulation of the
neuron. I merely hold to the idea that the activiuy of the first p'sychical
system is directed to the free outflow of the quantities of excitement,
and that the second system brings about an inhibition of this outflow
through the energies (Besetzungen) emanating from it, i.e. it produces a
transformation into dormant energy {ruhende Besetzung) involving a
raising of the levels" He nowhere says whether his Besetzungsenergie
is mental or physical, but we can hardly refuse to assume that it has at
least a physiological correlate in the form of nerve-energy; and since he
definitely states that the systems of the psychic apparatus have nothing
psychic in themselves^, being analogous to the lenses of a telescope
which produce virtual images corresponding to the objects of internal
perception {i.e. psychical objects), it is only to the anatomy and

^ Traumdeutung (Brill's translation, 85 ; italics mine).

■•* Op. cit. 475, I have made two slight alterations in Brill's translation.

3 See p. 484.

William Brown 279

physiology of the central nervous system that we can turn for further
explanation.

Now it seems to me that McDougall's interesting theories as to
the physiological processes underlying psychical activity throw much
additional light on the psycho-physics of inhibition, repression and
symptom-formation ^ McDougall regards the passage of nervous energy
{neurokyme) across the synapses of the cerebral cortex as the physio-
logical correlate of psychical process, and would explain inhibition as
a secondary effect of the act of attending. In attending to one object
or concentrating the mind on one form of self-activity, neurokyme is
concentrated, raised to a higher potential, in a particular system of
neurons, and by virtue of the lowered resistance of the intervening
synapses drains energy from all neighbouring systems along collaterals
which extend from their neurons to these synapses. McDougall supports
this theory by numerous observations on the psychology of sensation and
perception that are not easily explained in any other way. He regards
the special inhibitory nerves connected with the autonomic nervous
system as a primitive device which has been superseded by this more
efficient mechanism in the course of evolution of the central nervous
system. The repression of a mental tendency would thus correspond
to a withdrawal of neurokyme from the correlated system of nerve-
arcs ; and the resistance of the censor would correspond to an actual
heightened resistance of synapses that divide the wide system of
interrelated sub-systems functioning as the preconscious from that
functioning as the unconscious. It is at least probable that Freud
means by the censor something unconscious, for in his analogy of the
telescope he compares it to the "refraction of rays in their passage into
a new medium^"

McDougall's theory will probably need much further elaboration
and (possibly) modification to make it fit all the facts now known
about functional diseases. That the altered conductivity of certain
synapses plays a decisive part in the causation of these disorders
there can be little doubt. I recently had the opportunity of observing
a case of hysterical astasia abasia^ in a woman patient over forty years
old, the immediate or occasioning cause of which was an operation for

1 W. McDougall, *' The Seat of the Psycho-Physical Processes," Brain, 1901, xxiv. ;
" The Nature of Inhibitory Processes within the Nervous System," ibid. 1903, xxvi. ; " The
State of the Brain during Hypnosis," ibid. 1908, xxxi.

« Oy. cit. 484.

3 A functional inability to stand or walk.

280 Freud's Theory of the Unconscious

appendicitis. The woman had to re-learn, slowly and painfully, the art
of walking which she seemed to have completely forgotten. But close
observation showed that the chief feature of the symptom was a lack of
co-ordinating power of a particular kind. Whereas in normal walking
contraction of the flexor muscles is accompanied by automatic relaxation
of the corresponding extensors, and vice versa, in accordance with
Sherrington's law of reciprocal innervation, here contraction of both
sets occurred simultaneously. The patient while putting her leg forward
seemed at the same time to be trying to draw it back, and similarly
with other movements.

Since normal reciprocal innervation is best explained by McDougall's
theory as a reciprocal inhibition (this being caused by the drainage of
innervation energy from the less intensely charged chain of neurons
to the neuron-chain carrying the increased innervation necessary for
the initiation of a movement), our case is one of functional disturbance
of this mechanism in the form of altered resistances at the synapses.
Paralyses, contractures, and in fact all motor symptoms observable in
hysterical patients may be physiologically explained in exactly the same
way. It is a short step from this to a similar explanation of sensory
symptoms. Such explanation of course merely supplements, it does
not exclude, a psychological interpretation in terms of 'meaning,' such
as Freud gives.

{Manuscript received 29 November, 1913.)

THE ANALYSIS OF SOME PERSONAL DREAMS,
WITH REFERENCE TO FREUD'S THEORY OF
DREAM INTERPRETATIONS

By T. H. PEAR.

/. Introduction.

II. Points in the dreams which hear a relation to Freud! s theory.

HI. Some remarks on Freud's theory.

I V. The dreams ; their analysis and interpretation.

V. Conclusions.

L Introduction.

The last few years have seen a noteworthy change in the attitude of
psychologists towards the dream. It may fairly be said that from any
existing text-book of general psychology one can gain very little know-
ledge on this subject, which has always been of intense interest to the
non-scientific public. Even the sparse details which may be gleaned
are usually of such a vague and general nature that they are of little use
in helping the psychologist to understand the relation of the dream to
other mental processes, particularly to those of normal and abnormal
waking life. Since the publication of Freud's theory^, however, the
study of dreams has naturally received a great impetus. But it is
scarcely necessary to point out that the examination of a theory so
complex as this will involve the investigation of a very large number of
dreams of different people, in order to ascertain the extent to which the
dreams of various persons exhibit individual differences, as well as the
nature of these differences, and their relation to different types of mind.

The following article will attempt to analyse in detail, and to account
for, two dreams of the writer, and it will indicate their relation to

' Amplified from a paper read at the meeting of the British Association for the
Advancement of Science, Birmingham, September, 1913.

2 The Interpretation of Dreams (Translation by Brill of the 3rd edition of Die
Traumdeutung), London, 1913.

J. of Psych. VI 19

282 The Analysis of some Personal Dreams

Freud's theory of dreams without entering into a detailed discussion of
his general psychological theory, which is at present so much in debate.
At the present stage of the controversy it seems more profitable to
examine minutely fact after fact of mental life by which the special
validity of the single parts of his theory may be tested, and to postpone
the examination of his theories a& a whole until more evidence is forth-
coming from the experiences of many normal persons, of widely different
mental characteristics. That this evidence is by no means complete is
apparent to anyone who is conversant with modem psychological and
psycho-pathological literature S and this fact forms the justification for the
appearance of this paper.

One of the chief objections frequently raised against Freud's theory
of dreams is that the dreams upon which his explanations are founded
were either his own or those of persons whose mental condition was so
abnormal that at the time they experienced the dreams they were
undergoing medical treatment on this account. It has sometimes been
said that, even if the theory be true for abnormal patients it need not
be valid as an explanation of the dreams of normal people. But, since
it is generally admitted that mental normality and abnormality are
separated only by an infinite number of gradations, it is impossible to
believe that at some point in the transition the ' normal ' set of laws
gives place to ' abnormal ' laws. Freud says^ "The objection that no
deduction can be drawn regarding the dreams of healthy persons from
my own dreams and from those of neurotic patients may be rejected
without comment." But, whatever may be the theoretical justification
for this remark, it is obvious that we need a scientific statement of the
dream-phenomena occurring in ordinarily healthy minds.

Up to the present time, only a few workers have paid careful and
systematic attention to their own dreams. Their results have been of
great value in many ways : they have studied the «naterial of the dream,
its images, thoughts and feelings ^ but few of their findings can be used
as a means of testing Freud's theory. In the first place they have, to use

' See M. Isserlin, "Die Psychoanalytische Methode Frauds," Ztsch, f. d. ges. Neurol,
u. Psychiat. Bd. i. Heft i., also Ergeh. d. Neurol, u. Psychiat. 1911. A. Kronfeld, "tjber
die psychologischen Theorien Freuds und verwandte Anschauungen," Arch. f. d. ges.
Psychol. XXII. 2, 3.

2 Op. cit. 482.

^ Especially interesting examples of this kind of investigation have been recently
furnished by F. Hacker, " Systematische Traumbeobachtuugen mit besonderer Beriick-
aichtigung der Gedanken," Arch, f, d, ges. Psychol, 1911, xxi. 1-3, 1-131, and P. Kohler,
ibid. 1912, xxiii. 415-489.

T. H. Pear 283

a figurative expression, studied the minute anatomy of the dream rather
than its behaviour ; and few of them have paid atteation to the dream
when taken as a whole. This, however, is a point which is insisted
upon by Freud. Secondly, most of these detailed studies are concerned
with the ' apparent dream ' (the ' manifest content ' of Freud), and
criticism of the dream theory on the basis of such work misses Freud's
main point, viz. that his theory refers specifically to the ' latent content,'
viz. the thoughts which are at the basis of, the dreamt So we still need
a careful examination of the dreams of normal persons, noted without
delay on awaking.

The two dreams recorded here occurred in the sleep just before
awaking at the usual time in the morning, and in connexion with
them it should be remembered that, as Hacker has pointed out*, one
cannot assume that the dreams of deep sleep are of this nature.

II. Points in the dreams which bear a relation to
Freud's theory.

Full accounts of Freud's theory of dreams will be found in his own
book and in articles on the subject by Ernest Jones^ Ferenczi^ and
William Brown''. The main points of the theory which may be examined
in the dreams analysed in this paper are the following :

1. The relation of the ' manifest content ' to the ' latent content ' or
the dream thoughts.

2. The ' censorship ' of consciousness.

3. The ' dream-work,' which produces the distortion necessary to
evade the ' censor,' including the processes of dramatization, symbolism,
condensation and displacement.

4. The dream as the fulfilment of a wish.

5. The dream as the disguised fulfilment of a repressed wish.

6. The relation in the dream of the conscious to the unconscious
wishes.

7. The r61e of the infantile wish in the dream.

1 Freud himself says in The Interpretation of Dreams (p. 114), " It is quite incredible
with what stubbornness readers and critics exclude this consideration, and leave unheeded
the fundamental differentiation between the manifest and the latent dream content."

2 Op. cit. 123. 3 ^,ngr. j. of Psychol. 1910.
* Lancet, April 19th and 26tb, 1913 ; also this Journal, 1914, vi. 265-280.

19—2

284 The Analysis of some Personal Dreams

III. Some remarks on Freud's theory.

It is, unfortunately, not an easy matter to obtain a clear and un-
equivocal statement of Freud's own theory. The difficulty is increased
when one consults the expositions of the theory by other workers. In
the first place, Freud's own treatment of the subject is not free from
inconsistencies. Questions on which, in the earlier part of his book,
he expresses a guarded opinion are answ^ered more dogmatically in
the later chapters. In fact, the last chapter is rather a statement of
opinion than a scientific treatment of the subject. This appears when
we consider two important points :

(1) That the interpretations of his own dreams (which may be
considered the most valuable evidence for his theory) in the earlier part
of the book do not themselves form a factual basis for the extensions of
the theory made in the later theoretical treatment.

(2) That the later chapters occasionally conflict with statements
made in the earlier chapters.

It is instructive to take Freud's own statements concerning two
most important points in his theory, viz. the relation of the unconscious
to the conscious wishes in the dream, and the role of the infantile wish
in the dreamt ;

(1) The Wish in the Dream.

Page 100. " The dream represents a certain condition of affairs as I should wish
it to be ; the content of the dream is thus the fulfilment of a wish ; its motive is
a wish."

Page 102 (referring to the dream which was used on page 100). "I do not wish
to claim that I have revealed the meaning of the dream entirely, or that the inter-
pretation is flawless."..." When the work of interpretation has been completed the
dream may be recognised as the fulfilment of a wish."

Page 107. " The dreams of little children are simple fulfilments of wishes...."

Page 436. " The undisguised wish-fulfilments were chiefly found in children, yet
fleeting open-hearted wish dreams seemed (I purposely emphasise this word) to occur
also in adults."

Page 438. "I have a strong doubt whether an unfulfilled wish from the day
would suflSce to create a dream in the adult. It would rather seem that as we learn
to control our impulses by intellectual activity, we more and more reject as vain the

' I am well aware of the danger of unfair treatment in taking sentences out of their
context, but I believe that in the cases cited this can be done with scrupulous justice,
since the meaning of the sentences, and that of the connexions in which they occur, is so
clear that misunderstanding seems impossible. The quotations are taken from the latest
available edition of Freud's work cited on p. 281.

T. H. Pear 285

formation or retention of such intense wishes as are natural to childhood. In this,
indeed, there may be individual variations ; some retain the infantile type of psychic
processes longer than others. The differences are here the same as those found in
the gradual decline of the originally distinct visual imagination.

In general, however, I am of the opinion that unfulfilled wishes of the day are
insufficient to produce a dream in adults. — I believe that the conscious wish is a
dream inciter only if it succeeds in arousing a similar unconscious wish which
reinforces it^"

(2) The Infantile Wish in the Dream.

If we examine the chief statements concerning the part played by
the infantile wish in the dream, we find :

Page 160. " In another series of dreams we learn from analysis that the wish
itself, which has given rise to the dream, and whose fulfilment the dream turns out
to be, has originated in childhood, — until one is astonished to find that the child
with all its impulses lives on in the dream."

Page 162. "Another case establishes the fact that although the wish which
actuates the dream is a present one, it nevertheless draws great intensification from
childhood memories."

Page 166. "The deeper one goes in the analysis of dreams, the more often one
is put on the track of childish experiences which play the part of dream sources in
the latent dream content."..." As a rule, of covu-se, a childhood scene is represented
in the manifest dream content only by an allusion, and must be extricated from the
dream by means of interpretation. The citation of examples of this kind cannot have
a very convincing effect, because every guarantee that they are experiences of child-
hood is lacking ; if they belong to an earlier time of life, they are no longer recognised
by our memory. Justification for the conclusion that such childish experiences
generally exist in dreams is based upon a great number of factors which become
apparent in psychoanalytical work, and which seem reliable enough when regarded
as a whole. But when, for the purposes of dream interpretation, such references of
dreams to childish experiences are torn from their context, they will perhaps not
make much impression, esiJecially since I never give all the material upon which the
interpretation depends."

Page 171. "My collection, of course, contains an abundant supply of such
patients' dreams, whose analysis leads to childish impressions that are remembered
obscurely or not at all, and that often date back to the first years of life. But it is
a mistake to draw conclusions from them which are to apply to the dream in general ;
we are in every case dealing with neurotic, particularly with hysterical persons, and
the part played by childhood scenes in these dreams might be conditioned by the
nature of the neurosis, and not by that of the dream. However, I am struck quite
as often in the coiurse of interpreting my own dreams, which I do not do on account
of obvious symptoms of disease, by the fact that I unsuspectingly come upon a scene
of childhood in the latent dream content, and that a whole series of dreams suddenly
falls into line with conclusions drawn from childish experiences."'

1 The italics are those of Freud.

286 The Analysis of some Personal Dreams

Pages 183-4. " Since I have learnt, further, from experience in dream analysis
that there always remain important trains of thought proceeding from dreams whose
interpretation at first seemed complete (because the sources of the dream and the
actuation of the wish are easily demonstrable), trains of thought reaching back into
earliest childhood, I have been forced to ask myself whether this feature does not
constitute an essential condition of dreaming. If I were to generalise this thesis,
a connexion with what has been recently experienced would form a part of the
manifest content of every dream, and a connexion with what has been most
remotely experienced, of its latent content ; and I can actually show in the analysis
of hysteria that in a true sense these remote experiences have remained recent up to
the present time. But this conjecture seems still very diflScult to prove ; I shall
probably have to return to the part played by the earliest childhood experiences, in
another direction (Chapter VII).... The dream often appears ambiguous; not only
may several wish-fulfilments, as the examples show, be united in it, but one meaning
or one wish-fulfilment may also conceal another, until at the bottom one comes upon
the fulfilment of a wish from the earliest period of childhood ; and here, too, it may
be questioned whether ' often 'in this sentence may not more correctly be replaced
by ' regularly.' "

Page 439. " I say that these wishes found in the repression are themselves of
an infantile origin, as we have learned from the psychological investigation of the
neuroses. I should like, therefore, to withdraw the opinion previously expressed
that it is unimportant whence the dream-wish originates, and replace it by another,
as follows : The wish manifested in the dream must he an infantile one^. In the adult
it originates in the Unconscious, while in the child where no separation and censor
as yet exist between Foreconscious and Unconscious, or where these are only in the
process of formation, it is an unfulfilled and unrepressed wish from the waking
state, I am aware that this conception cannot be generally demonstrated, but
I maintain nevertheless that it can be frequently demonstrated, even where it was
not suspected, and that it cannot be generally refuted.

The wish-feelings which remain from the conscious waking state, are, therefore,
relegated to the background in the dream formation. In the dream content I shall
attribute to them only the part ascribed to the material of actual sensations during
sleep (see p. 185)."

Page 447. " The dream is a fragment of the abandoned psychic life of the child^."

Page 481. "/ will leave it undecided whether the posftUate of the sexual and
infantile m,ay also be asserted for the theory of the dream; Heave this here unfinished
because I have already passed a step beyond the demonsti-able in assuming that the
dream-wish invariably originates from the unconsciousV

The last statement quoted shows, therefore, that it is strictly fair to
conclude that in the development, in his book, of the two highly
important assertions, viz. that

(1) (page 438), "I believe that the conscious wish is a dream-
inciter only if it succeeds in arousing a similar unconscious wish which
reinforces it," and

1 The italics are those of Freud. 2 The italics are mine.

T. H. Pear 287

(2) (page 439), "The wish manifested in the dream must be an
infantile one,"

Freud has not proved his points. We see, too', that he is conscious of
this omission.

If we examine some of the most striking examples of his own dreams
we find that they are 'grown-up' dreams which are actuated by pro-
fessional interests (cf the second dream examined in this article), and in
them he demonstrates no infantile factors, nor does he show that the
wish at the bottom of these dreams was invariably a repressed, un-
conscious one. It is necessary to point out this fact, even at the risk of
becoming wearisome, since later expositions of his theory by others
state in a dogmatic manner what Freud himself expresses with diffidence.
Jones^, for instance, says " The latent content is always unconscious,
i.e. it consists of mental processes unknown to the person, and of which
he cannot become aware by direct introspection, but only by means of
certain indirect modes of approach.... The latent content is of infantile
origin, later additions being merely reinforcements of earlier infantile
trends."

It may be argued that subsequent work has justified this removal of
the limitations originally proposed by Freud in the statement of his
theory, but, so far as can be gathered from current literature, this work
has been performed mainly upon psycho-neurotic patients, and Freud's
own warning, with regard to this work when used as a basis for general
assertions concerning the rdle of the infantile in the dream of the normal
person, has already been quoted *.

It is quite clear then, that we need to know more about the dreams
of normal persons before the question of the importance of the infantile
unconscious wish in a general theory of dreams can be satisfactorily
answered.

IV, The dreams, their analysis and interpretation.

The first dream recorded below was noted immediately on awaking,
and as at the time no writing material was available the incidents in it
were repeated in words several times to himself by the writer until he
knew it by heart. This dream was unusually vivid and easily remem-
bered on awaking, and the immediate repetition, several times over, of
the very few points in the dream, combined with the fact that as soon as
possible it was recorded in writing, obviate the possibility of addition

1 Op. cit. 481. ^ Papers on Psycho-analysh, 1913, 367.

3 See quotation from Freud (page 171) given on page 285 of this paper.

288 The Analysis of some Personal Dreams

to it. The second dream was recorded in writing immediately on
awaking.

The analysis was carried out in the well-known way, by tracing the
dream material to its sources in waking life through the serial associa-
tion method, when the mind was freed from all criticism or conscious
guidance of the ideas which came to consciousness. Both dreams have
been submitted to psycho-analysis by a second person trained in psycho-
logy, but no dream thoughts other than those discovered by the method
of ' free association ' applied by the writer to himself^, were found.

In the first dream I have, for obvious reasons, omitted the names,
and altered the initials of the names, of the persons who appear in it
and in the dream thoughts. It is with reluctance that I publish this
dream, but the reason which impels me to do so is that I think it
important because it was the first dream analysed by me, at a time
when I knew only the bare outlines of Freud's theory. Further, at the
time of noting it, every detail in it appeared to me to be perfectly
remembered. When it was analysed every point seemed to be perfectly
accounted for, in terms of my past experience. This is a subjective
feeling which rarely occurs to me when considering my own dreams ^ but
it was very clear at the time.

In the second dream, which is constructed around psychological
matters connected with Dr C. S. Myers, I am permitted by his kindness
to use his name without alteration.

The First Dream,

The Apparent Dream. I was in an attic with a raftered roof. On
one side of the room the roof came down nearly to the level of the floor ;
on the other, it rose to a fair height. Psychological apparatus was dotted
about the room. I was experimenting with some apparatus {the character
of which I do not remember), being assisted by Miss G., a colleague at the
University of Manchester. Miss G. said suddenly " It's one o'clock, let us
go over to lunch^," and, moving over to some pegs, she took doiun a ' blazer,'
and put it on instead of her coat. The ' blazer ' was maroon in colour,
ivith two shields on the pocket. I looked surprised at this action, whei'e-
upon she said, " Oh, it doesn't matter what one wears over there." I
awoke, laughing.

* See Freud, op. cit. 414.

' I have been studying my own dreams for the last IJ years.

' In the dream I understand this to mean " in the refectory."

T. H. Pear 289

After waking, the dream was at once considered, in the manner
mentioned on pages 287, 288, and the ' free associations ' to the various
points of the dream were noted.

Sources of the apparent dream. (The items following in italics
refer to the apparent dream.)

Attic, raftered roof. The size and shape of the room, and the slope
of the roof, are those of my bedroom in Manchester, which is an attic.
Shortly before the dream, I had been discussing a proposed extension
of my laboratory, and a member of the staff had said to me, " One way
would be to put you in the attics in the main building. Go and see if
they will suit you." While this question of extension was still unsettled.
Miss G. sent for me to discuss the possibility of my taking one of her
classes in my laboratory. This change would necessitate the enlarge-
ment of my laboratory, and would make us members of the same
department (i.e. the department of Education). We should thus work
together. I was anxious, for several reasons, to teach this new class.
Whije visiting her room, which is one of the attics under consideration,
I had examined critically this room and the neighbouring ones. They
have raftered roofs.

Coat, Miss G. A friend, S. (whose importance will appear later),
had recently said to me, in talking of Miss G., " She looked cold to-day
at lunch, and was wearing her coat in the refectory." (The fact that the
dream occurred in the summer should be mentioned here, as it accounts
for the interest taken in this otherwise commonplace remark.)

Refectory, Miss G. A day before the dream I had lunched with
Miss G. in the refectory. As I got up to leave the table, a colleague
who is in my own department stopped me and said, " Is it true that X.
is leaving ? " (X. is another colleague in my department.) I answered
" Yes," and he said, " I wonder whom we shall get next ; the men in that
post have always been nice fellows." I immediately thought of my
firiend F., who had preceded X., and is now dead, and I said no more, but
hurried away, as the memory was painful to me.

The apparent inciting cause of the dream was a trivial event from
the dream-day, in which the centre of interest was my own coat. On
the night of the dream, a friend returned unexpectedly from South
Africa, and dined with me. From my house he telephoned to some
other friends, who replied, " Come along at once, and bring Pear, too."
On my arrival at their house I was slightly embarrassed at finding that
I had forgotten to change my coat, and was wearing a very old torn
coat which I was fond of wearing in my study. The incident, however.

290 The Analysis of some Personal Ih^eams

probably made more impression upon me than I would admit to myself
at the time, for although I knew that my hosts would not resent my
unusual dress, and laughed at myself for entertaining such a thought, I
was not comfortable all through the evening, and the thought of my
coat kept recurring to me. This is one reason for its prominence
in the dream ; the other one will appear later.

. Blazer, with two shields, maroon. At this time, I knew only one
blazer with two shields on its pocket ; a black one which was habitually
worn by F. when he lived with me. He occupied the rooms which I
have now — the attic bedroom and the study in which I was when my
friend, on the dream-night, took me away in my old coat. F. was also
fond of wearing a comfortable coat in his study — the blazer I have
referred to ; and the many intimate talks we had in this room are, I
believe, symbolized in the dream by the blazer which he wore on these
occasions. This blazer, however, was not maroon, but black. Maroon
is the colour of a Manchester University blazer, which was seen first at
the University sports, to which I was taken, on that occasion, by F.

F., Miss G. A few days before the dream, S. and I had called on
Miss G., who had introduced us to a Mrs F, This lady has the same
name as F., and it should be emphasised that the name is not a
common one ; in fact I know personally only three people of this name.
On leaving, S. had remarked to me that there was a striking resemblance
between the faces of Mrs F. and F., especially about the eyes. This
resemblance, together with the identity of their names, had also struck
me before it was emphasised by S. (It should be remembered that S.,
too, is responsible for the association between the ideas of coat and
Miss G.) Mrs F. had interested us very much by talking to us about
South Africa. The only other person who had lately discussed this
subject with me was the friend who was responsible for the prominence
of my coat in my mind on the dream-night.

Attic, F. As mentioned above, there is an association between my
attic bedroom and F., who occupied it before me.

The interpretation of the dream. One or two more remarks con-
cerning some experiences in the waking state which have obvious
reference to this dream will, I believe, prepare the way for a very
probable interpretation. F. died at a time when I had no opportunity
of talking about him to others, as I was then staying with people who
did not know him. When I came back to the University, little was
said to me about the sad event, for very natural reasons, and thus there
had been no chance to share my sorrow with others. But from time to

T. H. Pear 291'

time I was astonished by the fact that occasionally I forgot momentarily
that F. was dead. Once, while immersed in reading, I found a new
theory which would have interested him, and was astonished to find
that I had begun to write a postcard to him, to call his attention to the
fact. My belief is that I had persistently repressed the memory of his
death. There was no possible outlet for my sorrow at the time when
this painful news reached me, and, later, the feeling of others that little
good would be done by talking to me about my late friend closed all
possibility of effective reaction to the sorrow. On the dream-day, how-
ever, an indirect reference to him was made in conversation, and I
hurried away in order to avoid the subject. But the words of my
colleague must be remembered — "I wonder whom we shall get next;
the men in that post have always been nice fellows."

At the time of this speculation concerning X.'s successor, X. had
begun to give me valuable help in my laboratory. F.'s help in, and
sympathy with, my work was a feature of our friendship which I always
remember with the greatest pleasure. In the dream a colleague is
helping me in my laboratory, and I believe that the meaning of the
dream is that F. returns to his post.

I believe that Miss G. represents F., for she introduced me to
Mrs F. who at once recalled him, not only to my mind, but also to S.
who, by mentioning the resemblance, emphasised it in my memory.

She signals the end of work and the beginning of social intercourse
by putting on F.'s coat, just as he used to do.

In the dream she is a colleague in my department, as he was.

On two occasions S. has made an association in my mind with
Miss G.; once with F.'s name, once with the idea of 'coat.'

The processes of condensation, distortion and symbolism may be
clearly seen here. The scene of the dream is an obvious condensation
of the bedroom successively occupied by F. and myself, the laboratory,
and the attic which I hoped would form a laboratory in the future. The
blazer is composed of two such coats, and it should be noted that even the
incorrect colour is taken from a memory for which F. was responsible.
Without laying oneself open to the charge of uncritically accepting
Freud's theory of the distortion which is brought about in order to pass
the ' censor,' it may be pointed out that, had the blazer appeared in the
dream with all its characteristics correct, it might have been recognised
as belonging to F.

The symbolism, by means of which my friend, although not appearing
in the dream, is represented by the most characteristic feature of his

292 The Analysis of some Personal Dreams

dress, is simple and clear. The dramatization in the dream speaks for
itself. The superficial associations which arose from the chance re-
semblance of two persons, coupled with the coincidence of the identity
of their names, are just the material which we should expect, if Freud's
theory be correct, to form the core of the dream. Lastly, the sympto-
matic action which happened in waking life^ forms another powerful
piece of evidence that I had repressed the memory of my fi-iend's
death.

In this dream two wishes are fulfilled which were conscious and
fully recognised by me in the preceding waking state. Miss G. and I
become colleagues in the same department and the attic becomes my
laboratory. Behind both these wishes there was a relatively great
driving force ; the first wish-fulfilment represents a gratifying increase
in the number of students in my department and the second an increase
in the laboratory accommodation. These wishes are derived ifrom
professional and personal interests which are quite clear to me. But it
is important to notice the way in which the deeper-lying ' wish,' which
in waking life was never overt, but existed probably as a restless,
untiring conative tendency, underlies the whole dream.

In view of what Freud has maintained with regard to the action of
the 'censor' in waking life, it is important to note, too, the fact that
I awoke laughing. The real subject of the dream is one which, had I
realised it, would have been connected with an emotion very different
from that which I felt on awaking.

Second Dream.

The Apparent Dream. I was in St Anns Square, Manchester, early
in the evening, in the summer. The light was curious; impossible to
compare with any light effects seen in waking hours. The whole square
seemed to be one large arena (like the arenas used for bullfights), and
people were crossing and re-crossing it. At one end of the square {the
end opening into Market Street), in the right-hand corner, was a large
cinematograph screen, showing moving pictures, and the impression in
the dream {which seemed quite natural then), was that the square itself
was one vast 'picture-palace.' I was then in position 1 on Fig. 1,
uncomfortably close to the screen; i.e. the pictures were not easily seen,
and were distorted.

^ Cf. Freud, Zur Psychopathologie des Alltagslebens, Dritte Auflage, Berlin, 1910;
Jones, " The Psychopathology of Everyday Life," Amer. J. of Psychol. 1911.

CAFE
3

T. H. Pear 293

Suddenly I found myself in position 2 in the square. The scene
had narrowed down to the size of an ordinary room, about 12 feet
square, though I could see no walls. The light
was brighter, but not very bright, and I recognised
it as coming from, electric incandescent lamps.
I luas still in the square, yet people in evening
dress were passing and re-passing me, through
a partition like a screen. It was, except for the
feeling of not being ' walled in,' exactly like being
at a University soiree'^, for Professor and Mrs S.,
in evening dress, passed through the partition
and greeted me.

PARTITION ■=

CINEMATO- =
GRAPH SCREEN

Fig. 1. / found myself then at position 3. Here

the light was dimmer, and I ivas sitting at
a long form, amongst several other forms. People were eating and
drinking, and the place seemed like a South German cafe. The
' Gemutlichkeit ' was very apparent to me. (This feeling-tone, and
the eating and drinking, were the only ' cafe-signs,' yet they were quite
adequate to complete the perception of the place as a cafe"^.) At once
Dr Myers walked into the cafe, sat down by me with a casual greeting,

Cambridge

Bidwell

-i- — > •— London

Fig. 2. Fig. 3.

and took out several sheets of paper. (The impression was that we had
both been working in the same laboratory, and had seen each other quite
recently.) He began immediately to explain to me that he was beginning
a research on ' physical and metaphysical logarithms! {The work was
quite fidl of meaning and comprehensible to me at the time, and the
problem seemed quite familiar.) His first paper contained complex
algebraical, problems in which two problems were worked out in very
neat parallel lines^, in Dr Myers's handwriting, side by side, like the

^ The consciousness of the ' meaning ' of the scene was quite clear, although the
'scenery' would not have suggested a soiree to anyone in the waking state.

2 Cf. preceding footnote.

^ These lines were longest at the top and gradually decreased in length. See Fig. 2.
I sketched them immediately on awaking, but do not think I read them in the dream.

294 The Analysis of some Personal Dreams

creditor-debtor columns in balance-sheets {see Fig. 2). / cannot remem-
ber if I understood them. The reasoning did not seem difficidt. Then
he began to draw, on another sheet of paper, a map^ to illustrate his
remarks, which were, " You {meaning the dreamei'), go down to London
through {or from) Cambridge, and you get short-circuited at Bidwell, on
account of the suffrage question." I quite understood this at the time.
While he said this, a man bent over both of us. He had the general
appearance of a doctor {he wore a morning coat and dark trousers), but
was unshaven, and this fact was very unpleasant to me. He kept on
interrupting Dr Myers and laughing at both of us. Dr Myers was quite
friendly with him, but I wa^ annoyed and irritated at the interruption.
(Awoke here.)

Sources of the Apparent Dream^

St Anns Square, Cinematograph. Before going from Manchester to
Cambridge, where I had stayed with Dr M. on July 13th, 1912, ten days
before the dream, two business visits had to be paid on the afternoon of
the 12th, and the limited time available for them had caused some
excitement and interest in the ev^ts. The first visit was to my tailor,
whom I wanted to remind to send me a suit of clothes to take with me
to Cambridge next day. The second visit was to see a sound-proof
partition in a warehouse. This visit interested me greatly, as, if its
sound-resisting qualities proved satisfactory, this type of partition would
be erected in my laboratory. On my way between the warehouse and
a return visit to the tailor's I met two men carrying an advertisement
which announced that a 'picture-palace' was offering free refreshments
to its patrons. I had recently visited several picture-palaces, and had
discussed them with my father.

The cinematograph screen in the dream occupies the same position
in the square that my tailor's shop does in reality. (The advertisement
and the tailor's shop were seen a few minutes after each other.) The
actual screen and the unpleasant proximity of it are recollections from
an experience on June 29th, when, in a cinematograph theatre which
my father and I visited, there were no seats available, except some
directly under the screen. The increased flicker and the unusual
angular appearance of the figures were irritating to us, especially as

* The ' map ' was really a rough diagram which I have drawn as I saw it in the dream
(see Fig. 3). The names did not appear on the ' map,' bot I understood th£^t they referred
to the places marked on it.

T. H. Peak 295

the pictures were interesting. I felt some responsibility for the in-
convenience to my father, as I had suggested this particular theatre.
There was some interest to me in the fact of the increased flicker,
and its connexion with the unusually great visual angle subtended by
the pictures, also in the one-sided appearance of the flat human figures.

Why does the cinematograph screen appear in St Ann's Square ?
In St Peters Square, Manchester (the only other square in the town
which, so far as I know, is named after a saint), there is actually a
lanteim screen, upon which changing advertisements are projected at
night. I have often waited here for the tramcar, and have found the
pictures a welcome means of passing the time. We had waited in this
way on coming from the cinematograph theatre described above. (See
figures 6 and 7, page 301.)

The connexion of cinematograph — tailor — partition will now be clear.
In position 2 on the map of the ' dream-square ' the partition actually
appears.

As it happens, the only other member of the staff who is erecting
partitions of the same kind as my own, and in the same corridor as
mine, is Prof. S. Also, the carpenter who was awaiting orders to
proceed with my partitions had been entrusted with the task of making
a lantern screen for me, to be fixed in the partitioned corHdor.

Partition — Incandescent electric lamps. I had been compelled to
postpone giving orders to the electrician about the lighting of the par-
titioned corridor, owing to the rush on July 13th, although I had wished
to do this before going to Cambridge.

Incandescent electric lamps — Soiree — Prof. S. — Bidwell. The last
time that I had worked by electric light (the dream took place in the
summer), was at a medical soiree, a few weeks before the dream. I had
had some trouble with the electric bulb above my apparatus. This
apparatus had been arranged in such a way as to leave room for an
exhibit by Prof. S. The failure of the light, and its insufficiency when
attended to, were annoying to me, because we were carrying out Bid-
well's colour experiments, which need bright illumination. These
demonstrations had excited much interest and questioning.

Lighted-up partition — Refreshments. The association given above
(page 294) partly accounts for this, but the laboratory used at the
soiree (see above) opened into the refreshment room, into which we had
gone when our experiments failed ^

1 In St Ann's Square there is actually a cafd, which I frequently visit, in position 3 in
Fig.l.

296 The Analysis of some Personal Di'eanis

Forms — Bidwell — Dr Myers — Cafe. Before leaving Cambridge, the
last two subjects I had discussed with Dr Myers (on the station at
Cambridge) were the questions of sound-proof partitions and the
writings of Bidwell. At Cambridge, too, another psychologist had
spoken to Dr M. of the habit of the psychologists at a Germun uni-
versity, at which we had both studied, of discussing, and working out,
the results of their experiments in the cafe opposite the laboratory. Not
long before this dream (I believe, the day before), I had mentioned the
same fact to my father. At that time I was working in a large room,
and had arranged my books, including those dealing with BidwelVs
work, on a table which was surrounded by long forms. This room
would, for the next few weeks, represent my work. (Being a habitual
visualizer, I frequently represent to myself a whole side of my activities
by a visual image of one important thing connected with it.)

The connexion between Bidwell, refreshments and cafe therefore
seems obvious. There is, however, still another reason for their close
connexion, which will appear at the end of this explanation.

Dr Myers — Logarithms — Lines. I had remarked to Dr M. in our
conversation on the station that there seemed to be a probability that
a colour effect which Bidwell could not understand (reported in his
paper in the Proceedings of the Royal Society), was simply the violet in
' Fechner's colours.' Professor Alexander has said more than once to me
that there seems to be a more intimate connexion between Fechner's
Law and the general nature of logarithms than has hitherto been supposed.
He has also lent me a typed sheet of manuscript which deals with the
question, and in which, I remember, there is mention of the 'physical-
psychical relation,' and of the metaphysical concepts necessary in his
treatment of the subject. I remember saying to him, "I think I under-
stand it." The lines in Dr M.'s MSS. seemed in the dream to be parallel,
but the actual figures which formed them were not clear in the dream.
The most striking feature on the paper was the fact that there were
parallel lines.

The last time that Dr M. was in the Manchester laboratory, when
he was discussing the ' partition ' question with me, we discussed also
the ' Bidwell ' work in connexion with the familiar phenomena of
Benham's disc, in which black lines on a white ground (Fig. 4)
appear coloured when the disc is slowly revolved. As I revolved the
disc, he drew for me, on a sheet of paper, the lines of the top, writing
by the side of them the colours that he saw (Fig. 5). I was able to find
the actual paper, of which Fig. 5 is a reproduction. These lines, longest

T. H. Pear

297

at the top and gradually decreasing in length, are just like the lines of
the calculation seen in the dream. This paper was found filed with
those relating to the ' Bidwell ' work which lay on the table amongst
the long forms.

Fig. 4.

Fig. 5.

Cambridge — Bidwell — London — Suffrage. A few days before the
dream I had looked at the road map of the route between Wisbech, in
Cambridgeshire (near which town I was then staying), and London, via
Cambridge. I had noticed Great Shelford, where I had stayed with
Dr M., but the study of the map had taught me nothing new, and I
had noticed the relative positions of Wisbech, Cambridge and Shelford
simply because they had interested me,

J. of Psych, VI 20

298 The Analysis of some Personal Dreams

The last time that I had visited Dr M. at Shelford was at a time
when I had intended to go to London from Wisbech, and he had
invited me to take Cambridge and Shelford on the way. Dr M. had
also said in my hearing, when he was at Prof Alexander's house in
Manchester, to a lady who had asked him about his views on the
question of women's suffrage, " You should come down to Cambridge."

Until several hours after the dream the reason for the substitution
of Bidwell for Shelford did not occur to me. I may have noticed
before that Shelford Bidwell is the full name of the investigator who
had occupied my thoughts, but ' Bidwell ' seemed, in the dream, to be
quite the natural name of the village. The position of Bidwell on the
' dream-map ' was undoubtedly that of Shelford, for it was understood to
be three or four miles south of Cambridge, or rather, in the terms of the
dream, so many miles nearer London, on the way from Cambridge.

Short-circuited. While, during my stay at Cambridge before the
dream, I was asking a question about Benham's disc, an American
colour investigator came into the room. He apologized for coming in
late, and explained to us, "I've been side-tracked," meaning that he
had lost his way. Americanisms in psychology being, owing to James'
and Titchener's influence, especially interesting to me, I often use them
in my own thinking, and the words ' side-track ' and ' short-circuit ' are
frequently used in the same sentence when dealing with the psychology
of the thought processes, e.g. thoughts are 'side-tracked' or 'short-
circuited.' The American above had struck me at the time as^being
'very American,' and I was amused at his use of the word 'side-
tracked.'

The key to the whole dream, however, is given by

The man who interrupted. He is not actually a doctor, but is
intimately connected with medical work, and dresses just as he is
dressed in the dream. He is very dark and clean-shaven (but I have
never seen him unshaven, though his chin is very dark). The dark
chin has become, in the dream, an unshaven chin. The smile he wore
in the dream is the one which is, in actual fact, a distinctive feature
of him.

The man who interrupted — the German cafe. See the interpreta-
tion of the dream.

The interpretation of the dream seems to me to be conditioned by
the following facts. Before going to sleep, and during the day before
the dream, I had been planning my work for the Long Vacation. This
work, as I had been thinking about it, was represented visually by an

T. H. Pear 299

image of the work table spread with books and files. (I often use such
an image as a ' scheme ' in thinking ; cf. the map in the dream.) My
immediate interests were in the colour work (i.e. the ' Bidwell ' work),
on which I wanted to begin at once (partly because brilliant sunshine
was available at the time), and for which I had prepared some ap-
paratus. But on the day before the dream ^ I had made a decision to
leave this work alone for a time, and to begin to attack a problem con-
cerned with Memory, The reason for this was that I had recently been
asked for advice on this point by the ' medical ' man in the dream. In
fulfilment of a promise I had made him (made after the necessary work
had been planned with him over a German dinner in the German
restaurant of the Midland Hotel, Manchester, which occupies the same
position in St Peter's Square (the square which actually contains the
lantern screen) as that of my caf^ in the dream-square (compare Figs. 7
and 1)), I had decided to begin the ' Memory ' work before the ' Bidwell '
work. The data for this memory work I had obtained from this man
who, in the dream, ' kept on interrupting ' Dr M. These data actually
lie in the memory file on one side of my work table, Dr M.'s data in the
' Bidwell ' file on the other.

The meaning of the dream seems to be clear. On the ' dream-day '
I had actually decided (finally, as I supposed) to ' shelve ' the ' Bidwell '
work, although my mind was full of it, and to attack the ' Memory ' work
at once, because of ray promise. The dream throws valuable light on
the striving of impulses which may still go on, even after an apparently
final decision has been arrived at in waking consciousness. In the
waking state the conflict was brief, and apparently decisive. The
dream re-opens it in a characteristically vivid manner.

It is in a dream of such richness and complexity as this that one
may fairly seek for confirmation or negation of the existence of Freud's
alleged dream mechanisms — those processes which combine to form
what he calls the dream-work — the distortion of the latent thoughts
into the apparent dream. Let us, therefore, carefully examine the
material of the above dream.

In the first place, the dramatization in this case is well-nigh perfect.
The dream turns this mental conflict, which in a waking state would
have been one of mere thought (accompanied no doubt by some imagery),
into a kind of problem play, in which the two opposed influences be
come human beings. In the dream, the attractive ' Bidwell ' work is

^ The ' dream-day.'

20—2

300 The Analysis of some Per social Dreams

represented — for the many reasons given above — by Dr M. ; the less
attractive but urgent ' memory ' work by the man who ' keeps on in-
terrupting * him. My thoughts of the ' Bidwell ' work on the dream-day
had been continually interrupted by the thoughts of the ' memoiy '
work, until I had put an end to this state of things by deciding to
postpone the former work. In the dream I am annoyed and irritated
by the representative of the ' memory ' work, who persists in interrupting
the representative of the ' Bidwell ' work.

It should be noted that in the dream I am not irritated at the real
cause for annoyance, viz. the work which I have promised to do, but at
the man to whom I have given my promise — a man with whom I have
always been on friendly terms. Moreover, the dream seizes upon one
harmless feature of the man — his dark chin — to transform it into a
feature which is very unpleasant to me — an unshaven chin. Freud's
assertion that the emotional tone which is attached to a thought in the
latent content appears in the dream attached to another, related,
element which is not under the ban of the censure, must be considered
in connexion with this feature of the dream. It is quite true that
professional and scientific interests would oppose, in waking life, a
strong resistance to the temptation, which probably arose here, to con-
sider the work itself as irritating. As a matter of fact, the work was
very interesting to me, but the fact that my mind was full of the newer
problem had been sufficient to displace it temporarily from the focus of
my interest. Here the emotional tone seems to be displaced from the
work to the man, and in particular to one feature of him.

The condensation employed in fashioning the 'stage' and the
'scenery' of this dream is clearly visible. The most striking case is
that in which the two squares, the four restaurants ^ the two lantern
screens and the two laboratories fuse to form the scene of the events.
(A comparison of Figs. 6 and 7^ with Fig. 1 shows this clearly.) The
manuscript paper is a fusion of two papers, and the speech is a clear
condensation.

The kind of superficial association involved in the play on words in
the names Shelford and Bidwell, utilising a coincidence, is again a very
common factor in the dream- work of Freud's theory.

It seems undoubtedly true, then, that several of the processes which,

1 The German restaurant at the Midland Hotel, the cafe in Germany, the cafe in
St Ann's Square and the refreshment room at the soiree.

2 Actual plans of the relevant details, with their relative positions, in St Ann's Square
and St Peter's Square respectively.

T. H. Pear

301

according to Freud, are characteristic of the dream-work, are illustrated
in this dream.

What, then, can we regard as the meaning of the above dream ?
We may, I think, fairly describe it as the dramatic representation of a
mental conflict in which the opposed conative tendencies at work appear
in disguised forms. We must note, however, that I awoke before I
was able to see if the interrupter was completely successful. And here
we may ask, " What of the wish theory ? " It may be that my wish to
pursue the interesting work was actually being fulfilled in the dream
when the waking consciousness, gradually regaining its power, intro-
duced the counter-thought of my promise, and that this, depicting itself
as dramatically as the first thought, appeared as the ' man who inter-

cafe'

tailor's

SHOP

MIDLAND
HOTEL

Restaurant)

<^

LANTERN
SCREEN

ST. ANN'S SQUARE
Fig. 6.

ST PETER S SQUARE
Fig. 7.

rupted.' In connexion with this it must be remembered that it was
only at the end of the dream, just before I awoke, that he appeared.
I was able to find no cause for my awaking, such as a noise or a sudden
change of illumination. It should be noted, too, that the partition,
lighted up with incandescent lamps, represents the state of affairs as I
should wish it to be, and that at the soiree in the dream the lamps
are efficient, while in the event which caused this dream-episode they
were unsatisfactory.

The phrase, "You get short-circuited at Bidwell" is interesting.
Firstly, it strings together the words from several experiences to form
a new sentence, and is obviously a condensation ; secondly, it expresses in

302 The Analysis of some Personal D reams

a figurative way the new direction given to my thoughts by the ' Bidwell '
work. It seems possible that this expression contains the incipient
form of what in a more highly developed state might have been a
combination of two rather ' cheap ' forms of wit — the play on words in
using ' Bidwell ' for Shelford, and the use of an actually heard phrase in
a new way^

Another fact in connexion with this might be mentioned here,
viz. that Freud's theory of the short-circuiting or side-tracking of
emotional interests was known to me, through American writers, at the
time when I dreamt this dream. I do not w^ish to assert that I believe
strongly that the sentence expresses more than the ' short-circuiting ' of
my thoughts by the ' Bidwell ' work, but the completion of the sentence
should be noted — " on account of the suffrage question." It should be
remembered that one of the chief arguments used by the opponents of
Women's Suffrage is that they believe that certain highly important
interests may be " short-circuited on account of the suffrage " should
women take too intense an interest in public affairs. The existence of
this argument was well known to me at this period, and the psycho-
logical aspect of this controversy was at that time, for several reasons,
frequently in my thoughts. (But this possible interpretation occurred
to me at a time much later than that at which I analysed the dream, so
that I wish it to be considered quite separately from the facts given
above. I have no proof that such a thought entered into the com-
position of the dream, but from a consideration of my special interests
at the time of the dream, I am inclined to think it probable.)

V. Conclusions.

The two dreams which have been analysed in this article illustrate
clearly the processes which Freud has termed the 'dream-work.' In
them we find instances of dramatization, symbolism, condensation,
displacement, and the superficial association which, in both dreams,
makes use of a chance identity of the names occurring in two past
experiences.

The distortion in the dream seems to bear traces of evident pur-
pose, and its character, when taken in conjunction with the events in
waking life which are represented in the dream, supports the concept
of a ' censorship ' which is evaded by it.

1 Cf. Freud, Der Witz utid seiner Beziehung zum Unhewussten, Wien, 1905.

T. H. Pear 303

In the first dream, although it presents the fulfilment of two
conscious wishes, the important underlying wish which was fulfilled
was unconscious. Though in one sense this wish was of an infantile
character, it can scarcely be said to have emanated (at least in its
present form) from childhood.

In the second dream, the wishes which gave rise to it are clearly
seen on analysis, but there is no trace, even on careful psycho-analysis,
of an unconscious wish. The tendency suppressed in the waking life
was a conscious one.

On the whole, these dreams appear to support many of the main
assertions made by Freud. They do not, however, afford evidence for
his conjecture that the infantile and unconscious wish is a necessary
cause of the dream. The consideration of many other dreams of my
own, and of other ordinarily healthy persons, leads to the opinion that
this extension is a generalisation supported by insufiicient evidence.
Freud cannot be said to have shown that his theory is valid for all
dreams, for the dreams which are recorded when the dreamer is
awakened in the midst of deep sleep, as well as those which occur
immediately after falling asleep, have not yet been subjected to de-
tailed study combined with an examination of their latent content.
Neither has he proved the general validity of his theory for the dreams
which remain in consciousness on awaking, although many such dreams
are explicable by means of his hypothesis.

It is possible that future work may result in the further analysis of
the ' dream-work ' and the ' censorship,' and that they may be shown to
depend upon factors with which we are already familiar. But there
seems little doubt that such processes exist and play an important part
in mental life, and that Freud's striking demonstration of them is a
valuable contribution to psychology.

(Manvscript received 6 December, 1913.)

THE CONDITIONS OF BELIEF IN IMMATURE

MINDS^

By CARVETH READ.

§ 1. Introditction.

§ 2. The first ground of belief is perception.

§ 3. Classification of secondary grounds and causes.

§ 4. Immature minds perceive a« we do.

§ 5. The imaginative beliefs of savages are moulded by passion and

custom, and are allied to pla/y-belief.

§ 6. Effects of intense imagination, absence of criteria, and mental

incoordination.

§ 7. The ratiocination of immature minds.

§ 8, The weakness and gradual decadence of imaginative beliefs.

§ 9. The utility of imaginative beliefs.

§ 10. Black m^agic and red religion.
§ 11. Imaginative beliefs and scientific ideas.

§ 1. Introduction.

The expression 'immature minds' is here used to include children
and backward adults. Backward adults include the great majority of
all civilised nations and a still greater proportion of barbarians and of
savages. But the present essay is chiefly concerned with savages ;
because amongst them the conditions of belief to be examined operate
most intensely and with least qualification or restriction. And the
word ' savage ' is used, not offensively, but for the sake of brevity, to
denote the more backward peoples, and as less misleading than ' primi-
tives ' ; since no people is known to us that must not have had a
history as long as our own.

^ An abstract of this paper was read at the meeting of the British Association for the
Advancement of Science, Birmingham, Sept. 1913.

Carveth Read 305

The general conditions of belief have so often been discussed that I
must ask to be excused for restating them in the way most convenient
for the following discussion^ It seems to me that they are the same for
all minds, except that (1) their relative influence varies at different
stages of development, and that (2) it is only for some minds amongst
civilised peoples that there exists a Logic or Methodology : definite
categories of judgment (such as ' cause ' and ' quantity ') and canons of
evidence, without whose sanction they do not fully believe anything —
at least, within a certain sphere of investigation or special study;
though, outside of it, they may be credulous enough. Of course, our
Logic may be faulty, or we may make mistakes in applying it, and so
adopt erroneous beliefs in spite of our care.

§ 2. The First Ground of Belief is Perception.

The ground of all belief is perception, directly or indirectly ; and in
perception certain sensations have a certain order of prepotency ; — pain,
tacto-kinesthesis, vision or smell, etc. This is common to men and
animals ; and is such a matter of course that we are apt to overlook its
significance and necessity. Belief has sometimes been discussed as if it
were chiefly concerned with relations of ideas; and systems of philo-
sophy have sought justification in the coherence of their ideas, with
little or no regard (not to say with contempt) for the coherence of ideas
with perceptions. But nearly the whole of every man's life (savage or
philosopher) passes in an attitude of unquestioning belief in the
evidence of his senses.

Methodologically we know that perception is fallible; but in the
long run it overrules everything else ; and experimental methods
consist in taking precautions against the errors of perception, and in
bringing every hypothesis to the test of perception.

§ 3. Classification of Secondary Grounds and Causes.

Further grounds and causes of belief are divisible into (1) the
evidentiary, which (though often misleading) may generally be justified
by reflection as raising some degree of probability ; and (2) non-
evidentiary, which (though very influential) cannot be justified by
reflection as having any logical value.

1 Perhaps indebtedness should be acknowledged especially to my old friend Prof. James
Sully. See The Human Mind, c. xii. and Sensation and Intuition, Ess. iv.

306 The Conditions of Belief in Immature Minds

(1) Evidentiary grounds of belief are (a) memory, shared by the
higher animals and indispensable to ourselves; (6) testimony, which
must be trusted if language is not to be useless : both of these grounds
are supposed to rest upon previous perception. And (c) inference, also
shared by some of the higher animals, and necessary to all original
adjustment to the future or to unperceived circumstances, but highly
fallible, and constituting the chief problem for the exercise of Logic
when that science arises : especially to distinguish, amongst inferences,
valid illations from merely verbal substitutions and from incoherent
imaginative analogies. Incoherent imaginative analogies become, with
the growth of reflection, relegated to the region of play-beliefs and
poetry ; but where, in the immature mind, necessary beliefs and play-
beliefs are very imperfectly differentiated, such analogies are effective
in practical affairs.

Testimony gathers force with the numbers and consideration of
those who support it, and especially with their unanimity.

As the constructive instinct deals with beliefs and play-beliefs, they
are piled up into systems of Science, Theology, Philosophy, Astrology
etc. : in which systems, each belief strengthens, and is strengthened by,
the rest. Even without systematization, the mere structural similarity
of beliefs formed upon the same implicit principles of causation, or of
magic, or of animism, throws them into apperceptive masses ; and such
systems or masses readily assimilate and confirm new inferences having
the same character, and offer resistance to all inferences having a
different structure.

(2) The non-evidentiary causes of belief are chiefly the following :
(a) The agreeableness or disagreeableness of any judgment draws

attention to, or diverts it from, such a judgment and the evidence for
it : except that some disagreeable feelings, especially fear, by a sort of
fascination of attention, are favourable to belief in the imagined danger.

(6) Every desire fixes attention upon beliefs favourable to it and
upon any evidence for them, and diverts attention from conflicting
beliefs and considerations. Thus every desire readily forms about itself
a relatively isolated mass of beliefs, which resists comparison and,
therefore (as Ribot says), does not recognise the principle of contradic-
tion. Incompatible desires may be cherished without our becoming
fiilly aware of their incompatibility ; or if the fact obtrudes itself upon
us, we repudiate it, and turn away.

The more immature a mind, and the less knowledge it has, the less
inhibition of desire is exerted by foresight of consequences that ought

Carveth Read 307

to awaken conflicting desires ; the less compassion one has, the less is
desire inhibited by its consequences to others : therefore the less check
there is upon belief.

(c) Voluntary action in connexion with any belief, whether of a
rational kind or in connexion with rites and ceremonies, favours that
belief: (i) by establishing the idea-circuit of means and end, the end
suggesting the means to it, and the thought of means running forward
to the end ; a circuit that resists interruption : (ii) by the general effect
of habit and prejudice; for every habit of action or of thought has
inertia and, moreover, it is agreeable, and to break it disagreeable ; so
that, again, a relatively isolated system is formed, which resists com-
parison and criticism.

On the influence of desire and rites etc. depends the " will to
believe." We cannot believe anything by directly willing it; but we
can will what to attend to, or what to do.

(d) Finally, belief is determined by certain social influences other
than testimony and tradition : especially by sympathy and antipathy
between families, parties, tribes ; and by imitativeness and suggestibility
(qualified fortunately by contra-suggestibility) ; so that beliefs become
fashionable, endemic, coercive, impassioned and intolerant.

§ 4. Immature Minds perceive as We do.

All these grounds and causes of belief, evidentiary and non-eviden-
tiary, are common to both civilised and immature minds ; but their
proportional influence is very different at different stages of develop-
ment ; and in immature minds the power of the non-evidentiary causes
is excessive.

Perception, in normal circumstances, is accepted by all as a matter
of course : it controls the necessary activities of practical life in
hunting and in industry, in making weapons, hoeing the ground,
building houses, etc. ; however these processes may be modified, or
interrupted, by the intrusion of beliefs derived from other sources. If
the savage sings a spell to his prey, or his weapon, or tool, or buries a
slave under his house, he may thereby increase his own confidence in
the work ; but, otherwise, if it be no better, neither need it be the worse
for such hocus-pocus. The properties of matter exact practical obser-
vance, without which nothing can be done. Even magical practices
presuppose a sane perception of the central facts : as who is acting, for
what purpose, when and where, with what and with whom. Upon this

308 The Conditions of Belief in Immature Minds

basis there may be an astonishing accretion of imaginative belief; but
we shall see that there are limits to the effectiveness of such beliefs.

M. L6vy-Bruhl, indeed, in his very interesting book, Les Fonctions
Mentales dans les Societes Inferieures, maintains that, under the
influence of social ideas {representations collectives), the primitive mind
actually perceives things differently from what we do. Whilst we
succeed in attaining an objective presentation, eliminating subjective
associations, for primitives proprietes mystiques, forces occultes are
integral qualities of the object. He grants that in certain cases of
immediate practical interest, we find them very attentive and able to
discriminate slight impressions, and to recognise the external signs of
an object on which their subsistence or even their life depends (p. 40) ;
but in the tres grande majorite des cas, their perceptions are over-
weighted by subjective elements. This doctrine reverses (I venture to
think) the real relations between perception and other causes of belief
and their proportionate influence in savage life. It is not only where
subsistence or life is at stake that backward people see things as they
are : in merely experimental tests, Dr Rivers found amongst both
Papuans and Todas, that, as to suggestibility in perception, they
showed a high degree of independence of judgments

§ 5. The Imaginative Beliefs of Savages are moulded by
Passion and Custom and are allied to Play-Belief,

The peculiarity of savage beliefs, then, is not due to corrupt and
clouded perception, but to the riot of imagination, unrestrained by
criticism and reinforced by the popular concensus. The savage's
imagination is excited by the pressing needs of his life in hunting, love,
war, agriculture, and therefore by love, hate and grief, by fear; suspicion
and anxiety. Imaginations spring up in his mind by analogy with
experience ; but often by remote or absurd analogies ; and there is no
logic at hand to distinguish the wildest imaginative analogies from
trustworthy conclusions. The same pressing needs and the same
emotional storms affect a whole tribe and stimulate everyone's imagina-
tion ; and, tracing its origin (no doubt) from ancient times, and slowly
clearing and solidifying, there grows up a mass of public imaginative
beliefs, which are inculcated into every individual by tradition, sugges-
tion, imitation, sympathy. Such beliefs are embodied in rites, cere-
monies, formulae ; and are, in fact, customs. They have, therefore, the
» This Journal, Oct. 1905, 393.

Carveth Read 309

strength of custom in the habits of individuals and families and in
public respect ; and the weakness of custom, inasmuch as the obser-
vances may continue whilst the beliefs are forgotten, or may decay and
disintegrate by social fatigue and neglect. In their flourishing period
they extensively modify the behaviour of tribesmen in all practical
affairs, sometimes helpfully or harmlessly, sometimes injuriously and
destructively. In general, imaginations are prevented from modifying
a tribe's conduct beyond certain limits by biological necessities; but
exceptionally they result in tribal insanity, tending toward, if not
accomplishing, the tribe's destruction, as in extreme cases of the
practice of human sacrifice, or of the ordeal by poison.

Indeed, so violent and tyrannous is the effect of superstitious beliefs
in many cases, that it may be difficult to understand how they are
almost entirely born of the imagination : the qualification ' almost ' will
be discussed presently with reference to Magic (§ 8). In a civilised
country there are always current some beliefs as imaginative as any
to be found in the middle of Africa; but whenever the imaginative
character of a belief has been recognised, we class it as ' make-believe '
or (better) ' play-belief,' and it passes into the region of fine art, fiction,
sports and pastimes. If such things have any place in our life, we go
to them of personal choice in the intervals of business ; are fully aware,
for the most part, that the matter which absorbs our thoughts for the
time is not really important — whether the hero or villain will prosper,
whether the fox will be killed or get away; and perhaps excuse our
condescension on the ground that the tone of fiction affects public
morals, that fox-hunting maintains the breed of horses, and that (at
any rate) the bow must not be always bent. Under the influence of the
fine arts and literature our emotional states may be intense ; but they
are dissociated from action, exist for their own sake, have a special tone
and require only an imaginary satisfaction. With a backward people
a much smaller portion of their imaginative possessions and pursuits
has been differentiated as play, and much that seems to us absurd
seems to them necessary ; the actions and observances that express their
beliefs are not performed as a matter of personal choice, but of public
custom : the ends to be obtained (they think) are the same as those of
what we call business.

To understand how, in spite of these contrasts, the magical or
religious beliefs of savages and the play-beliefs of civilised men, having
a common ground in imagination, are closely allied, we must call to
mind the many degrees of intensity of play-belief in ourselves, varying

310 The Conditions of Belief in Tmniatare Minds

from the momentary entertainment of playing with a child, through
various grades of fiction or ceremony, down to a deeply serious frame of
mind, a profound movement of dread or compassion that may long
outlast our play. A child's absorption in such beliefs is far more
intense than ours ; but his circumstances prevent him from attaining to
the convictions of a savage. The child of civilised people has little or
no support in tradition (except sometimes from nursemaids) ; he is not
driven by the desires and anxieties of subsistence ; and he is frequently
interrupted by his seniors. The savage has an overwhelming tradition
and authority, pressing anxieties, and no seniors. Until the civilised
sceptic reaches his shores, there is nothing but tardy experience or
social fatigue to check his vagaries. His imagination vies with the
sense of reality, often overpowers it ; yet (we shall see) his imaginative
beliefs show many signs of their insecure foundations.

§ 6. Effects of Intense Imagination, Absence of Criteria,
AND Mental Incoordination.

It is not only the influence of society and tradition that renders
imaginative beliefs coercive to the savage ; in the immature mind of
the individual there are certain characteristics favourable to their
prevalence.

(a) The process of imagination itself, the picture-thinking of
savages, seems to be more vivid, sensuous, stable, coercive, more like
perception, than our own normally is. " The Australians," say Spencer
and Gillen, " have the most wonderful imagination \" They often die
of it. So do Hindoo peasants, Maories, Negroes and others, if they
know that they have been cursed or have broken a taboo. Hence there
is a tendency to believe in imaginations as perceptions are believed in ;
and to believe in the efficiency of rites, because the mere performing of
them with a purpose makes their purpose seem to be accomplished.
When a man of intense and excited imagination makes an image of an
enemy, and stabs it, that his enemy may suffer, his action gratifies the
impulse to stab with its associated ideas, as if he actually wounded the
enemy himself; and so revenge seems to be a present fact.

The same intensity of imagination is found in civilised children, and
is greater than in ordinary adults^. Savages, too, seem to dream
more vividly and convincingly than is usual amongst ourselves; and

1 Native Tribes of Central Australia, p. 462.

2 cf. R. R. Rusk, This Journal, Dec. 1910, v.

Cakveth Read 311

they are said to be more liable to hallucinations. Their dreams are
attributed by travellers sometimes to fasting, voluntary or involuntary,
sometimes to extreme repletion when there is an opportunity for it.
Physiological conditions of the immature brain may explain the intensity
of imagination, the vividness of dreams and the hallucinations.

(6) But more important than any intensity of picture-thinking to
the growth and persistence of imaginative beliefs, is the absence of a
standard by which they might be discredited. One reason why we
believe our memories, and not our imaginations, is that, whilst in both
cases the images (or elements of images) are derived from experience,
in memory the relations of images in place, time, and context are also
derived directly from experience; whereas, in imagination, images or
their elements are reconstructed by analogies, often very vague analo-
gies, of experience, or by condensations the most capricious. Hence, to
make imaginations credible for us, the relations of experience must be
faithfully imitated, as (e.g.) in Robinson Crusoe. But outside the
practical, repetitive, necessary course of life, observation of fact by
immature minds is not exact and coherent; and, therefore, their
memories are not coherent, especially as to relations of time ; so that
imaginations suffer little by comparison with such memories. There is
not enough orderly memory or general knowledge to discredit even
absurd imaginations; for so far as observation and memory are disorderly,
generalisation, conscious or unconscious, is impossible. Hence not only
traditionary myths may be monstrous and arbitrary, but occasional
tales of private invention, amongst both children and savages, usually
exhibit disconnected transitions and impossible happenings. Yet they
satisfy the immature mind.

But even so far as standards of judgment exist amongst savages,
derived from repeated experience in their practical life, there are
certain conditions of the immature mind that hinder the comparison of
ideas and, therefore, the criticism of beliefs.

(c) About every imperative need, such as success in hunting, with
its correlative desires and anxieties, rites and ceremonies gi'ow up to
gratify imaginatively the desires, and relieve the anxieties; and the
ideas of these observances form relatively isolated systems. To us these
ideas generally seem absurd and irrelevant, when compared with the
facts of the savage's own experience. We see a hunter, for example,
endeavour to gain his ends by two distinct series of actions. In one he
fasts, enchants his weapons, casts spells upon his expected prey ; in the
other he carefully prepares his weapons, patiently tracks his prey, warily

312 The Conditions of Belief in Immature Minds

approaches, and slays it. The latter series we approve and appreciate
as causation ; the former we ridicule as hocus-pocus, contributing
objectively nothing to the event; and we pity the 'heathen in his
blindness.' And, indeed, he may be said to be mind-blind ; for in
observing the rites, his attention is so occupied with means and end,
and caught by the circuit in which these ideas revolve, and he is so
earnest in carr3'ing out the prescribed actions, that he cannot compare
them with the really effective actions, so as to discover their absurdity
and irrelevancy. In short, a state of mental dissociation is established
for the system of magical ideas. So far does illusion go, that probably
he regards the rites as the most important part of his proceedings.
But that is not really his deepest conviction : he trusts in Magic, and
keeps his bow-string dry.

{d) In the case of children we may assume, and in the more
backward races of men we may suspect, that the comparison of judg-
ments is difficult or, in the more abstract cases, even impossible, because
of the imperfect development of the cortex. There must be some
structural conditions of the free flow of energy through all organs of the
brain, corresponding with the associability and comparison of all ideas.
We may doubt if these conditions are complete even in good, cultivated
minds; since everybody finds one or another study or art especially
difficult for him ; or the freeing of himself from this or that sort of
prejudice especially repugnant. Such imperfections of structure, great-
est at the lowest levels of organization, and gradually decreasing as ideal
rationality is approached, we may call incoordination ; and, so far as it
obtains, the results must be somewhat similar to the discoordination,
the breaking down or interruption of organic efficiency, that occurs in
hysteria, hypnosis and some forms of insanity.

Effective incoordination may, however, be merely functional, from
defect of education, or for want of practice in thinking.

In either case, whether from defect of structure or from high
synaptic resistance, there will be failure of comparison and, therefore, of
criticism, and also (we may suppose) a greater intensity of imagination
and of dreaming and a liability to hallucination, such as is said to be
generally the case with immature minds.

Carveth Read 313

§ 7. The Ratiocination of Immature Minds.

So far as failure of comparison and criticism occurs there must be
an absence of Logic. Our Logic consists of a few universal principles,
generally accepted (their full recognition implying the widest compari-
son of types of judgment), with which any more particular judgment
may be compared in order to test its validity. One man may be a great
student of Logic and a very inefficient reasoner; another may never
have opened a text-book, or even a primer, and yet show, by the
definiteness of his judgments and the adequacy of his plans, that he is
a sort of incarnate Logic, that his mind works according to reason or (in
other words) according to the order of facts. Such men occur amongst
backward peoples. The Basuto chief, reported by Casalis in 1861
(quoted by Prof Haddon in Magic and Fetishism), said : " Sorcery only
exists in the mouths of those who speak of it. It is no more in the
power of a man to kill his fellow by the mere effect of his will, than it
would be to raise him from the dead. This is my opinion. Neverthe-
less, you sorcerers who hear me speak, use moderation." He knew the
falsity of the belief, and he also knew the force of the illusion.
Congenitally of superior mind, the instructions of experience had
brought him thus far ; and from his high level down to the region of
imbecility there are all grades. The average level is shown by the
prevalence of Magic and Animism. How are such beliefs arrived at ?

Two accounts of savage reasoning have been given by those who
admit that savages reason at all : one is that they reason correctly from
absurd premises; the other that they reason absurdly from correct
premises. There is (I think) some truth and some error in both these
doctrines. For the sake of comparison, let me make bold to take the
usual example of the syllogism, slightly altering the way of writing it
for reasons that will presently appear :

All men are mortal

II
.'. Socrates is mortal.
The sign of equality, written vertically, marks the minor premise —
" Socrates is a man " ; that is to say, for the purpose of the argument,
he is the same as other men.

It is now (I suppose) admitted that our reasoning, in everyday life,-
does not take place in this explicit formal way. We do not think first
of the major premise and then of the conclusion ; we need not think of
J, of Psych. VI 21

314 The Conditions of Belief in Immature Mifids

the major premise at all. If someone doubts the judgment that
Socrates is mortal, and asks for evidence, we may think of the major
premise, and then put it into words for the first time. In order that
premises may determine our judgments it is not necessary that they
should ever have been formulated ; as we see in the premise — " Magni-
tudes equal to the same magnitude are equal," which determined the
use of a common measure, the five fingers or the cubit, ages before it
was explicitly stated. Premises of this high degree of generality are
called ' forms of thought ' ; but forms of thought are established by
experience in very concrete material, either by repetition of experiences
constituted by similar relations, or even by a single impressive case ;
and that ' men are mortal ' is one of these forms\ Such forms deter-
mine our judgment by the assimilation of relations, ratiocination, or
analogy (in the widest sense of that word); and that 'Socrates is
mortal ' is a judgment so determined.

As for savages, experience settles also in their minds forms of
thought, latent major premises, which determine their judgments, and
which, as types of experience, are roughly true. They know that for a
man to fall into the fire, or to put his hand into it, hurts him ; that
when a man dances and sings he feels more energetic ; and that he
thrives by eating solid food. From such truths they draw inferences by
analogy. According to contagious Magic, thus :

Latent premise : To throw a man into the fire hurts him :

'I . .

.'.To throw in a lock of his hair hurts him.

Or, according to mimetic Magic, thus :

■ . . . "
.•.To throw in his image hurts him.

Of course, as the major premise is not consciously referred to, nor is the
minor. Explicit premises are no part of primitive reasoning, which
proceeds dynamically upon analogies and, if upon bad analogies, knows
not how to check or rectify them. The minor premise especially is
an invention of Logicians, because it is necessary to their province,
which is not reasoning, but proof.

But how can a man believe that a lock of hair, or nail-parings, or
the footprint of his enemy may, for the purpose of his action, be treated

^ It is often said that savages do not believe that man is naturally mortal ; and this
seems in many cases to be true. But the belief that man is not naturally mortal must not
be mistaken for a belief that, in fact, man does not die.

Carveth Read 315

as the enemy himself? Because he passionately desires it; because
the substitute intensifies his imagination of the enemy, and the action
he performs upon it gives him relief. How can a man believe that the
image of his enemy may be substituted for him ? Because by conceiving
it so (it need not really resemble him), it excites the same reactions
and gratifies his rage.

It will be noticed that the famous maxims of Magic, that whatever
has been in contact with a man — or that any likeness of a man — may be
substituted for him, appear to be grounds of the minor premise. It is
not to be supposed that the savage, who seems to act upon these
maxims, is explicitly aware of them, or has ever generalised and
formulated them. In whatever way they exist in his mind, their
derivation seems to be another case of illusory analogy. To injure an
integral part of a man injures the man himself; therefore to injure
a detached part of him has the same effect. The extension of this
inference to include clothes and footprints and its exaggeration of the
resulting injury are due to the force of hatred. Again, as a reflection
or shadow implies the presence of the man who casts it, so does his
picture or imaged

As to the invisible force, mana, or whatever it may be called,
pervading things that have once been in contact, and operating at a
distance, which many savages undoubtedly believe in, it is easy to point
to facts of experience, — light, sound, odour, infection — from which it
may have been abstracted. But surely so refined a notion cannot lie at
the foundation of Magic : we must begin the explanation with some
much simpler mental processes, such as have above been indicated, and
which seem to need no further explanation.

Again, there are certain other reasonings implied in savage practices,

where the error lies not so much in the minor premise as in the minor

terra. Thus :

Latent premise : Sense of energy is increased by singing, dancing etc.

II
.'. Magical power is increased by the same means.

Or, according to Animism, thus :

Latent premise : Men eat solid food.

.'. Ghosts eat ghostly food.
Here the minor premise seems to us plausible, in the light of savage

1 Cf . H. G. Spearing, The Childhood of Art, 92.

21—2

316 The Conditions of Belief in Immature Minds

ideas; but the minor terms are imaginary. There are no magical
powers and no ghosts ; and, therefore, the minor premise is still absurd.
In saying that there are no magical powers, I do not mean that the
magician has no professional powers, but that such powers as he has are
not magical.

It follows from these considerations that in the reasoning of Magic
and Animism, for the most part, if stated in our recognised forms, the
major premise is empirically true ; the minor premise is false : the form
of reasoning is the same as ours, in so far as it is analogical in the most
general sense (a comparison of relations) ; but it is invalid, because it is
analogical in the narrower sense in which, whilst the relations of terms
are similar, the terms themselves are not enough alike to justify the
inference. There are three types of ratiocination: (1) equations;
(2) parallel cases of causation and of class-attribution ; (3) analogies
of imagination — on a level (as far as proof goes) with poetical and
rhetorical ornaments, simile and metaphor. The natural progress of
reason consists in advancing from the third type to the second and
first, in which the minor premise becomes true; and this takes place
because greater definiteness of thought has high biological value.
Immature man, outside the necessary practical life (which might be
called the biological life), is not a rational but an imaginative animal ;
and most savage beliefs about Magic and Animism are derived by
analogies of imagination ^

§ 8. The Weakness and Gradual Decadence of
Imaginative Beliefs.

There are, it is true, some facts connected with superstition which,
with immature minds, may easily pass for good evidence. Some of
those who pray to Neptune are saved from shipwreck, and the drowned

^ A vague notion of causation plays a considerable part in some magical reasonings.
Amongst other things, it seems to be assumed that opposite effects have opposite causes.
In rain-rites the savage holds that

"Who makes wet weather must himself be wet ;
Who makes dry weather must himself be dry.

He sacrifices a black bull (or goat), the colour of clouds, for rain, and a white one for
sunshine (Examples from Dr Frazer's Spirits of the Corn and the Wild, c. 5). The
experiential grounds of these fancies lie in the opposite effects of rain and drought upon
the crops, of noise and silence in hunting, of light and darkness upon all the ongoings
of nature, etc.

Compare with this section "The Function of Relations in Thought," This Journal,
1911, IV., and Dr C. Mercier's New Logic, cc. xxii. and xxiii.

Carveth Read 317

are forgotten. By suggestion the sick are often healed, and the hale
struck down. Curses and incantations, if known to the intended
victims, fulfil themselves. If a magician has the common sense to make
rain only when the wet season approaches, the event is likely to confirm
his reputation. Sleight of hand and the advantages of a dark seance
are not unknown to a savage tutored in an old tradition of deceit. The
constant practice by a whole village of both magic and industry for the
same end, makes it impossible for ordinary mortals to see which of
them is the real agent of success. For the failures of magic or sacrifice
there are always notorious excuses. Hence the qualified expression
' almost entirely imaginative ' has been used to describe those vast
congeries of ideas which are characteristic of the immature mind as
contrasted with the civilised : the foregoing analyses have shown their
intrinsic structure.

Now, imagination-beliefs may seem indistinguishable in character
from perception-beliefs ; in immediate feeling-quality they are certainly
very much like them ; and, on a first consideration, they appear to have
at least as much influence over men's actions : but this is not true. In
course of time, they change, though the 'evidence' for them may
remain the same. Moulded from the first by desire and anxiety, they
remain plastic under the varying stress of those and other passions. In
a primitive agricultural community, preparation of the soil, hoeing,
reaping and harvesting go on (though with inferior tools and methods)
just as they do with us; and from age to age the processes are
generally (not always) confirmed or slowly improved. At the same time,
every such process is surrounded by a sort of aura of rites, which seem
to be carried out with equal, or greater, scrupulosity and conviction;
yet, age by age, the rites slowly atrophy and lose their meaning and
influence.

This weaker character of imagination-beliefs, their close alliance
with play-beliefs, is shown in yarious ways :

(a) The rites which express them are often carried out with
deception — practised on the crowd in a public performance, as by
obtaining from heaven a shower of rice, which (over night) has been
lodged in the tree tops, and is shaken down at the decisive moment;
or, in private practice, played off on the patient, by bringing a stone in
one's waist-belt and then extracting it from his body.

{h) Religious beliefs often combine incompatible elements, such as
acknowledgment of the superior wisdom and power of a god, whilst
employing devices to cheat or threats to punish him.

318 The Conditions of Belief in Immature Minds

(c) These imagination-beliefs break down under various trials :
(i) Economy, as in selling the Rice-mother when the price of
grain rises; or offering the gods forged paper-money,
instead of good,
(ii) Self-preservation, as in substituting the king's eldest son

for himself,
(iii) Compassion, as in burying with the dead puppets instead of
slaves (though in this considerateness economy may have
some part), or substituting in sacrifice a bull for a man.
(iv) Social indolence and fatigue ; whereby the meaning of rites
is forgotten, and the rites themselves are gradually
slurred and abbreviated. This must be an important
condition of the degeneration of rites, as it is of language,
(v) Foreign influence ; even, perhaps, repeated experience of
failure, etc.
{d) The beliefs of magic and animism are supported by intense
emotional excitement during the performance of the rites and cere-
monies that express them. Emotion is artificially stimulated, and
probably is felt to be necessary to sustain the illusion.

(e) The specific connexion of such beliefs with the play-attitude of
mind is shown :

(i) By their rites being accompanied by games, such as leaping,
swinging, spear-throwing — supposed to have some magical
efficacy,
(ii) The ceremonies themselves are often dances, dramas,

choruses,
(iii) With the degeneration of belief, the rites remain as
dramatic and musical pastimes; whilst the myths survive
in epic poems, fairy tales and ghost stories.

§ 9. The Utility of Imaginative Beliefs.

These imaginative beliefs and practices were the necessary result of
desires and anxieties about the necessaries of life, at a time when men
could not know any better, whilst a great development of fi'ee ratio-
cination was accompanying the growth of the brain from the anthropoid
to the human scale. To us they seem so absurd, and the practices
seem to waste so much time and care (at the best), whilst (at the worst)
they are so cruel and destructive, that we wonder what utility can
justify their prevalence and persistence.

Carveth Read 819

There must be some great utility to compensate for the mischief, not
in each case, but generally and on the whole ; nor yet consciously aimed
at, but accruing by the way. The utility of ethnic beliefs involved in
the worship of gods, I have indicated elsewhere {Natural and Social
Morals, c. ix. § 3), especially the political, artistic, and moral conse-
quences ; and I will now add some considerations as to the utility of
those rites of magic and ceremonies of religion which more particularly
subserve hunting and industry.

(a) They gratify the desire to do something, or to feel as if some-
thing were being done, toward the end desired, especially in the
intervals when really effective w;ork cannot be carried on, as whilst the
crops are growing and after the harvest : they allay anxieties and give
hope and confidence.

(6) So far as needs and interests are common to a tribe, village, or
other group, these ceremonies encourage social cooperation and unity,
and preserve tradition and the social integration of successive genera-
tions.

(c) But equally important it is that they are organized pastimes.
The men of backward societies, during a considerable part of their time,
have not enough to do. Social ceremonies keep people out of mischief,
and, at the same time, in various ways, exercise and develope their
faculties. With us industry is a sufficient occupation, or even too
engrossing, and circumstances keep us steady; so that, in leisure,
pastimes may be treated lightly. With the savage some pastimes must
present themselves as necessary periodical religious duties, whose
performance, in his belief, encourages and enhances industry. Our
games are free ifrom practical hopes and anxieties and are, for the most
part, a merely personal recreation ; but the more elaborate, such as
horse-racing, have still a social function ; or, like cricket and football, a
tribal character. The school, college, county or even the nation feels
deeply concerned about them.

§ 10. Black Magic and Red Religion.

As for the dark side of magic and superstition, it needs no other
explanation than crime, fanaticism and insanity. Love, jealousy,
hatred, greed, ferocious pride and the lust of power, are amongst the
causes that mould belief Any calling pursued in secret, under a social
ban, is of course demoralised. Where the interest of an organized
profession stands in a certain degree of antagonism to the public

320 The Conditions of Belief in Immature Minds

interest, it may become the starting point of unlimited abominations ;
and of this truth the interests of magicians and priests have furnished
the most terrific examples. The retrospect of human culture fills you
with dismay, but need not excite astonishment ; for human nature is
less adapted to its environment (chiefly social) than anything else in
the world; the development of the mind and of society has been too
recent for us reasonably to expect anything better.

§ 11. Imaginative Beliefs and Scientific Ideas.

If the character of those beliefs which distinguish the immature
mind of savages from the mind of scientific culture has been rightly
described in the foregoing pages, it becomes impossible to assent to the
derivation of science and scientific ideas from magic and animism.
Scientific ideas are rather to be considered as implicit in the practical
life of animals and men in their dealings with nature, and as having
been elicited especially from the practical pursuits of men by the
' dissociation of variable concomitants.' They are elicited from the
experiences of hunting and war, but especially of industry and com-
merce ; where the necessary succession of events of engrossing interest
is felt in every action, and can also, for the most part, be followed by
the eye, as in making and using weapons, tools, boats and in building
houses ; or where land and goods must be measured or counted and
exchanges carefully compared. Causation and number are inherent in
these operations ; and when the man of genius arrives, in whose mind
those .relations have become isolated, he sees them in the facts of
experience, and assimilates the facts under their appropriate relations.
But superstitious beliefs, though (as we have seen) they obtain a certain
empirical verification upon the assumption post hoc, propter hoc, are far
jfrpm being verified constantly and continuously ; nor is the process of
their fulfilment, step by step, capable of being felt or perceived. And
we have seen that the ratiocination by which such beliefs are arrived at
in detail always takes place by a false or groundless illation.

As industry becomes more and more continuous and exact and
arduous, it is more and more differentiated fi-om play ; and as it obtains
control over natural forces, superstition declines. Uncontrolled forces
are its stronghold; and the last prayer, like the first, is a prayer for
rain.

t

{Manuscript received 4 October, 1913.)

AN EXPEEIMENTAL INVESTIGATION OF
PERCEPTION.

By frank SMITH.

(From the Psychological Laboratory, University of Cambridge.)

1. Subjects of experiment.

2. Apparatus employed.

3. Method of investigation.

4. Results :

(a) Immediate interpretation of objects as a whole.
(6) Analysis of the objects perceived.

A. Experiments on Adults :

(i) Procedure adopted by the subjects,
(ii) Subjective factors.

(a) Imagery.

(b) Subjective additions.

(c) Objective changes in the pictures.

(d) Subjective certainty.

(e) Readiness of decision.
(y) Associations.

(g) Self-projection.

(A) Improbability.

(i) Influence of the first idea.
(Hi) Perception of colour,
(iv) Feeling tone.

B. Experiments on Children :
{i) Six years of age.

(ii) Twelve years of age.
(Hi) A mentally defective child.

5. General conclusions.

What are the factors involved in the process of perception ? Are
we justified in making a classification of individuals according to their
' type of perception,' as has been attempted by Binet^ ? Do adults and
children show any fundamental differences in their manner of per-
ceiving ? These are the questions which the following experiments
Avere undertaken to answer.

' L'Etude experimentale de V Intelligence, Paris, 1903, ch. xi.

322 An Expetnmental Investigation of Perception

1. Subjects of Experiment.

The subjects in these experiments were thirty-two adults and thirty-
eight school children of different ages. The majority of the adults were
Cambridge graduates and undergraduates ; with the exception of two
domestic servants, all of them had had excellent education. There were
seven women among the adults. Nine of the thirty-two were science
students. The children belonged to three neighbouring schools —
the — Grammar School, an elementary school in a good residential
district, and an elementary school in one of the poorest parts of the
town.

Experiments on the adults were carried out at the Cambridge
Psychological Laboratory ; in the case of the children I visited each of
the schools in turn, and worked in one of the class rooms.

2. Apparatus employed.

In the Laboratory the apparatus I used was a form of tachistoscope
devised by Mr Hales, which has already been described in this Journal^.
By means of this instrument transparent objects can be exposed for a
very brief period, and where nothing is said to the contrary it may be
assumed that the time of exposure in the present experiments was
about one-thirteenth of a second. Variations of the time interval were
tried occasionally, and will be mentioned in their place. In the school
experiments I used a smaller and simpler apparatus, which had a
photographic shutter in place of the pendulum. The time of exposure
in this case was longer, and probably not quite so constant, the average
time being about one-fifth of a second.

The effect of these very brief exposures was to spread out the process
of perception (and apprehension) over a longer time than usual, so that
the different factors involved therein could be noted with fair ease.
Even subjects who were quite unaccustomed to introspection were able
to indicate accurately what progress in the process they were making,
and how the progress came about.

The objects were presented in the form of lantern slides. They
varied in nature and in difficulty. They were chosen for various
reasons, — simplicity, complexity, familiarity or strangeness. The pictures

1 1908, n. 244fif.

Frank Smith 323

numbered 1 to 7 were coloured; the remainder were in black and white.
The following is a brief summary of their contents :

1. A railway track with an engine and passenger train travelling quickly towards
the right hand bottom corner. There are two pairs of rails, and also a pipe running
parallel to them. On the left are several posts and a green hedge. On the right the
ground falls away.

2. The platform of a small station, with the rails also showing on the right.
Several groups of people are standing about, the most prominent being an old lady
and her son, who are going to part. Half of a bridge is \'isible, rising from the
platform on the left and turning at right angles to cross the line. Behind the bridge
is a mass of station buildings, and in the left foreground are rails and a lamp-post.

3. Two men : a sweep on the left with a bundle of brushes on his shoulder, and
a gentleman on the right with frock coat and silk hat. A vertical line down the
centre indicates that they are about to collide with each other in turning the corner.

4. The sequel to No. 3. The corner is now to the right and the sweep's brushes
have struck the gentleman in the face and knocked his hat off. His glove is also
falling.

6. A child's face with a circular background of blue. The child's hair is unkempt
and his right eye is screwed up mischievously, thus drawing the mouth to one side.
The upper part of a white garment is shown round the neck but nothing further can
be seen.

6. A man is busy tarring a wall with a long brush, the tar being in a bucket on
his left. Another man seen in profile on the right has just begun to expostulate with
him and the former, raising his brush on the left and turning his head to the right,
has tarred the face of a lady who is just coming round the corner of the wall on the
left.

7. On a. table at the bottom of the picture are a cat, an iukpot, a book and a
vase of flowers. Above the cat, and quite near it, is a bird in a cage. At the top of
the picture is a scroll containing the words : " A bird fancier."

8. A pug dog on the left and a Persian cat on the right — lying on a table, with
their faces towards the observer.

9. A three-masted ship of the Victory type, lying broadside-on in calm water.
The port holes are very prominent. On the right and beyond is another ship of the
same type.

10. A lifeboat standing on the slips in a harbour. The slips are very prominent
in the foreground. Beyond the bay is a town with several large buildings showing
indistinctly.

11. A donkey, standing nearly in profile, with its head to the right. The ground
is rough and rising to the right. The donkey's ears are stretched out, the hind feet
together and the fore feet slightly apart.

12. A Brittany bed : consisting of a richly carved cupboard containing two beds,
one above the other. The doors slide apart and are open. Kneeling on a ledge by
the upper bed is a woman in the costume of a Brittany peasant, and there is someone
sleeping in the lower bed. Outside the cabinet is a carved settee.

13. A muzzled bear in the middle of a suburban street, standing on its hind
legs and holding a high pole. There are houses with prominent bow windows on

324 An Experimental Investigation of Perception

each side and the sun, evidently shining brightly on the right, is casting strong
shadows.

14. An elephant standing in profile, head to the right. The trunk is curved
upwards and near it is the keeper. There are trees in the background and some
people show faintly on the right.

15. Six yachts, at varying distances, sailing up a lake which is backed by ranges
of hills. All the yachts are incUned towards the right and there are ripples on the
water.

16. A chaos, made by a few blots of ink. To the left are a few straight lines
which suggest a box or a small hut, but the rest is without meaning.

17. Queens' College, Cambridge, as seen from Silver Street bridge. The college
buildings are on the right, and dense trees on the left. The river flows between, and
is spanned by a wooden bridge near which are two small punts.

It may be here remarked that it was generally impossible to foretell
whether any given picture would prove easy or difficult to a given
subject, — some subjects finding pictures very difficult which caused no
trouble to others, while the latter perhaps stumbled at those pictures
which appeared easy to the majority.

3. Method of Investigation.

The experiments were carried out in the following way : first, the
subject was told something about the nature of the experiment, and
the part he was to play therein. He was then shown exactly where the
image would fiall on the screen, and was also habituated to the speed of
exposure. If desired, a small fixation mark was affixed to the screen.
A lantern slide was then put into place, a warning signal was given,
and about a second later the exposure was made. The subject reported
as fully as possible what he had seen; he was asked also to describe
any associations the picture had called up, any feelings, emotions, etc.
he had experienced, any inferences he had made, and so on. This
introspective record I wrote down almost verbatim. After the subject
had told me all he could remember, I questioned him in order to clear
up any valgueness of statement and to make sure that nothing had been
forgotten. When the report on the first flash was completed, a second
exposure was made of the same picture, and the subject reported the
additions or corrections he wished to make. This procedure was re-
peated several times, until the subject had seen the whole picture, or
had become tired of it, or had found it too difficult to solve completely.
Hence the number of exposures varied with each picture and each
subject, sometimes five or six being found sufficient, and in some cases
a hundred not being found too many. . Occasionally, after the warning

Frank Smith 325

had been given, the picture was not shown but the subject was asked
to report his experiences while he was preparing to attend to the
coming flash. The records of these ' fore-periods ' threw light on the
subjects' activity in planning out their work, and showed how far they
attacked the problem systematically.

4. Results.

An analysis of the introspective records leads to the following
results :

The process of perception, as it here occurs, takes place in two
stages. There is first (a) an immediate interpretation of the object as
a whole, and next (b) an analysis of this vaguely apprehended whole
into its component parts. Subjects of all ages showed striking uni-
formity in the first process, whereas in the second there was a no less
striking variability.

(a) The interpretation is immediate, and my experiments did not
throw much light on this point, though with other material it might be
possible to investigate it more completely. " We begin," as Stout says,
"by apprehending a whole in its distinctionless totality, and then
proceed to unfold its details^" It is important to insist that previous
experiences, which give meaning and distinction to this complexity,
need not themselves be present to the mind. Some writers have
insisted very strongly on the necessity of the revival of past experiences
in perception, but these experiments support the view that the mind
can give a meaning to experience without any such revival taking placed
For the time being we may go no further than this general interpreta-
tion. " It is possible to think of a whole in its unity and distinctness,
without discerning all or even any of its component details^"

As examples of this instantaneous recognition I quote from the
introspection of different subjects :

" Recognition is qtiite instantaneous. I didn't notice any process."
" I couldn't distinguish between the instant of seeing and recognition."
" Recognition was immediate."
" Recognition is very rapid, and the word is as quick as the recognition."

1 G. F. Stout, Groundwork of Psychology, 1911, 70, 71.

2 Cf. "What is Perception?" by C. H. Judd, J. of Philos., Psychol., and Sci. Methods,
1909, VI.

^ G. F. Stout, Analytic Psychology, i. 78.

326 A7i Ex2yerimental Investigation of Perce2)tion

In rarer cases recognition, although immediate, developed greater
clarity after a perceptible interval. Thus :

" Recognition is quite immediate, but it gets much clearer as one thinks about it."
" Recognition is instantaneous, though there's something like an unfolding."

A few subjects (at least three of the thirty-two, — probably more, as
I did not question all about the matter), distinguished between the
recognition of differences in luminosity and the recognition of form.
That is, they first saw an illuminated patch with certain black marks,
and then these immediately assumed some definite shape. Examples
of this distinction are found in the following :

"A white background and black things in front. It looked like acrobats in all
sorts of attitudes. There was a slight searching for an explanation, but not long."
(No. 16.)

" My first impression was of black masses without outline. Then it suggested a
dog."

I am not sure that these cases are as genuine as they look. Two of
the three subjects were quite untrained in introspection, and the third
gave very few answers of this nature. In one of the subjects I think it
was due to excessive caution. This subject always began with the
most general interpretation he could, and only approached the particular
after great deliberation. Thus in picture No. 6 he described the central
man as follows : " I saw something like an inverted Y. I take it to be
a man," And in picture No. 11 he said : " The picture of a quadruped.
There is something sticking out of the head — horns or ears. I associated
it with a cow."

It will be seen that by this method of answering he left room for
retreat in case of mistake, and gave first an interpretation broad enough
to avoid nearly every likelihood of error. In one of the remaining two
subjects I think more was seen than she reported at first — either through
vagueness or forge tfulness. In the picture of the bear (No. 13), she
said:

" Nothing but white and black. I couldn't say what, imless the white is snow.
I was quite ready, but not so concentrated as usual. It might have been a boy
standing up in a snowy street. But I dismissed that as impossible. It was all too
blurred. He came up involuntarily."

I then asked whether any more came back to her, even vaguely, and
she replied: "Well, the boy might be standing with his legs apart."
It is therefore probable that the perception of form here was more
complete than the first report indicated.

Frank Smith 327

The third case of this kind was one who had had long training in
psychology and in experimental work. With the nonsense picture (No.
16) she described her experience as follows :

" I can't say much of the first stage. The picture rolls on, but doesn't take a final
interpretation at first. I feel there is a preparation for a definite final something.
I had confidence of something definite which would come up when the picture was
gone. Then I saw other parts, and the previous interpretation was choked down by
this new sensation, and so the other never appeared in full consciousness."

Perhaps I have unduly emphasized these few exceptions, but it is
because they are the more striking in face of the uniformity found with
the remaining subjects. In my own case I had the experience of a
blinding flash of light from which the awareness of certain forms
immediately emerged. That is, while I was still in a state of surprise
at the speed and brightness of the exposure, I found I was conscious of
what I had seen, without the least effort being necessary. The whole
process was one and indivisible.

In order to try and discover something more about this point I
modified the conditions of the experiments by cutting down the time of
the exposure to considerably less than one-hundreth of a second. Yet
even with this very short exposure the immediateness of recognition
was unchanged. I quote the following reports :

" Recognition was as immediate as before."

" A three-masted ship of the type Victory. There are port holes and white band.
It is facing my right. The exposure was considerably shorter. Open sea. No land
in the background. I have an impression of small boats, but I am uncertain. My
certainty about the general thing is just the same, but the details are more uncertain.
The recognition of the three-masted ship was just as immediate as it would be
with a longer exposure." (No. 9.)

Subjects were usually unconscious of any change in the speed of the
flash in the cases where I secretly altered it from the normal to the
very rapid time. They often reported, however, that the picture had
become " very dull," and conversely, when the duration of exposure was
lengthened, they would remark that the picture had become "much
brighter."

I also tried the effect of showing a picture which was outside the
experience of the subjects. This proved to be a difficult matter, and
pictures Nos. 12 and 16 were the best I was able to obtain. Both were
interpreted very differently by different subjects, but it is significant
that there was never any hesitation about interpreting them as something
of every-day experience. The former of these pictures received such
descriptions as the following :

328 An Exi)erime7ital Investigation of Peixeption

" It looked like the balcony of a window and somebody standing on it."
"Something like the frontage of the Laboratory, but much more pretentious.
I was not aware of any interval between the sensation and the interpretation."

" It was like a glass case, standing in a sort of trough. Immediate interpre-
tation."

In the case of the latter picture similar fantastic accounts were
given :

" It looks like a most extraordinary coat of arms, made up of quite vmintelligible
figures."

" Like a cartoon in Punch."

" It might be a lot of rapiers and boxing gloves — the paraphernalia of a gymnast."

" Three little animals with funny little heads."

Although interpretation was so immediate, doubt might afterwards
arise as to its accuracy, but this never prevented an immediate judgment
being made in subsequent experiments. Even the faintest sensations will
bring about the perception of complex things. Picture No. 12, by its
regular form, might suggest a door or a window, or the part like a settee
near the floor might even suggest a fireplace and fender. Hence when
the picture was incompletely and imperfectly seen, ideas connected with
one or other of these interpretations at once arose, and the picture was
filled in mentally by association and expectation. In this way so strange
an object was interpreted in terms of very familiar objects.

Our experience in normal life is often very similar. No doubt, we
repeatedly misinterpret the things about us in our rapid glances to and
fro, and we only look more closely when doubt arises, or when we are
greatly surprised at what we have seen. I have occasionally made most
grotesque mistakes in glancing quickly along a crowded street while
thinking of something quite remote from the scene ; a second careful
look has failed to reveal any very obvious reason for the first interpre-
tation.

It is interesting also to note that many young children, who had at
first seen No. 12 as a window or a door, persisted in the same interpre-
tation when they saw the picture properly at the end of the experiment.
Here the mental factors at work were apparently much stronger than
the visual sensations which were responsible for the process.

(6) After the immediate recognition, which, as I have shown, is of
universal occurrence, the mind seeks to know the details, and to analyse
the vague whole into its various parts. In this process of analysis,
enormous individual differences appear; indeed, at first sight, it is
difficult to discover any uniformity whatever. I have found it vain to
attempt to set up 'types' of individuals with distinguishing and

Frank Smith 329

constant characteristics, because the peculiarities of what I thought
might be a ' type ' rarely turned out to be constant, and often appeared
in other subjects whom I had decided on other grounds to belong to
another type'.

A, Expei'iments on Adults.

I propose to give, first, an account of the introspection from adult
subjects, reserving the children's introspection till later, in order that
a clearer comparison may be made between the two groups.

(i) Procedure adopted by the subjects. Subjects may be roughly
divided into two classes according to their method of procedure during
the analysis. Some continued to see each picture as a whole, except
when an outstanding feature attracted their attention or happened to
fall on the focus of vision. Sometimes one detail would appear so
prominent that it was seen continually, even when there was a faint
intention to look away from it. Other subjects made a more or less
systematic search for details. In the most pronounced cases of this
class, they were able to concentrate on an exceedingly small area of
the screen, and so saw only a very minute portion of the picture clearly
and the rest very vaguely :

" The clear part was a small semi-circle at the bottom."
" One seems to see just one part at a time."

The subjects in the first class made, in general, very little progress
in the solving of the pictures, and, indeed, often asserted they were
making no progress at all. Typical reports are :

" I haven't got much more than at first."

" It seems to me one gets more out of it the first time. I suppose that my mind
gets confused."

" I think the second time was as clear as any time since."

These subjects did not usually persevere much. They began to feel
baffled after a few exposures, and unable to proceed :

" I am gradually getting the feeling I am stuck and can go no further."
" I am beginning to feel hopeless about it. With increased hopelessness comes
increased impatience to see the picture in a long exposure."

1 "Moreover it happened that any subject whatsoever, who from his description of one
picture must have been classified as an 'obversation type,' would behave in the description
of the next picture as if he were of the 'description type.' We have therefore preferred
to abandon this attempt (to classify by types), rather than undertake an arbitrary
and artificial classification." Marie Borst, " Untersuchungen iiber die Erziebarkeit und
die Treue der Aussage," Beitr. z. Psychol, d. Aussage, 2'« Folge, 1905-6, 105-6.

J. of Psych. VI 22

330 An Experimental Investigation of Perception

Those, on the other hand, who searched closely for details made
steady process and were content to spend a long time with each
picture. They showed very great anxiety over their accurate pro-
cedure :

" I was too high. I don't get it unless I hit the exact spot."
" I seem to see only the point where my eyes are concentrated."
"After the first impression 1 should naturally go for the details."

This power of minute observation was most strikingly developed in
students of science, and the two best cases were those of a University
Lecturer who has spent many years in scientific investigation, and a
Research Student of some years' experience. I will quote from the
report of the latter, given with picture No. 3, and in order to show the
development of his analytical method I will present his answers after
the first flash and after the tenth flash. It will be seen that all parts
of the picture are mentioned in the first, but only a very small area in
the second answer :

" A sweep and a gentleman. The sweep is coming round a corner apparently.
The well-dressed man has a top hat and a cut-away coat. He can't see the sweep
apparently. It may not be a cornei-, but there is a line between them which
suggests one. The sweep is black, and has brushes over the left shoulder. I believe
he had something in his left hand. It looked like a bag without handles, as though
for soot. The upper part of the picture was more definite on the whole. The
gentleman had a moustache ; I am not so certain about a beai-d, but I think he had
one. There were no bricks marked in the wall. I think it was a corner, because
each one seemed ignorant of the other's approach. There is something on the right
of the sweep in the distance which gave me the idea of a lamp. The gentleman was
a tiny bit bent. The sweep had the attitude of a man carrying something. He
looked slovenly. I couldn't see the top of the wall : it went straight up out of the
picture. I didn't see the edge of the pavement if there is one."

I asked him if the picture was coloured or plain, and he replied :

" Plain. It was not a photo. It may have been done with ink and brush. I didn't
see the picture in detail enough for that. It reminded me of a picture I saw two
years ago in Birmingham."

I further asked him if he saw all these things at the moment of exposure,
and he answered :

" I keep recalling the image, and I can recall almost as faithfully as I saw. The
association with the Birmingham picture came, however, at the moment of exposure,
but got dropped for a time."

This long description, it will be seen, gives a comparatively accurate
and full grasp of the picture as a whole, with a very probable explanation

Frank Smith 331

of the relation between the two main figures. But at the tenth flash
his attention was turned to one part only :

" I saw the sweep's face and eyes. They are not very dirty, but a little black.
They are in pro61e. He seems to he in a hurry judging from his face. His eyes are
fairly wide open. He is too boyish for a beard. The nose is fairly prominent. I am
not quite certain whether there is a moustache or not. It may be the shadow of his
nose. I did not see his ear. I haven't seen his boots at all. And I'm not certain
about the pavement. I'll look at the right bottom corner — then I shall want to go
to the other corner and compare."

When we compare the following answers from other subjects with
those just quoted we see clearly the difference between the two
classes :

"A face ; I think a woman's." (No. 5.)
"A street." (No. 17.)

" Two men fighting. Then I thought they might not be fighting, but clasped
in one another's arms. The one on the right had light trousers on." (No. 4.)

These are the complete reports after the first exposition of a picture.

Subjects of the second class, who analysed fairly thoroughly, usually
reported that the pictures gradually lost their unity and became mere
collections of parts :

" It is made much more now of pieces put together than at the beginning. This

is due to seeing different parts separately."

" It is made up of parts and distinctly less unified."

" The picture seems to have got divided into two parts." (No. 3.)

"The more I see the picture the more I take bits of it. If you showed it me

twenty times I should never take it all in."
" It needs an effort to recall the whole."

The same effect was produced when I showed pictures to some subjects
with an exposure of two or three seconds : " I got no impression of the
whole." Once, when I changed the time of exposure from one second
to one-thirteenth of a second, a subject remarked : " The picture was
more of a whole then, without time to concentrate."

There was one case, and only one, where mention was made of the
reverse effect of seeing the picture as a unity more and more as the
experiment proceeded : " The picture is now a whole, and I see every-
thing vanishing away to a point about the middle ot the picture. There
is more unity than at the beginning."

The group with more strongly marked analytical powers showed
more systematic procedure and mental activity^ than the others in

1 Cf. A. J. Schulz : "We minht speak of an 'active' and 'passive' type. The first
analyses the exposed object — compares, unites, separates. The other behaves more

22—2

332 An Experinuntal Investigation of Perception

many ways. They constantly planned out their procedure beforehand
and fulfilled their aim in spite of all obstacles :

" I was going to look on my right. I had a sort of mental struggle first and then
decided to look for the people on my right."

" My attention to the man was intentional." (No. 14.)

" I'll look to the stern." (No. 9.)

" I can't make out the people on my left. I must give attention there." (No. 2.)

This systematic procedure is well illustrated by the following extract
from the introspection of the University Lecturer already mentioned.
The picture is No. 1, and the figure before each answer refers to the
number of the exposure :

(35) " Let me think what to do. I think they are ordinary sleepers. I'll look.'

(36) " I got a general impression then of green lines — actually on the lines. It
doesn't seem quite normal."

(37) " Yes, the same. I was a bit too low."

(38) " I couldn't get the nature of the sleepers. There's a tremendous lot of green
which seems out of place."

(39) " I am not sure if there aren't more than two sets of lines. I can't be
certain about the number."

(40) " My impression then was of four lines and a rail to one side. I had that
before. It may be a pure illusion."

(41) " Yes, an impression again of more than two sets of rails. I can't be certain
about the number."

(42) " I am sure there are more than two pairs."

(43) " I think the train is on the far rails. But I think there is more than one
pair there. It is frightfully difficult to decide."

(44) •' I can't be sure whether 1 am taking this rail on the left from its mate.
I have an idea of more than four. I'll give it up for a bit."

This is an excellent instance of consistent application to the solution
of a small point ; it is but one of numerous similar instances given by
this subject throughout the experiments.

Some subjects were only guided to an active search by some kind of
outside stimulus. The stimulus might come from a part of the picture
which had aroused curiosity, or arise from the monotony of seeing the
same thing several times :

" I feel now I want to look at the expression, to find what it is."
" I am bored with the train, and am looking round it."

passively, and lets the picture simply 'work on him,' in order to enumerate the elements
afterwards one after the other, without concerning himself about their relations. How-
ever, a sharp distinction is impossible. Under favourable conditions, and for certain
relations (especially identity or strong similarity) all subjects were active — of course in
different degrees. But the same subjects were not constant." " Untersuchungen iiber die
Wirkung gleicher Reize auf die Auffassung bei momentaner Exposition," Ztsch. f. Psychol.
Lii. 251.

Frank Smith 333

In total contrast to this active direction of the perception process,
some subjects showed a listlessness and lack of method which is best
called ' passivity.' Sometimes it arose merely from lack of interest, as
in the following :

" I feel more passive and less curious about details."
" I have been letting things come as they will."
" I haven't tried to find any explanation."

The difficulty of a picture was sometimes enough to prevent active
effort:

" It's an awful effort to look at one thing. It's easier to be passive than active."
" It is more difficult to concentrate on a detail in this picture — there's such a lot
to see." (No. 2.)

Certain outstanding features of a picture would often attract
attention, and would prove stronger than the subject's intention to look
at another part :

" The curve of the trunk attracted me again." (No. 14.)

" I meant to look that time at the side, but the engine again attracted me."
(No. 1.)

" I meant to look at the two people, but the other platfonn caught my eye."
(No. 2.)

" I can't get away from the bear in the foreground." (No. 13.)

Indeed sometimes this attractiveness of striking parts prevailed
over the opposing intentions of subjects who showed systematic power
to direct the process :

"The rails always catch my eye. I didn't want to see them then." (No. 1.)
" The forepart is so conspicuous that it attracts my attention."

I agree with Schulz^ that most subjects show the characteristics of
either ' type ' at different times. But among my subjects a few could
clearly be designated as ' active ' (nearly all of these were science
students) ; while at the other extreme some subjects were almost wholly
passive (science students were not found in this class). There remained,
however, a large group, the members of which showed at various times
certain characteristics of both classes (among these were a few of the
science students). The majority of this group were often active in the
fore-period : they would, for example, decide to look for some definite
detail, or to settle some doubt ; but in the actual exposure their
attention would be passively diverted to some outstanding feature,
and their intention would not be carried out.

1 See footnote, pp. 331, 332.

334 An Eocperimental Investigation of Perception

The advantages of a scientific training were well-marked : a rough
calculation shows that the science students made three times as many
inferences as the others ^

Several of the almost wholly passive subjects made no inferences
whatever, not even, for example, the obvious inference that when there
was a shadow marked in a picture the sun must have been shining.
This was the commonest inference of all, and was made occasionally by
children. No doubt all subjects could make this inference if directly
asked ; the important point is that they did not do so spontaneously.
One subject noticed that the feet of the central man in picture No. 6
were turned away from him, yet he only remarked on it as a curious
fact : he certainly did not infer that the man's back must also be turned
towards him.

Active subjects not only made more inferences than passive ones,
they also carried them further. Thus, they not only saw that the sun
must be shining in picture No. 13, they also added that it must be
rather high in the sky, because of the fairly short shadows. Two subjects
decided that as there were three or four posts on the left in picture
No. 1, therefore they must be telegraph posts and not signals; whereas
most subjects were content to leave the question open. Other typical
inferences that were made were the following :

" The wind must be blowing from the left." (No. 15.)
" The train must have been photographed from a low position." (No. 1.)
" It must be a street comer, because the sweep and the gentleman do not see each
other." (No. 3.)

" The l)ear's face must be towards the stick, because it is all in shadow."

Inferences were not always so correct as these I have quoted, and some-
times led the subject quite astray. Thus in No. 14, one subject
thought he saw fur on the animal's legs, and inferred a shire horse, and
this idea persisted for a very long time. The ears of the donkey were
occasionally seen as horns, a cow being inferred. The bear was often
seen as a man in the first flash, holding a brush, and the picture was
inferred to represent a crossing-sweeper. This interpretation occurred
in a surprisingly large number of cases, and occasionally could only be
dismissed with difficulty.

Hence a wrong inference at the outset might add very considerably
to a subject's difficulties, though, as a rule, one who could infer quickly
and at the same time had independence of judgment and activity of

^ These facts, while they do not prove that a scientific training will produce an analy-
tical power of observation, suggest a high degree of correlation between the two.

Frank Smith ^35

outlook, had a strong advantage over others. By inference he was able
to arrive at various possibilities and hypotheses to which he might turn
his attention.

So far as I could tell, inference was weaker in women than in men.
Of the five women subjects whom I investigated fully, inference was
practically absent in four. By the remaining subject only very simple
deductions were made, such as the position of the camera which took
the picture and the length of the exposure (this subject was interested
in photography). The limited number of cases, however, makes it
impossible to speak on this matter with much certainty.

Pictures were divided by the subjects into two classes : simple and
complex. The simple were those whose subject-matter was only one
thing — a donkey, a bear, etc. The complex were those which related
an incident, and whose meaning was more or less hidden, and depended
on the discovery of the relations of the figures to one another (Nos. 2, 3,
4, 6). Now detail obviously plays a smaller part in the complex picture
than it does in the simple, and subjects who could make an accurate
and exhaustive survey of a simple picture did not always arrive at a
satisfactory solution of the incident pictures. Hence it was that some
subjects excelled in the incident pictures without any very careful
analysis of details. This was well expressed by one subject in passing
from No. 1 to No. 6 :

" I feel here a marked difference of perception. In the engine picture it was a
question of details. Here it is a question of the sense of the story, for which
details are unnecessary. This is a proof in the case of art that pictures that tell
stories almost necessarily force to superficial contemplation. I have a feeling that
I have seen less of this than of the train, for the reason that I was occupied with
the story it tells."

Another subject said :

" The parts hardly exist in a story picture. Details would be a nuisance. So
I keep to the story interest."

One subject, however, said that details were far more important in
a problem picture " as the problem is greater : the figures might be
doing a thousand things." But this subject was not strikingly successful
in solving such pictures.

The differences between scientific and other students can be expressed
more accurately by figures. Thus, if the number of objects and qualities
of objects correctly mentioned by a subject in a given number of pictures
be tabulated, we shall be better able to see how far a division into
classes is justified. It is significant that of the five subjects who stand

336 An Experimental Investigation of PercejMon

highest in such a table (with an average of 306 marks for each picture),
four had had scientific training, whereas of the five lowest on the list
(with an average of 13'3), only one was a scientist, and he was the
highest of those five. The remaining three science students occupy the
sixth, ninth and twentieth places respectively on the list. This shows
that science students possess a decided advantage, in general, over other
students as regards the discovery of the contents of an object, but that
this advantage is not invariably possessed by all such students.

We may carry out this method of tabulation still further. In giving
correct reasons for things, four of the five best subjects were scientists,
but only one of the five worst. Similarly, of the five who mentioned
the largest number of correct positions occupied by different parts of
the pictures, three were scientists, and of the five worst in this respect
only two were scientists. In describing the movements or actions of
the figures in the pictures, four of the five best subjects and only one
of the five worst had had scientific training.

Of the five who made the fewest mistakes and the five who made
the most mistakes in each picture there was only one person of scientific
training in each group. But to get a correct measure of accuracy we
must divide the total number of mistakes by the number of objects
correctly seen, since subjects who mentioned a large number of objects
ran a far greater risk of error than did those who mentioned only a few.
Of the most accurate five determined in this way three were scientists,
whilst among the least accurate five there were no scientists.

In every respect, therefore, the scientists taken as a group proved
superior to the others, but this superiority did not necessarily apply to
them individually.

(ii) Subjective Factors. There are a number of subjective factors
which can be best treated separately before an attempt is made to
deduce from them more general conclusions.

(a) Imagery. In varying degree visual imagery was, of course,
very common among the subjects. It was absent in two cases only. In
one of the cases it was replaced by " a sort of verbal memory." In the
other case there was a nearly complete lack of all kinds of imagery, and
this subject frequently complained of his disadvantage. Thus, he found
great difficulty in locating special parts of the pictures in the fore-periods.
His method was to examine very small areas at a time, and he tried to
fixate the exact spot before the exposure, but as he had no image he
frequently made mistakes of localisation. He was also more liable than
most subjects to forget to tell me all he had seen unless he told me at once.

Frank Smith 337

As a rule visual imagery was a very useful aid. There were cases
where it seemed to be almost as clear as the percept, and could be
resorted to, even after a considerable interval, in order to settle some
doubt or answer some question, much as one would look, under normal
conditions, at the actual object. I quote some examples :

" Jiist after I said that I wondered where his feet would be. 1 argued that
his right foot would be behind, so I recalled the image, and I think it is so."

" One moment ! There are railings, I remember now. I was thinking, and saw
them in my image." (No. 14.)

" I believe I project the image on the glass and compare the two."

" I can recall the image almost as faithfully as I saw it."

The subject from whom I have last quoted thought he saw the sweep
coloured blue in one exposure, though he was not able to see the blue
again in later exposures. Yet he could revive either of the two images
— with and without colour — quite easily, and he realised that he had
only seen the blue sweep once. The two images seemed to remain
quite separate and constant in his mind.

In many cases imagery was selective, and might be good for form
and not for colour, or good for outlines and poor for detail, and so on.

A most interesting discovery was the peculiar behaviour shown by the
visual image in a few subjects, in that it seemed to possess a 'self-activity,'
and began to change and develop almost at once. That a mental image
changes as time goes on is well known \ but the change in this case
seemed to begin immediately. There were five clear cases of this kind,
and one was specially remarkable. The pictures were intensely real to
this subject, and her image had all the movement and change that
reality itself has. I quote first the most important parts of her report
on picture No. 13 (the numbers again indicating the different flashes) :

(1) " There was a street and it was sunny. Whether there was a man or a bear
with a pole or not — it's something with a pole. It's something too big for a
man, and looked like a bear standing on its hind legs. It felt hot and sunny."

(2) " I feel sure it's a bear with a pole in its paws. It is still hot and dusty and
sunny."

(3) "I am quite sure it's a bear. It seemed brighter and sunnier. I had
rather a feeling of fear — I didn't like to see a bear in a thoroughfare. There was
nothing round about it."

(4) " I see the shadow of the bear distinctly on the left. The background seems
fearfully indistinct. I can't get away from the bear in the foreground. Some

1 "The imagery tends, with the lapse of time, toward the imagery of the object
represented by the picture, and with this change takes on characteristics that belong to
the object, but which are not represented by the picture." F. Kuhlmann, "On the
Analysis of the Memory Consciousness," Amer. J. of Paychol. 1907, 411.

338 An Experimental Investigation of PercejJtion

movement seems to go on in my visual image on the left. There is a man in a straw
hat, and people crossing the road a long way behind the bear. I still feel it's a fear-
some thing."

(6) " Yes, there are houses on the left and something on the right. Still dusty
and sunny and a foreign land. I don't like seeing a bear on the road."

(8) " ...I have a feeling of wishing they would take the bear away."

(9) " ...I don't know how much I see and how much I imagine, liecause I feel it
is in a big town, and the bear has no business there."

(12) " I feel there are two men about, but I don't see them. There's still that
man in a straw hat walking across the road, though he's only in my image."

(16) " I see the men in the after-image — the keepers. I don't see them in the
picture. There's a lot of movement afterward in my image."

In this case the cause for these additions seems to be her own fear
because of the unusual sight of an unattended bear in the streets. It
will be noticed that she is conscious of the ' activity ' of her image.
I quote a second example from the same subject where the additions
seem to be caused more by associations than by imagination. The
picture is No. 1 :

(1) "A Great Northern train coming towards me. I can hear it and I feel that
it is going towards a timnel. It is certainly Great Northern and going to Scotland.
I imagine I saw a child waving a handkerchief on the left. I stiU hear the noise of
it — puffing and going. 1 imagine luggage and people and everything. There's a
tremendous lot of movement."

(2) "...It is full of people in all the carriages. I have a pleasant feeling of
excitement about it. I'm sure it's on a long journey."

(3) " ...I still feel somebody is waving out of the window to my left but I can't
see them. I still feel it is going, and hear and smell it."

(4) " . . .1 was looking for the person on my left. I couldn't see him, but I feel he's
there... I see the people in the compartment — the way they are dressed and every-
thing."

(6) "...I still feel it's about to go through a timnel. I feel the landscape
changing as the train goes on. The people all look like traveller — as though they
have settled down for a long journey. It couldn't be a local train."

(7) "I think some of the windows are open. But I don't see them very dis-
tinctly. It's slowing down a little. I don't know why."

When I showed the picture at the end she added :

" I was going to Edinburgh by day by the Great Northern — though I have never
been to Edinburgh by day. Yet I felt it was the same journey — though I was sure
I was not going to London in this case as I was when I travelled on that route by
day. The people were a mixed crowd — Americans and such like. It was not a
crowd on a local train. There was a woman with a veil. I have to pull myself
together to see it as a picture — it's so absolutely real."

This case was rather exceptional, though there were four others who
belonged to the same category. One of them, in looking at the donkey,

Frank Smith 339

got an idea that there was also a pony's head somewhere attached to the
donkey, and this idea persisted with varying strength through several
flashes, though he added :

" I am not sure whether I see this or whether it is imagination. It suggested
that to me. I saw an image of a pony nibbling. It is a matter of perception at first
but, may be, worked upon in my image."

Evidently he was not sure exactly what had been perceived and what
had been imagined.

Subjects with mobile imagery may easily be unconsciously un-
truthfulS and mistake imagination for fact. Such people are known in
ordinary life, wholly unconscious of their failing. But, as I have shown,
a subject with mobile imagery may be fully conscious of the difference
between image and percept.

Other kinds of imagery than visual were not very common. Olfactory
and auditory images were only mentioned by one subject, verbal-motor
imagery by three. Unpractised subjects probably overlooked the nature
of their imagery in many cases.

(6) Subjective additions. Very closely connected with the pheno-
menon of self-active imagery was the tendency shown by some subjects
to make additions from their associations and imagination. This
tendency was shown by all the five subjects who had self-active imagery,
and also by a few others in a less marked degree, one of whom possessed
either no visual imagery at all, or so little that it was of no importance
in his mental life. However, it was scarcely more than a tendency in
his case, about which he was cautious it is true, but which he could
keep in control. Thus, he said : " I am almost ready to persuade myself
I can see a man on the smallest provocation."

Other typical reports dealing with subjective additions were the
following :

" I feel now I could see anything I wanted in this face, as regards expression."
" I thought I saw a turban on the figure." (No. 14.)
" I think there are some trappings on the elephant — part of a howdah."
" My mind wanders afterwards, and seems to raise up new things."
" I thought bulldogs wouldn't have ribbons on, and so I seemed to see heavy
collars."

^ See "Sur las transformations de nos images mentales," by Jean Philippe, Rev.
philos. XLiii. 482: "When a new state of an image has replaced the former state of this
same image, it presents itself alone in the place and instead of the former one : we cannot
therefore compare these two successive states of the image, nor see, consequently, that it
has changed — unless a fortuitous meeting reveals the transformation that has occurred."

340 An Experimental Investigation of Perception

It would be an interesting question for future inquiry whether self-
activity of imagery is a necessary cause of subjective additions. In
some cases, of course, they are the same thing. Then it is the image
which is described, not the object.

(c) Objective changes in the pictures. Ten of the subjects reported
that there were inconsistencies in the successive percepts, as though
the picture itself changed during the course of the experiment. This
phenomenon occun-ed, of course, mostly in people of strong subjective
tendencies. The cause probably was that they would see a part of the
picture, interpret it immediately, and complete the unseen parts mentally.
In a later flash they would see a different part and find certain dis-
crepancies. Their experience was thus like seeing two different pictures.
Typical reports were :

" I am not quite sure it is the same picture."
" My notions are continually clashing."

" The train always appears more in the foreground of the picture than I expect it
to be."

" It seemed a different picture of another donkey."

It is interesting to note that the two scientist subjects who showed
such great similarity in their method of procedure belonged to different
classes in this respect : one spoke of objective changes several times in
each experiment, the other never mentioned them.

In a few cases the pictures seemed strikingly larger or smaller at
the end of the experiment than they appeared in the flashes. With
one subject each was definitely smaller at the end. With different
subjects this change occun-ed most often in the case of the train picture,
and many were surprised to find how small the engine really was.

{d) Subjective certainty. Subjects varied considerably in the certainty
and confidence they had of their decisions and achievements. It must
be remembered that subjective certainty is no guarantee of accuracy;
it is merely a mental attitude which may accompany sufficient or
wholly insufficient reasons, and may be found accompanying utter
inaccuracy.

Thus, one subject, who was very certain throughout, mistook the
bear for a man. I allowed the mistake to go on for some time before
I told her she had made a mistake — that it was not a man at all. She
was quite unshaken in her opinion : " But it's a human being. If it's
a woman she's in a man's get-up. You are fooling me." I added that
it wasn't even a human being, and she replied : " That upsets all my
ideas then. You haven't shaken my faith though." This certainty

Frank Smith 341

persisted in subsequent flashes : " It still looks like a human being."
"Still the*same exactly."

The certainty of some subjects, on the other hand, was obtained by
reasoned argument, and by deduction from observed phenomena : " I saw
three posts. They must be telegraph posts then, and not signals."

Some subjects showed a complete degree of uncertainty, which
sometimes persisted through the whole experiment. Those who were
uncertain accepted corrections readily, whereas the certain often refused
them, or only accepted them after a struggle. One who had seen the
bear as a man for twenty-one flashes was then told that it was not
a man. After the next flash he said :

" Well, then, it's a l)ear dancing. The position of the legs gave me that, and it
explains the shortness of the legs which I attributed to an apron."

This attitude of certainty or uncertainty did not seem to run parallel
with any other characteristic, though on the whole the certain subjects
made far better progress than the uncertain ones.

(e) Readiness of decision. Some subjects decided easily and at
once, however insufficiently they had seen the picture ; others remained
undecided, divided between alternatives for a long time^ The former
were, as a rule, the least careful in their examination of the object, and
wrong decisions did not affect their procedure in subsequent experi-
ments. Children belong almost uniformly to this class, as we shall see.
Many adults were quite uncritical as regards their own opinions. And
when such subjects did change their opinions they decided on a second
interpretation quite as readily as they had previously done on the first.

Sometimes a subject would begin his report in an undecided
attitude, but would achieve a final decision by the end of it :

"It might be a railway station. The left part is suspiciously like a station.
I saw people. Yes, I believe it's a railway station. Yes, it's a railway station."
(No. 2.)

The subjects who only arrived at their decisions after an interval of
some duration were generally very painstaking in their attempts to see
everything of importance. They carefully weighed alternatives and
examined possibilities. I quote from a case where, after deciding that the

1 "With a few subjects the answer once given must remain unchanged. Whether it
was right or wrong, whether a complicated figure was nearly round or angular — nothing
or exceedingly little could be said. Other subjects, on the contrary, were able after a still
longer time to criticise and correct the given answer, to find excrescences (Ausgelassenes),
and so on." Schulz, op. cit. 287-8.

342 An Experimental Investigation of Perception

bear was a man in the first flash, the subject began to suspect a mistake
after the eighth flash. His subsequent reports were :

(9) " I am beginning to suspect it isn't a man at all — it's a performing bear. It's
the way bears hold a staff. I was looking then to see about this."

(10) " I saw his legs. They seem to be short and thick for the body. He's
not a typically built man if he is one. The legs are a considerable distance
apart."

(11) " The whole suggestion is that of a bear. I didn't see any ears, but the
head and especially the shape of the figure suggest this."

(12) "I seemed to see the face, which is in profile. It is elongated and more
like an animal's."

(13) "I saw the other leg. It seems to be short and without calf."

(14) " I seem to see that the back of the head is hairy. But I'm diffident about
details of that kind. It's either an animal or a man dressed as an animal. From the
figure I should say an animal, and probably a bear."

In extreme cases there was complete indecision. This occurred both
with ignorance of the subject matter, and also with complete knowledge
of it. I quote briefly from one case — the actual introspective record is
very long — where the subject was very well acquainted with animals.
The picture was that of the donkey (No. 11):

(1) "A donkey."

(2) " More like a cow. Hind legs are those of a cow at any rate. Head in-
distinct."

(5) " I can't associate the neck with any beast."

(6) " It seems to be a cow's legs."

(7) " I should still say a cow, but I am undecided."

. (8) " Still a cow by the body. I was thinking of various breeds to test it by."

(9) " A cow by the head. The ears are not up high, but they are not drooping
like a donkey's. They may be horns."

(10) " It may be a bull from its head."

(11) " I think it's a bull from the neck — an arch on the top."

(12) " I doubt whether it's a bull."

(13) " The shape of the body makes me hesitate between a cow and an ass."

(14) " Distinctly more like a donkej' from the whole impression."

(16) " Nose and mouth look like a donkey."

(17) " By the nearest hind leg I should say it's a cow."

(18) " Fore legs are those of a donkey."

(19) " Donkey or pony from the nose."

This wavering went on for thirty-six exposures, during which the subject
made mention of bulls, horses, cows and donkeys. At the end he was
still a bit uncertain :

" I think a donkey by the way it stands. I would express my decision as five to
three in favour of a donkey."

Frank Smith 343

Curiously enough, indecision was never attributed to any fault of the
pictures; it was the accurate and decided subjects who made adverse
comments on them.

(/) Associations. Subjects in whom the various subjective ten-
dencies were strongly marked generally had so abundant a flow of
associations that they were unable to report them all :

" I just get a flash, and associations crowd in."

" Vague fleeting associations passed very quickly — figures in Pilgrim^s Progress —
one of a man with a whip trying to pull a donkey along, though I can't see its
connexion with the picture." (No. 6.)

"Associations so vague that I can't tell them to you."

Subjects in whom subjective factors had less sway, on the other hand,
had usually few associations.

It is interesting to note that associations very often ran along some
definite line, such as localities visited, or pictures seen, or events of
childhood. Thus, of seven associations detailed by one subject, five were
definitely of places. In another case, of nine associations six were
definitely remembered events of childhood.

Associations might be vague and general, or individual and vivid.
They were nearly always a source of pleasure. In a few cases self-
projection occurred with them :

" I felt myself actually showing the picture book (an association) to my niece."
(No. 5.)

'" This is a picture, but I was thinking of an actual bear dancing, not so much
of the picture here. I felt it to be ludicrous to watch a bear standing up like
that."

Like imagery, associations were sometimes a source of error. This
occurred when they were so strong that they replaced the actual percept.
I quote from a few cases of this kind :

" It was associated very vividly with a station in Devonshire. I think I may
have imagined the steps from that. The associated station took the place of this one
immediately." (No. 2.)

" I find it hard to distinguish between what is seen and what is associated."
" I saw a peculiar expression on the dog's face, and then some association inter-
pijgted it as contempt."

Associations had, to some extent, a reciprocal relation to inferences.
Those subjects who were able to proceed by inference did not usually
get many associations, whereas those whose thoughts were one long
train of associations rarely made inferences at all.

{g) Self -projection. In a few cases the pictures were seen as real
things, and the sqbjects projected themselves into the events and became

344 All Expermie}ital Investigation of Perception

actors instead of onlookers. I have already quoted at length (page 338)
from the most pronounced case of this kind, and add a few quotations
from other subjects :

" There's a kind of excitement about it that gives me pleasure."

" It reminded me of a square in Coj^enhagen. I was looking at it from my hotel

window. It was very pleasing."

" I am in the street but not frightened. He looks quite nice and tame, I am

prejiared to believe the master is near. I am some little distance away. He's not

looking in my direction, and doesn't see me." (No. 13.)

Self-projection occurred only rarely with adults, and with children
I think it is quite impossible to say whether it occurred at all, because
I am not sure that they understood such questions as " Did you feel as
though you were there ? " or " Did you seem to be in front of a real
thing ? " although they answered " Yes " or " No " to such questions
quite readily. Moreover it was easy to make them change their
answers. A few cases occurred where the projection was not quite
complete — thus, one subject reported of the station picture that
although she was not actually there she was near enough to feel the
excitement of the crowd.

In all cases of self- projection, of course, the picture was seen as
reality, but the pictures might also become real without the subject
projecting himself into it. I quote different cases:

'• I think on the whole it is a real train. Any picture would do that. But I am
cei-tainly not in it." (No. 1.)

" It isn't a picture. It's a real station — the Quai d'Oi-say." (No. 2.)

" It is a pure picture."

" I don't know whether a picture or real. I wasn't concerned with that question,
and it didn't arise."

This question has been investigated recently by Mr Aveling^, whose
results show that pictures were perceived by his subjects in three
different ways :

(a) As spots of colour — 37 times.

(6) As pictures (where the object of thought was the picture jo^ws

the thought that it pictured a thing) — 282 times. *

(c) As symbols (where the object of thought was a thing which
the picture represented) — 295 times.
Compared with my own observations, class (c) here seems to be
abnormally high in numbers, for I found such cases to be quite the

. 1 "The Relation of Thought-Process and Percept in Perception." This Journal,
IV. 211 ff.

Frank Smith 345

exception. However as these results were obtained from a very small
number of subjects they might be considerably modified if a larger
number were investigated. I agree with Mr Aveling when he says
(p. 227): "The same picture may be symbolic for one. observer and
asymbolic for another.... The explanation of this phenomenon would seem
to lie in the facility with which a symbolic subject assimilates, in the
perception, previous experiences of many years anterior date."

(h) Improbability. Some subjects were influenced by the improb-
ability of their answers :

" The whole idea was of a sweep carrying brooms. Then I thought of the light
blue suit, and so inhibited the idea." (No. 3.)

Others were not influenced in the least in this way. They might
mention the fact that what they were saying was improbable, but that
did not seem to them a sufficient reason for rejecting it.

(i) Influence of the first idea. The first idea was often of very
great importance in determining the subsequent course of the inter-
pretation, and frequently led subjects astray when it was wrong. Such
a subject, an ex-schoolmaster, on seeing picture No, 14, took it to be
some school children drilling. In the succeeding flashes he saw, under
the dominating influence of this idea, the regularity of the lines and the
division into squads. This illusion held during forty-four exposures
when I told the subject he was wrong. He had had no doubt about the
matter.

Other subjects seemed to lay no more importance on the first inter-
pretation than on any other. They might change completely at the
second flash, or even after a large number of flashes they showed
themselves just as open to modification as at the beginning.

In concluding this section on subjective factors it is not an easy task
to draw any general conclusions from such complex results. But the
five subjects who had what I have called a 'self-active' imagery
possessed many other characteristics in common. All of them made
frequent subjective additions. Four of them spoke very often of objec-
tive changes in the pictures. All had a strong flow of associations.
Four were fairly undecided. None made really good progress along one
definite line. All were drawn passively to outstanding features of the
objects, though they showed activity of intention in the fore-period.
Inference was weak in four. None was influenced by improbability,
and all seemed to receive a strong impetus from their first interpretation.

J. of Psych. VI 23

846 A7i Experimental Irwestigation of Perception

There was also one subject who showed many of the same characteristics
as the other five, but who did not make mention of any image changes.
Whether these took place and were overlooked, or whether they were
altogether absent, it is impossible to say, but his connexion with this
group seems a very close one.

Other factors appeared in very different arrangements among the
remaining subjects. Thus, those who analysed the pictures carefully
were generally those who made the best progress, but there are striking
exceptions to this statement. Other conclusions I have already indicated.

I did not investigate all the subjects with equal thoroughness, so
that among the thirty-two there were several of whom I can say very
little. But it seems justifiable to separate a small group with marked
subjective tendencies, and another class with a more objectively accurate
and exact attitude. A large number of subjects were hard to classify,
for they did not possess any strongly-marked characteristics.

(iii) The Perception of Colour. The experimental conditions under
which the investigation was carried out were not calculated to favour
accurate colour perception, but several interesting points appeared which
are worth mentioning.

In a few cases colour was commented on even before form, and
aesthetic judgments were then often made :

" That's in colours."

" I found the attractiveness of the colour, especially of the sky, almost a dis-
turbance at the moment of recognition." (No. 1.)
" I foimd the strong light and shade very confusing."
" A cheap process."
" The sky is a horrid unpleasant green colour." (No. 1.)

Other subjects showed a marked indifference to colours, and only
mentioned them fairly late in the experiment, after other matters had
been investigated and settled. One subject correctly saw green in the
engine at the first exposure, but made no mention of the blue sky, which
was very bright and prominent, till the fifty-eighth exposure. Another
missed a very prominent bluish-green colour in picture No. 7 till the
forty-fourth exposure. I then told him to look in a certain direction,
and he at once saw the colour.

When I used the shortest exposure, about one-hundred-and-thirtieth
of a second, colours tended to disappear altogether with the majority of
subjects, except where the colours were very highly saturated, or where
the subjects were interested in and attracted by them. Thus, if I began
an experiment with a coloured slide at the shortest exposure, I often

Frank Smith 347

had the report "Not coloured." This occurred ahiiost always with
pictures 2, 3 and 4. In the case of picture No. 2, I showed it to ten
subjects at this quickest speed, and eight of the ten reported that there
was no colour in it. As I slowly lengthened the time of exposure this
judgment generally remained unaltered for a long time; sometimes,
indeed, colour was not mentioned till the picture was permanently
exposed. Occasionally, a colour seemed to be just on the margin of
perception. Thus, in picture No. 3, at the normal times of exposure,
a subject saw blue in the third flash, but not again till the fourteenth,
when he specially looked for it.

Fatigue had great influence on colour perception. This was well
shown by one subject with whom I was working fairly late at night.
He failed to see any colour in No. 3 at normal exposure time, though
I asked him about it. As he was tired we did not finish the picture
that night, and on continuing it next morning he saw at once that it
was coloured.

Although I did not use any quicker flash than one lasting one-
hundred-and-thirtieth of a second, yet even at that speed some subjects
continued to see almost as much in the first flash as they had previously
seen with the normal time of exposure. On the other hand, I had one
subject where the time threshold for colour perception seemed to be
exceptionally high.

(iv) Feeling Tone. These experiments were not very favourable
to a satisfactory investigation of feeling tone. As Kuhlmann says :
" Pictures as such have not the interest and emotional colouring that
belong to objects. Such emotional colouring, when it takes possession
of consciousness, brings with it the visual imagery with which it is
connected. Again, in so far as meaning and interpretation is read into
the picture at all, so far, of course, the picture ceases to be what it is
and becomes the object. The picture to this extent creates the tendency
to substitute the imagery of the object in its natural setting\" One
subject said : " Feelings are an after thought with me. I have to recall
the picture."

At the outset the shortness of the flash was a disturbing factor,
because subjects felt somewhat hurried, and their chief attention was
given to the subject-matter of the pictures. Owing to the pre-eminence
of this claim the emotional tone was apt to be ignored by the less
practised observers. The shortness of the flash irritated quite a number
of subjects :

1 F. Kuhlmann, op. cit. 415-6.

23—2

348 An Experimental Investigation q/ Percej^tion

" I have a feeling of discomfort because it is so short."

" I am aggravated because the picture disappears so quickly."

" The instantaneous exposure is aggravating."

One or two subjects, on the other hand, were stimulated :

" Speed gives piquancy to it."

" The shortness of exposure does not irritate me. It arouses my curiosity."

Subjects who experienced some difficulty in solving the pictures
often developed disturbing feelings of self which rather added to their
inability :

" I am beginning to feel irritated at my inability to decipher it."
" I have a nascent feeling of shame that I have to make two great corrections."
(The corrections, however, were not 'great.')

B. Experiments on Children.

I now pass to an examination of the results obtained with children.
These were kept separate from the rest, so that the two groups might
be the more easily contrasted and compared.

The three schools were chosen because of their typical differences,
and at each I worked with a group of children of the average age of
twelve years, and also at the two elementary schools with a group whose
average age was six years. Unfortunately I was unable to get children
of this age in the Secondary School.

Stem has worked out with some completeness the mental develop-
ment of children, and the modifications due to age and sex\ but my
experiments are much less complete as I was working at a more general
problem. Hence I confined myself to two ages in order to see generally
in what direction changes occur with the advance of age. Moreover,
I did not use many girls in these experiments. But it is possible that
this same method of investigation might throw light on the question of
mental development in school children, if it were applied uniformly to
groups of both sexes at many different ages.

(i) Six years of age.

I will deal first with the results of the youngest children.

^ " The time from seven to ten years is for boys a period of strong development, whilst
girls during the same period virtually remain at a standstill and their reliability even
recedes; between ten and fourteen years girls, by rapid development, recover the lost
ground and even overtake the boys a little, whereas the latter show no improvement.
Therefore the difference between boys and girls is greatest at the age of ten." Stern,
"Die Aussage alsgeistige Leistung and als Verhorsprodukt," Beitr. z. Psychol, d. Aussage,
Erste Folge, 291-2.

Frank Smith 849

(a) An elementary school in a very poor district.

In this school, where the children were from the poorest homes, and
of parents of a very inferior type, I worked with four normal children,
two boys and two girls, whose average age was a little over six years.

The most outstanding feature of the boys' results was a very pro-
nounced tendency to add imaginative detail. Nothing seemed too
grotesque or too impossible to apply to any picture. This tendency
was not so marked in the girls, but in the case of the boys there seemed
to be no ability to distinguish between what was accurate and what
was inaccurate, between what was seen, and what was imagined. I will
quote from one case :

(1) " A house and windows and door. A little boy peeping out of the window."
(Picture No. 1.)

(2) " A house and a bedroom."

(3) " A little boy outside."

(4) " The little boy is hiding. I saw him peeping."

This continued until the eleventh exposure, when I told him it was not
a house. He continued :

(12) " It was a house."

(13) " A train. And a woman in it peeping and waving her hand."

In the Queens' College picture this same subject saw a little boy,
and described his " hat and blue sailor coat." When I suggested there
was also a woman he agreed, and added that she was waving her hand
to the little boy, and he was waving his hand to her. At the end
I showed him the picture, and asked where the woman was. He replied
very confidently : " She's in her house."

Many of the boy's replies seemed to be faint memories of past ex-
periences. Thus, he described the bear as a " woman," and added the
curious description that she was deaf and dumb. When I asked him
the reason for this, he replied that "she carried a pole." On further
questioning, I found that he had once seen a deaf and dumb woman
with a long stick, and undoubtedly the idea had come from this recollec-
tion. And again, his first answer in the station picture was : " Little
boy. Green hill. Stick in his hand." All these three phrases had
been used by him previously in other pictures and undoubtedly came
back to him from them.

The other boy showed the same tendency. In the picture of the
child (No. 4), he described shoes, stockings and coat quite readily. When
I showed the picture at the end, and asked him to point to these things,
he touched the bottom of the glass screen vaguely, and said they were

350 An Uxperimental Investigation of Perception

" down there." In the same picture he said there was " green grass,
and in the bear picture he spoke of a " train." These answers, too, were
undoubtedly due to the other pictures. Some of the additions cannot
be explained in this way. Thus, when he saw a house in the bear
picture he added quite spontaneously : " There's a house with windows,
where the little boy lives, and he's got a father and a mother and a baby
and a cow."

Sometimes these boys developed a kind of ' additive ' description of
some object, which consisted of adding together all the possible qualities
of such an object, utterly regardless of objective accuracy. Thus, in
picture No. 6, this second boy described the central figure, after an
exposure of one second, in the following way : " A man. Red. Green
hat. Black shoes. Black trousers. Red scarf Blue waistcoat. Blue
eyes. White ears. White nose. White mouth."

The two girls were much more exact and reliable than the boys in
describing what they had seen and in repeating what they had said,
though one of them showed a passive tendency to repeat former sayings
regardless of their meaning. Thus, in the child picture, she first
described it as a " man " and gave several qualities, including "whiskers."
When she discovered it was a child, she corrected her mistake, but
added the other features, including "whiskers," in full, and only with-
drew the impossible ones when I drew her attention to them.

It would be interesting to discover the cause of this indifference to
a strict line between reality and imagination. It may be that both
boys had a ' self-active ' imagery, but this explanation would not cover
all the answers. I think, rather, that the problem is one of environ-
ment. The mistress of the school agrees that the peculiarity I have
described is far more common among slum children than among those
fi'om better class homes. She suggested that the explanation is that
these children have to give answers at home which will please their
parents. The strict truth of the answer is not an immediate concern.
If the answer displeases, the child is punished. Hence with a stranger
he is willing and even anxious to see what he thinks he ought to see,
and this brings about an attitude favourable to ' creative imagination.'

This explanation also accounts for the very marked susceptibility of
these children to accept suggestions quite readily, and even to add
descriptive details to the objects suggested. The two boys and one girl
did not show the least inclination towards caution, but caught up each
suggestion I gave immediately and fully. When I suggested a dog in
the station picture, one boy said he had seen it and that it was barking.

Frank Smith 351

I asked one of the girls if she had seen a swan in the Queens' College
picture, but she did not know what the word meant. I also told her
there was something going right across the picture (meaning the bridge),
which she had not yet seen. After the next flash she showed how very
passively suggestible she was by saying : " It's a swan up there."

The other girl was more independent and critical. She refused
suggestions often, and only accepted them when there was some justifi-
cation of their probability, and at the end she was careful to correct her
mistakes.

These subjects had no problems and showed no indecision. In the
donkey picture one boy mentioned six animals successively, and each
was given with an air of confidence and finality. There was never any
postponement of decision, no matter how incompletely the picture had
been seen. Each judgment, when made, was final for the time being,
though if it had to be abandoned another would be offered with the
same finality. One of the boys interpreted picture No. 12 as, in turn,
a mountain, a ladder, a motor car, a little boy, a chair, a tram, a bus, and
a picture. There was no hesitation about each judgment, no balancing
of alternatives, no consciousness of the need of deliberation.

The pictures were looked at as a whole, and there was no analysis \
Practice did not seem to bring about any modification of method, or to
develop any caution. Inferences were quite absent. Associations were
rare, though this was partly due to the inability of the children to
recognize them as such, and to their neglect in reporting them.

As was to be expected fi-om the experimental conditions, the
perception of colour was very difficult, as was also the exact localisa-
tion of the colours seen for so short a period. This accounts in some
measure for the many mistakes made. The child was described as
having a " blue coat and black hair," whereas the blue was round the
child, and the hair was brown ; the train was said to have " grey "
carriages, whereas in reality they were brown.

So far as it is safe to distinguish between the sexes in such a
small number of cases,, the girls showed some superiority over the
boys in several ways. Thus, if we take the average number of facts
correctly observed in each picture and divide the two boys from the

^ As Binet says, "What is lacking in the child is that he does not possess to the same
degree as ourselves the power of analysis." " Perceptions d'Enfants," Rev. philos. xxx.
592. In the same article Binet relates that a child of one year and nine months was able
to comprehend the drawing of a horse, but even at four years was still unable to under-
stand the drawing of a nose, or of an eye.

352 An ExpenmenUal Investigation of Perception

two girls, we find that while the boys only observed 7*1 facts in each
picture, the girls observed 10"9, and that whilst the girls made only
8"0 mistakes in each picture, the boys averaged 9*6. In describing
correctly the positions of objects, the girls also had the advantage,
the figures being 2"1 as against 1'5.

(b) An elementary school in a good district.

At the other elementary school, where the children were of a higher
social status and from more cultured homes, I had five boy subjects
whose average age was six years.

Here there was very little tendency to add imaginative detail. It
appeared occasionally, as when one boy called the donkey a "horse" and
then added that there was a cart fastened to it; and also when he
said he saw telegraph wires on both sides of the railway lines. But
the first of these additions was afterwards corrected quite voluntarily.

Suggestions, too, were not accepted so readily and were often
resisted quite firmly. In the case where a mistake had been made
(such as calling the bear a man), these children, when told they were
wrong, did not make their correction immediately and rashly as had done
the others : they looked again. One boy, when I told him he had made
a mistake, looked a few more times and then said : " I don't know.
I don't think I shall get it."

In many particulars there was close agreement, of course, between
the children of the two schools. The method of solution was the same :
that is, the whole picture was looked at each time, and no analysis was
attempted. Decisions were made at once, and the siibjects were soon
satisfied without making very close examination. Alternatives were
not expressed, and a judgment made with confidence at one moment
was often cast aside almost immediately for a contrary one, which would
be given with equal certainty.

But the superiority of this group over the previous one was marked.
Two children made the inference that as there was a shadow the sun must
be shining, and another said he saw a steeple, and so argued a church
from it. Reasons were often given for statements : thus, in the station
picture one child naively argued there was a bride, because he saw a
" lady with a veil and baby " ! He afterwards corrected himself and said
she could not be a bride because she had a bonnet on and had no roses.
To another boy the donkey was a cow " because it had horns," then a
horse " because it had reins." Another saw what he thought were
chimneys at the bottom of the bear picture "as though the picture

Frank Smith 353

was upside down." This tendency to give a reason for a statement,
even if the reason is a wrong one, shows a much higher mental process
than do the unreasonable answers given by the first group.

The language was very similar in the two cases, though at the
second school more verbs were used, such as " curves " and " walks " ;
and longer sentences were formed.

Some of the subjects had a great power of seeing very small things :
thus, the engine driver, who can just be distinguished looking out of
the engine, was first noticed by one of these children. Colour perception
was again rather poor. One of the five refused for some time to change
his opinion when I told him he was wrong. Another told me before
one of the exposures that he had made up his mind where to look;
this, however, only occurred once. The complex picture No. 6 was
thus interpreted by one subject : " He's dabbing the brush in her face
on purpose, and the old man is telling him not to do it." Associations
were often mentioned and appeared to be much more frequent than at
the first school.

The average number of facts correctly observed by each subject was
10*5, and the number of mistakes 6'15, The average number of positions
given correctly was 3*35. Hence these boys showed marked superiority
over the two boys, and are slightly better than the girls of the other
school.

We may conclude generally that this group showed much more
reliability and activity than the other. There was a more critical
attitude, a greater accuracy and much less sway of subjective ten-
dencies.

(ii) Twelve years of age.

I now pass to the older children, a group of whom came fi'om each
of the two schools just described, and a third group from the —
Grammar School ^

(a) An elementary school in a very poor district.

In the first elementary school, with children of the poorest class,
I had seven subjects — four boys and three girls — whose average age
was twelve years and four months.

There was greater evidence of activity than among the younger
children at the same school. This was shown by the occasional refusal

' This school is well known for its adoption of modem methods of teaching, and for
its educational experiments. A spirit of independence among the scholars is more marked
than at other schools.

354 Ati Uxj)erimental hivestigation of PercejUiou

to accept the suggestions I made. Thus, one boy saw something white
coming under the bridge in the Queens' College picture, but my sug-
gestion that it was a swan did not influence him, and he was content
to leave it with a non-committal " I can't tell." Another boy showed
quite commendable caution in the Brittany picture : he agreed it was
" something like a window," but he refused to go any further without
more evidence. Occasionally, some of the subjects decided what to
look for in the fore-period, though this was exceptional. A few
inferences were made, too, and at the end some of the subjects
enumerated their chief omissions, and corrected the most obvious of
their mistakes.

But on the whole these subjects must be described as largely
passive. The first idea that came to them seemed to determine very
largely their subsequent answers. Thus, in picture No. 6, the male
figures, which were seen as boys in the first flash by one subject,
remained boys throughout. Many more suggestions were accepted
than were refused. The following answers to three successive flashes
of picture No. 2 show how completely susceptible this subject was to
every suggestion :

(8) Have you seen the dog ? " Yes, near the little boy on the right."
What colour is it ? "I didn't quite see."

(9) " The dog looked brown, with two patches of white

on its front feet."
Is it like any dog you know ? " No."

(10) " It seemed like Mr Brown's dog, with a short

tail."

There was a clear case in this gi-oup of ' self-active ' imagery.
In the picture of the train he made the following reports (I omit
details) :

(1) "A little boy crossing the lines at the left side, waiting till the engine passes.
He's waving his hand."

(2) " It looked as if the stoker had just made up the fire — smoke was coming out
of the chimney on the side were the boy was. The boy was waiting on one side, and
the engine was nearly past him."

(3) " It looked as though people were popping their heads out of the window and
waving their hands at the little boy."

(6) " The engine seems past the little boy now, and he's just going to step across
the lines."

Objective change was present in several of the subjects, as the
following answers indicate :

" This is a different boy." (No. 5.)

"The train is nearer and bigger every time." (No. 1.)

Frank Smith 355

Colour perception was still poor and inaccurate. Four of the seven
did not mention the blue circle behind the child's head, though the
colour was very striking : they may have seen it, but it evidently
did not seem of importance to them. There was also a vagueness
about colour answers which indicated inaccurate localisation. Thus,
in the train picture, one said there were " red, blue and white flowers
on the floor," whereas the stones on the left were a vivid reddish-
brown, and the hedges and ways were a bright green. Many said the
child had " dark " hair and a " blue " coat, as did the younger children
at this school.

If we compare the girls and boys in the group, the former show
a decided superiority in several ways, a result which is in agreement
with many other observers. Thus the average number of correct facts
discovered in each picture was 12*58 in the case of the boys, and 14*5 in
the case of the girls, whilst the latter olily averaged 3*58 mistakes, and
the former made 7*0 mistakes.

The girls gave more correct reasons for their statements than did
the boys. They also seemed better able to discover the explanation
of the complex pictures. Thus, of the station scene, one girl said :
" The man looks like the woman's son. He may be going away." This
is the correct explanation, and was only given very rarely even by adult
subjects. Picture No. 6 was also * explained ' very creditably by the
girls, whereas the boys rarely solved or attempted to solve these
complex puzzles.

Girls had more associations than boys, and made better progress in
getting out the details in succeeding flashes. They also showed greater
power of analysis. The number of subjects is too small to warrant any
certain generalisations being drawn from these differences between the
two sexes, although the distinctions were fairly consistent in the cases
observed.

(6) An elementary school in a good district.

At the second elementary school I had six boys as subjects with an
average age of eleven years. This lower average was due to the fact
that three of the boys were only about ten years of age.

Mental activity was more generally noticeable than in the group
just mentioned at the first school, and was shown in several different
ways. Thus one boy, the youngest of the group, constantly worked
down to causes :

" His hair is rather rough, as though he is just up." (No. 5.)

" His lips are only a pale red, as though not completely healthy." (No. 5.)

356 A7i Experimental Investigation of Perception

Another subject said of the same picture :

" He is closing his right eye, as though somebody has thrown something at him,
or is going to hit him."

Probable things, not yet seen, were specially sought. Thus, in the picture
of the donkey, one subject said :

" There ought to be some sky. I haven't noticed any,"

After the next exposure he added :

" I couldn't see any sky. The cow seems to be on a hill, and the ground rises up
behind."

Another subject noticed that he had not seen a signal in the train
picture, and so looked for one.

Two subjects, on the other hand, were amazingly uninquisitive.
They were satisfied after a very, few flashes that they had seen every-
thing, and would keep saying at the end of subsequent exposures, as
though tired of the picture : " That's all." These subjects accepted
suggestions as readily as any in the previous group.

Introspection was more exact at this school than at the first, as the
following examples show :

" I think I noticed the colour first. I saw it was white. Then I thought a white
cow." (No. 11.)

" It reminded me of going from Cambridge to Yarmouth. I felt in the train but
it was still. I also thought of my brother and sister in the train." (No. 1.)

Colour perception was again weak. Two missed the colour altogether
in picture No. 5, and all subjects made mistakes.

There was an excellent example at this school of a subject showing
his predisposition to one special aspect of the pictures. In this case
it was to geometrical notions and concepts, and the boy compared
lengths, proportions, sizes and angles in almost every picture :

" It did not seem to have long legs compared with the body." (No. 13.)

" The sides of the road seemed to vanish quickly." (No. 13.)

" The elephant seeined about one and a half times as high as the man."

When I asked the headmaster to place the boys in the order of
their ability in mathematics, he gave this subject the top place, without
knowing anything of these results.

There were no cases of self-active imagery, and objective change
was reported by one subject only :

" I have a different idea." (No. 13.)
" I get a different view." (No. 1.)

Frank Smith 357

The average number of correct facts discovered in each picture was
seventeen in the case of the three older boys, but if we take all the
boys together the average was only 14" 2. These results show a signifi-
cant superiority over the average of 12'58 gained by the older boys at
the first school. (Unfortunately I had no girl subjects in the second
school.) The average number of mistakes made in each picture was 5*65
for all the six subjects, this again comparing favourably with the other
group. More correct reasons for statements were discovered, and more
exact positions of objects were stated by the second group than by
the first ; also, more associations were mentioned.

Besides this general superiority which is evident in the second
school, there is also evidence of more individual variation, and
generalisations are more difficult to make. Some subjects showed a
fairly independent active attitude, as I have shown, whereas others
were amazingly passive.

In this group, too, the subjects had the benefit of a wider ex-
perience than in the other, and seemed better able to deal with the
strange and unusual. Although • the school is much further away
from Queens' College, two of the four to whom the picture of the
College was exposed recognised it at once, a third knew it was a
Cambridge College but could not name it, and the fourth recognised
it correctly at the end.

(c) A Grammar School.

At the — Grammar School I examined nine boys, whose average
age was twelve years and two months (varying from eleven years and
ten months to twelve years and six months).

Two striking differences from the results obtained in other schools
may be mentioned at once. The first was the caution used in describing
the nonsense picture (No. 16). From the children in other schools, as
indeed from many adults, I got many very elaborate and extravagant
interpretations, as the following :

" Some ducks walking about on land."

" A picture of St George and the Dragon."

" A boy on a donkey's back, and some more children standing at the side."

And to these interpretations further details were often subsequently
added. But at the — School the boys gave a very meagre and re-
stricted interpretation to the picture. The following answers were
returned by different boys, all after the first flash :

" A whole lot of jumbled figures with a box in the middle. Done in blue ink. All
dancing round a box. Little ink smudges."

358 An Experimental Investigation of Perception

" A pattern, isn't it ? Blue colour."

" Like a landscape puzzle. Green colour."

" I don't know what it is. Blue things. Might be Japanese writing. It goes in
all directions."

" A lot of blots. It reminded me of when I put ink on a lantern slide and showed
it on a sheet."

" Black figures on a white background."

These answers show great objective accuracy, and the ability to keep
very close to the bare percept — an attitude which is the antithesis
of the tendency to subjective addition so clearly marked at the first
elementary school.

The second difference was the very systematic and careful examina-
tion of each picture at the end of the experiment. Corrections were
made with great accuracy, and omissions were noted, the subjects
showing a very general absence of bias and prejudice. I quote two
examples :

" Oh ! it's a bear performing in the street." (The subject had previously thought it
was a man.) " There are houses and shadows — not a sledge. A chain, nuizzle, trees
either side, big buildings, some very tiny vehicle at the end of the street. The
white thing on the bear is a muzzle. The head is shagg}' on the top. There's a
chain attached to the pole, and I believe a bit of chain attached to the paw. There's
a little boy in a white jersey and black trousers looking at the bear ; he's leaning
against the wall of the house nearest to me on the right. The trees extend the
whole way up."

The second is the picture of Queens' College :

" There are no swans. It's a College I think." (This subject was not a Cambridge
boy.) " At the far end are two boats and men rowing. On the left a low sort of
wall. A clump of trees. Stone buttress pointed at the top. Behind the bridge
and on the right a part of the building sticks out. A pointed roof. Three chimneys.
Windows deep set and round. Buttress supports bridge. Brick house. Two
rows of windows. At the end of the picture there are two windows at the top, then
three windows, then a division, then three windows. They are square and have
bars on them. I can see the roof of the part with round windows. It is black and
has garrets."

Children at other schools generally ignored the mistakes they had
made, and some would stare at a picture, exposed to them at the end of
the experiment, without adding a single correction.

Suggestions were most often refused quite decisively, though they were
occasionally accepted when there was some ground for their probability.

Several boys looked out specially for objects which were likely to
be found, but which they had not yet seen, and chose special parts of
the pictures for examination :

Frank Smith 359

" I haven't seen the tender yet." (No. 1.)
"There ought to be some posts nearer me." (No. 1.)

" I am going to see if there are houses on the same side as the pavement."
(No. 13.)

" I was going to see what the woman was sitting on." (No. 12.)

One boy is to be distinguished from the rest as showing much less
care and caution. He had been attending the school for a short time
only, though this, of course, may only be a coincidence. He did not
correct his mistakes at the end. He went far from the actual percept.
He said there were "about six people" in No. 6, and "a cat" between the
man and woman. In the bear picture he saw " a crowd of people on the
right." He made generalisations on very insufficient data. In No. 6
he said the figures were "dressed in bluish stuff." In the station
picture there were "no porters or officers about." However, he made
inferences with fair ease, and of the Queens' College picture he said
very shrewdly : " The trees form a very deep grove, and make the
place very dark. The windows are not much good because the light
doesn't get to them very well. The river is not very wide."

Complex pictures were not very well solved as a rule, and it was
evident that the main interest was not in the hidden story of a picture,
but in the observable details it contains. One boy was remarkably
clever in discovering the important features of a picture at once. He
seemed to be able to ignore instinctively what was unimportant, and to
concentrate on the significant parts.

Language was more complex and varied than among the other groups,
and expressed finer shades of meaning. There was one case where a
subject added so many qualifications to his answer that I could not
write them down quickly enough.

The average number of correct facts discovered by these subjects
in each picture amounted to 19'4, and the mistakes numbered
3"15. These numbers show a considerable superiority in discovering
detail and in caution possessed by this group as compared with the
others. Associations, too, show a marked increase, evidence, it may be,
,of a fuller experience. The ability to state the exact positions of
objects and to give the correct reason for things was also possessed in
a higher degree at this third school than at the others.

Objective change was remarked upon by one subject, but subjective
addition was quite absent. The power of analysis was present to a very
much more marked degree than in the other groups.

To sum up, we may say that the children in this group showed
evidence of remarkably advanced mental powers in dealing with the

360 An Experimental Investigation of Perception

problems before them. They were systematic in the fore-period, and
careful during the experiment. They showed caution against subjective
errors, and independence of judgment. They based their answers on
good reasons and did not lightly dismiss alternative possibilities without
good cause. In all these respects they approached the characteristics
of adults, and showed a very pronounced superiority over the other two
groups of similar age.

This fact of variation from school to school is of importance in
psychological investigation, and offers possibilities of further research.
It is easy to see that the second of the three schools occupies a position
midway between the other two. At the first school the older children
did not seem to have advanced very far, in some respects, over the
younger children at the same school.

The figures which I have already given may be set out in tabular
form for greater ease of comparison :

Av. no. of

Av. no. of

Av. no. of

Av. no. of

School

No. of
subjects

Age

correct
facts in
a picture

mistakes
in each
picture

correct

reasons

for things

positions
correctly
described

1. Elementary
(poor district)

2 girls
2 boys

6
6

10-9
7-1

8-0
9-6

0
0

21
1-5

2. Elementary
(good district)

5 boys

6

10-5

6-15

0-15

3-35

3. Elementary
(poor district)

3 boys
3 girls

12
12

12-58
14-5

7-0
3-58

0-25
0-4

4-75
3-8

4. Elementary
(good district)

3 boys
3 boys

10
12

11-4
17-0

3-8
7-5

01
0-58

4-2
5-25

5. Grammar

9 boys

12

19-4

3-15

0-4

7-6

(iii) A mentally defective child.

An interesting case presented itself at one of the schools, where I
had a mentally defective child as a subject. This was a boy of eight-*
and-a-half years of age. He was inferior to children much younger
that himself, and was very open to suggestion, but his chief fault was
his inconsistency. He showed no power of making gradual progress
over an extended period, no ability to solve a problem by degrees.
Lack of memory largely accounted for this weakness, though not
entirely, I think. His attention would wander off in a distressing way.
When left to himself he merely gave enumerations of the things he had

Frank 8mith 361

seen and imagined. Thus, in a picture of two bulldogs (not included
in the descriptions given earlier), I did not interfere at all with him,
and his answers were :

1.

" Two dogs."

12.

" Sim " (imaginaiy).

22.

"River" (imaginary).

2.

"Twodogg."

13.

" Moon " (imaginary).

27.

" Dog's nose."

3.

" Sitting down."

15.

" Ears."

28.

" Other dog's nose."

5.

"Dog's legs."

17.

" Head."

31.

" Pavement."

7.

"Tail at the back"
(imaginary).

20.

" Sky " (imaginary).

32.

" Road " (imaginary).

0.

" Road " (imaginary).

21.

" Star " (imaginary).

33.

" Sea " (imaginary).

It will be noticed how one answer often suggests the next (12 and
13; 20 and 21; etc.). Others are probably vague memories of previous
pictures : the " river " from Queens' College, the " road " from the bear
picture, and so on.

When I asked this subject questions in order to make the reports
more explicit, I was soon in a maze of contradictions and inconsistencies,
for he would accept even the slightest suggestion and then promptly
forget it in favour of something else.

This research is incomplete at many points, but the problem is so
vast, and involves so many factors, that this fault could not be avoided
in the time at my disposal.

In conclusion, I should like to express my very great indebtedness
to Dr Myers, who first suggested the experiments to me, and who has
helped me unceasingly during their progress. My thanks are likewise
due to Dr Rivers and to Mr E. Bullough for much valuable criticism
and advice ; to Dr Rouse, Miss Wing, Mr Hajmes, Mr Inglis and
Mr Mullett for their courteous help in the school experiments ; and
also to my numerous subjects, who often visited the Laboratory at great
personal inconvenience.

5. Conclusions.

The process of perception begins with an immediate interpretation
of the object, and this is strikingly uniform in different subjects. It is
followed by an analysis of the object, and, in this, very marked individual
differences emerge.

Among adults, persons of scientific training show advantages in the
method of analysis, and are generally better able to direct the process
systematically and carry the analysis further than are other subjects,
though there are exceptions to this statement.

J. of Psych. VI 24

362 An Ex^jei'imental Investigation of Perception

Some persons with strongly marked subjective tendencies behave
very consistently in many respects, and it is probable that the
cause of their similarity is to be found in a ' self-active ' imagery,
which plays a more important part than the actual percept. There
are also subjects who show very little tendency to be influenced
by subjective factors, and with whom the image has little or no
importance.

Among children, perception becomes more complex as age increases.
Children of six years of age have no power of analysis, and see the
object as a whole. They cannot balance alternatives or postpone
decisions. They are very passive to suggestions. Children of twelve
years of age show a greater power of analysis, a more active mental
attitude, and an improvement in the discovery of detail. Compared
with the younger children there are gi'eater individual differences.

Children of the same age from different schools vary in a marked
way. At a secondary school children of the age of twelve have already
many of the powers and methods of adults. At a 'slum' school they are
much inferior and even show some of the characteristics of the youngest.
Mental activity is as marked at the first type of school as passivity is at
the second. An elementary school of a good type shows characteristics
midway between the other two.

THE COLOUR PERCEPTION AND COLOUR PRE-
FERENCES OF AN INFANT DURING ITS
FOURTH AND EIGHTH MONTHS^

By C. W. valentine.

/. Purpose of experiments.

II. Results of previous investigations.

III. A new method of experim,ent.

IV. ResvZts of experiments.
V. Discussion of results.

(a) The presence of colour sensations,
(h) Colour preferences.

VI. Experiments hy the 'grasping method.'

VII. The method of grasp and reward.
VIII. The use of the hands in grasping.

IX. Summary of results and conclusions.
Appendix. Determinations of the brightness values of colours used.

I. Purpose of Experiments.

The experiments here described were begun upon my little boy
when he was just over three months old. I wished to discover what
colours he preferred at that age, before their 'natural' effect could
have been materially affected by any disturbing associations. In the
course of the experiments I was also on the look-out for evidence as
to the development of the colour sense.

II. Results of Previous Investigations.

The most important recent investigations known to the writer are
those of Baldwin^ Miss Shinn^ McDougall*, and Myers*.

^ An abstract of this paper was read at the meeting of the British Association for the
Advancement of Science, Birmingham, Sept. 1913.

2 Mental Development in the Child and the Race, Chap. in.

3 The Development of the Senses in the First Three Years of Childhood, 148.

* "An investigation of the colour sense of two infants," This Journal, ii. 338.
^ "Some observations on the development of the colour sense," Ibid. 353.

24—2

364 The Colour Perception of an Infant

Most of the experiments referred to by these investigators were
performed upon infants between the ages of six and twelve months, and
with the exception of Baldwin, the writers show considerable agreement
as regards the colours preferred. Myers found that light grey was pre-
ferred to dark grey, but that in spite of this partiality for brightness,
yellow was preferred to white. McDougall, whose investigations were
the most extensive, concludes that, in the sixth month, red, green and
blue were all preferred by one of his children, to greys of the same
brightness and even to white, but that no one of these three colours,
red, green and blue, was markedly preferred to the others. His figures,
however, suggest that red tended to be preferred to blue by one child
during the sixth months and that yellow was slightly preferred to red
by another child of seven months ^

Miss Shinn, reviewing the evidence of several investigators, concludes
that during the second half-year of life, red, yellow and orange are
the most attractive colours, but she thinks that white is even more
attractive ^

Baldwin's results are somewhat different. They suggest at first
sight that blue is slightly preferred to red by his child. But Baldwin
himself shows that his child was less inclined to grasp the coloured
papers the further they were from him. Now red was placed 22 times
at one of the far distances used by Baldwin — 13, 14 or 15 inches — and
blue only 16 times ; while at the shorter distances of 7, 10 and 11 inches
red was placed 8 times and blue 10 times. Thus the placing was obviously
in favour of blue. Yet Baldwin masses all the results together for the
purpose of calculating percentages. This conceals the fact that, at the
long distances, red fares distinctly better than blue.

Thus Baldwin's results suggest after all that red was at least as

1 The figures are : red chosen 82 tiroes, versus white chosen 33 times, when the two
were presented together, the scores for blue versus white being 17 to 13. McDougall
mentions a possible source of error, op. cit. 341.

2 Op. cit. 345. If the scores for and against yellow and red given in Table III are
added they give : for yellow 32, against 12 ; whereas the figures for red are : for 42,
against 31. The scores when red and green were presented together are omitted, as yellow
was not presented with green.

In considering the long series of experiments summarised in Table II (p. 343), one
must take into account the scores against each colour as well as for, owing to the fact that
they were not all presented the same number of times. The total scores are, red : for 128,
against 67; green: for 107, against 86; blue: for 114, against 88; white: for 96, against
118; grey: for 31, against 117. This shows an appreciably higher proportionate score
for red.

3 Op. cit. 158.

0. W. Valentine S65

attractive as blue to his child. On the whole, then, the results of these
investigators indicate that red and yellow are the colours best liked by
infants of this age (as far as this can be judged from the examination of
five children), though Miss Shinn concludes that white was liked even
more than yellow and red.

III. A New Method of Experiment.

I wished to make a test as to the colour preferences of my child, W.
at an age considerably younger than that of the infants tested by
previous investigators, viz. at three months. At this age, however,
grasping had not developed sufficiently to make Baldwin's method, or
the modified methods of McDougall or Myers, possible. But I convinced
myself, by some preliminary experiments, that another method was
capable of giving fairly reliable results.

Briefly, the method was to measure the time during which W. looked
at either of two coloured wools held before him for two minutes at
a time.

The wools were selected from Holmgren's wools for testing colour-
blindness. The infant was placed in a comfortable position among
cushions in an arm chair. The chair was placed near the window on
dull days, but near the middle of the room when there was strong
sunlight. The direct rays of the sun never fell on any of the colours.
Greater constancy of illumination could have been obtained by means
of artificial light, but the variation was but slight, and this very varia-
tion provided interesting evidence on at least one point to be mentioned
later. I sat on a chair in front of the child and held two of the coloured
wools about one foot from his face, in such a position that they would
appear against the dark grey background of my coat. The wools were
held quite close together for a few moments ; then I slowly drew them
apart until they were some eight inches distant from one another and
equidistant from the central line of vision of W. They were then held
motionless. As soon as the child looked at either of the colours, or if
he followed one of them with his eyes, I called out its name to my
assistant, who was provided with a stop watch and with a record sheet,
three vertical columns on which were headed thus (supposing the
colours to be red and blue):

Red Off Blue

The stop watch was kept going continuously. When I called out the
name of the colour looked at, my assistant noted the exact second

366 The Colour Perception of an Infant

indicated by the watch, and put it down under the name of the colour
on the record sheet. When the child turned his gaze away from the
colour, I called out ' off,' and again the exact second was noted and
recorded under ' off.' When again a colour was looked at its name was
called out. Thus the record would appear somewhat as follows :

Red Off Blue
5

11

25
30

21

28

45

and so on. From such a record it was easy to calculate that red was
looked at from second 5 to second 11 {i.e. for 6 seconds); that neither
colour was looked at between second 11 and second 21 ; then that blue
was looked at from second 21 to second 25, thus scoring 4 seconds, and
so on.

The experiment was stopped two minutes from the time at which
W. was first noticed to look at one of the colours. He was then played
with for a minute or two, after which another two minutes' test was
performed, the position of the colours this time being reversed, blue
now being on his right, red on his left. The total scores of the two
tests were added together, and the colour that scored the higher was
reckoned the more attractive.

One further note must be added as to the method of scoring. If W.
looked at a wool for less than three seconds, the score of course appeared
on the record sheet, but these scores were not included in the totals.
Thus a mere turning of the eyes for a moment or two upon the colour
does not count. It might indeed be suggested that this in itself
indicates some power of attraction possessed by the colour for the child.
But such a glance scarcely seems deserving of a score unless the colour
holds the attention for at least three seconds. Results are also given
based upon the times when the baby looked for not less than eight
seconds continuously at one colour.

Now there are obvious difficulties and dangers in such a method,
but I think that they are not so great as might be imagined. In the
first place it might be thought difficult to be certain whether the baby
was looking at the wool or at an object beyond it. But the marked
convergence of the baby's eyes, necessary to look at the colours, was of
great help to me here. It was impossible to mistake a look directed at
the background of my coat for one directed towards a wool.

C. W. Valentine 367

I gained confidence in the use of the method by practice obtained
in a preliminary series of twenty such experiments, during which, when
in any doubt, I moved the wool slightly to make quite certain that the
baby's gaze was fixed upon it. But the results of these experiments
are not included in the following records, for, as every observer of infant
life knows, a moving object is likely to attract the attention merely
because of its movement. In the subsequent experiments the wools
were held motionless.

I do not think that my hands were any serious source of attraction,
as the wools were held from behind. Further, W. looked almost in-
variably at the broad extended top end of the skein of wool where no
part of the hand was visible. Any occasional errors would tend to
spread themselves equally among all colours in a long series. Sus-
pension of the wools upon a screen might have avoided this difficulty,
and this plan was followed in some subsequent experiments. But it
was not found quite so easy then to follow the child's gaze, and my
impression was that the hand method was preferable.

The baby was perfectly free to move his head to right or left, as it
rested easily against a cushion. Very occasionally if he seemed to have
settled down somewhat to one side, the wools were moved slowly round
to that side (when he was not looking at either wool) so that neither
might have any unfair advantage of position. There is no record of any
occasion on which such slow movement attracted the baby's attention
to either wool. Indeed the very fact of his having moved far round
indicated as a rule that neither wool was showing much attraction for
him on that particular occasion.

A further danger was not overlooked, viz. the possibility of the
operator's favouring any colour in respect to position. The writer can
only say that he was extremely careful to avoid this and that he did not
start out with the object of supporting any particular theory of colour
vision.

In the later series of experiments just mentioned, the screen was
held steady throughout the experiment. But my final judgment was
that with a trained and conscientious observer it is preferable to be
able to adjust the position of the colours to any very marked change
in the baby's position \

The colours used were black, white, red, yellow, green, blue, violet,
pink and brown. The red, yellow, green, blue and violet were the

1 The plan of seating the baby upon his mother's knee was tried, but he obviously did
not like this so well, and tended to wriggle more than when seated on cushions.

368 The Colour Perception of an Infant

purest obtainable from the Holmgren wools. Of these the yellow was
only moderately intense. As I had not perfectly white wool at hand
I used a piece of dull white linen made into a roll of about the same
size and shape as the wools.

I should like to have included grey and orange. But as I used the
method of comparison in pairs (each colour being compared with each
in the course of the experiments), the addition of one further colour
would have meant eighteen further experiments and, as it was, the
series was prolonged as long as seemed advisable. For the baby tires
of the game if the colours are presented too often ; moreover it was
necessary that the last experiments should not be separated from the
first by too long an interval if the results were to be added together, for
rapid changes and development are taking place in the mind of the
child at this early period^.

At each sitting those colours were selected for use which had been
used least often previously, so that any effects due to novelty were
reduced to a minimum. For this reason also the same colour was never
used twice on the same day. Occasionally two sittings were taken on
the same day, usually one being early in the day and the other towards
the end. If the baby appeared restless or discontented, the sitting was
abandoned.

A very important question was the relative brightness of the colours
used. Owing to the generally acknowledged attraction of brightness for
infants it is obviously unsafe to draw inferences as to colour preferences
if the colours used are of unequal brightness. It does not seem satis-
factory for the experimenter to rely upon the general immediate
impression of brightness produced by the colours. Three different
tests were therefore used to determine the relative brightness of the
colours (see Appendix). The results of the three tests were in sub-
stantial agreement, and indicated that the yellow and pink were equally
bright and equivalent in brightness to a grey of 185° W. + 175° Bk.
The green, blue, violet, red and brown were also practically equal in
brightness, and equivalent to a grey of 32° W. + 328° Bk.

^ I regret especially, however, the omission of grey. At first I had in mind solely the
question of colour preferences. It was only after I had performed a fairly long series of
preliminary experiments that I became interested in the question of colour sense develop-
ment, and then results obtained with a new wool suddenly introduced would have been
unreliable. Otherwise it would have been well to have included a dull grey equivalent to
the red, green, blue, and violet in brightness and a bright grey equivalent to the yellow in
brightness, dropping perhaps the pink, and black or brown.

C. W. Valentine

369

IV. Results of Experiments.

Table I shows the number of gains for each colour. The sign +
when placed horizontally opposite to any name indicates that this
colour proved superior to the colour named at the top of the vertical
column in which the sign stands ; similarly the sign — indicates that
that colour proved inferior in attractiveness ; while the sign ? indicates
that the scores were so small or so nearly alike that the experiment is
regarded as indecisive.

Table I.

Yellow

White

Pink

Bed

Brown

Black

Green

Blue

Violet

Yellow

-

+
+

+
+

+
+
+

+

+
+

+
+

+

+

+
+

?

?

+
+
+

?
?

+

+
+
+
+
+
+

?

+

White

+

Pink

Eed

+
+

Brown

+

Black

?

Green

+

Blue

9

Violet

This gives us the following order

i

Table II.

Gains

Losses

Balance

of gains over losses

Yellow

8

0

8

White

6

2

4

Pink

6

2

4

Red

4

4

0

Brown

3

4

-1

Black

2

4

-2

Green

2

4

-2

Blue

1

6

-5

Violet

0

6

-6

In these tables, however, as much value is given to a gain by only
a small majority as to a gain by an overwhelming majority. It is
important therefore to compare the above order with that given in
Table III. Here all the scores (in seconds) of each colour have been
added up, together with the scores against them.

It would be unfair to draw up an order merely upon the total scores
of each colour. For such conditions as the child's mood, or the light,
may have been exceptionally favourable on those days when some
particular colour, say pink, was used, and exceptionally unfavourable on

370

The Colour Perception of an Infant

the days when some other colour was used. Thus green in Table III
has 165 to its credit — about the same as brown (151). But far more
seconds were scored against green (421) than against brown (275). We
therefore get a fairer basis of comparison by adding the score of green
(165) to the score against green (421), which gives us the total scored
during the experiments in which gi'een was used, viz. 586 seconds : and
then finding what percentage of this total was scored by green. Such
percentages are also given in Table III.

Table III.
Showing the total scores for each e^eriment.

^
o^

g

5

M

a
'A

c

1

P3

.2

"o

>

Totals A

Totals B

Percentage
scores

•^ o

0) u

i

1

C

'S

o

CD

a

A B

Yellow

White

Pink

Eed

Brown

Black

Green

Blue

Violet

42

4
23

8

0
23
37

0

76

96
24
3
38
9
0
0

17

48

26
0
39
19
37
0

106
70
46

28

10

20

8

0

31

104

79

40

18
0
3
0

88
53
19
45
33

0
13
12

133
153

77
6
4
3

22
23

50
22
102
39
-37
37
0

13

24

188
53
39
38

4
94

7

525
680
476
242
151
149
165
122
48

137
246
186
283
275
263
421
300
447

282

487

289

75

65

152

54

18

9

48
108

68
153
276
109
276
106
287

79-3 85-4
73-4 81-8
72-2 80-9
45-3 32-8
37-8 19-0
35-7 58-2
28-2 16-3
28-9 14-5

9-7 30

1

TotalB

137

246

186

283

275

263

421

300

447

1
i
i

In column A all scores of 3" and over are reckoned. In column B only scores of 8" and
over are reckoned. That is, when W. only looked at a colour 7" or less continuously, that
particular score was ignored.

Table III (percentage scores A) gives us an order almost identical
with that of Table I, the only change from Table II being that green
and blue are now brought almost on a level with one another.

Ordei' housed on Table III.

yellow 79-3 7o
white 73-4
pink 72-2
red 45-3
brown 37-8

black 35-7°/o

blue 28-9

green 28-2
violet 9-7

Consistency Test. By comparing Tables I and III we can obtain
some indication as to consistency. For example, judging from Table III,
brown ought to be the winner in the experiments in which it was com-
pared with black, blue, green and violet, and the loser when compared

C. W. Valentine 371

with yellow, white, pink and red. On inspecting Table I we find that

all these results were obtained, except that the experiment with brown

and green was regarded as inconclusive (brown 4 — green 0). Hence the

consistency score ma^^ be reckoned as at least 7| out of 8.

Proceeding in this way we find the following consistency scores :

11 8 , . 7 . , 6 , 7 , H . 1 1 6 , , 7^ 6

yellow 3 , white ^ . pmk - , red ^ , brown -— , black 3 , blue -~ , green ^ ,
0000 o 00 o

7
violet 3 , a total of 31 cases (out of 36) in which the results of individual

o

experiments were consistent with the order given by the total scores.
If indecisive experiments are ignored the result is only three incon-
sistent results in 32. When one considers the possibility of changes of
interest and mood from day to day this degree of constancy seems very
satisfactory.

V. Discussion of Results.

Two kinds of inferences can be made from such results as we have
obtained — (1) inferences as to colour preferences, (2) inferences as to
the development of the colour sense. We cannot of course infer from
the absence of preference between two colours the absence of any
difference of sensation^ But from evidence showing that one colour
is markedly preferred to another, we can infer that the colours are
sensed as different colours, unless the preference can be ascribed to
differences of brightness.

{a) The Presence of Colour Sensations. W. is obviously attracted
by the brightness of objects. White comes second in the list, and
yellow and pink, by far the brightest of the colours, are first and third
respectively.

But there is, I believe, adequate evidence that brightness is not the
sole cause of some of the coloured wools being preferred to others. The
strongest evidence of this is to be found in the great difference between
the score for violet (9*7 °U, on the one hand, and the scores for the other
colours of equal brightness, especially red (45*3 °/„) and brown (37*8 °/„).
Further evidence is afforded in that the score for red is higher than the
scores for blue and green, although they are each as bright as the red.

Among the warmer colours, then, red and brown must owe their
high position to the fact that their colour was both sensed and liked.
Green and blue are almost equal on the scale. But both of them were

^ Cf. C. S. Myers, op. cit. 358. McDougall himself explicitly states the logic of the
question on page 346 of his article.

372 The Colour Perception of an Infant

distinctly preferred to violet, so that sensations of blue and green were
presumably also experienced ^

The fact that yellow is more attractive than the equally bright pink
and even more attractive than white suggests tljat also yellow is both
sensed and liked. The difference between the percentage scores of
white and yellow is small, it is true ; but even if yellow had only scored
the same as white one might reasonably suppose that the colour of
yellow was appreciated. Otherwise, if the yellow wool appeared only as
a bright grey, one would expect a superior score for white owing to its
superior brightness. It may be suggested that a bright grey might be
preferred to a pure white owing to the latter being too glaring in
a strong light, and that this might account for the yellow, even if only
sensed as grey, being preferred to white. But evidence against this is
found in the fact that white actually scored more on the bright days
than on the dull days, the totals being as follows :

On four bright days On four dull days
Score of white 352 sees. 328 sees.

Furthermore, as has so often been observed in the case of infants of this
age, W. showed a delight in looking directly at very brightly lit windows
and at incandescent lights.

In any case we have seen that W. was almost certainly sensitive to
red, brown, and probably blue and green as such, and according to every
important theory of colour vision it is most unlikely that these colour
sensations would be developed in the absence of yellow.

There is no appreciable difference between the scores for blue and
green, and consequently no direct proof of discrimination. But the
difference between the scores for blue and green on the one hand and
for violet on the other indicate that some colour was perceived in the
blue and green wools.

That blue and green should have similar affective values for an
infant is not surprising in view of the fact that such seems to be the
case also with much older children ; at least blue and green are frequently
confused by them. It has however been shown by Miss A. W. Tucker^
that the weakness of young school children in the discrimination of blue

^ Unless one can attribute the different positions of these colours to the fact that the
colours themselves are sensed, one must introduce some hypothesis to the effect that
the colours have brightness-values for the eye of the infant which are different from their
brightness-values for the eyes of adults. This has never been disproved, but it seems
highly improbable that the great differences between the scores of red and brown (for
example) and that of violet are to be so explained.

2 This Journal, iv. 33.

C. W. Valentine 373

and green cannot be traced to lack of sensitivity to either blue or green,
but must be referred probably to psychological causes. A similarity in
the affective values of the two colours would doubtless aid in such
confusion.

Summing up the results gained so far we may conclude that probably
the sensations of red, yellow, green, blue and brown may all be experienced
by a child about three months old^

There remains the case of violet to be considered. These experiments
afford no direct proof that violet was sensed. One can only fall back on
the argument already referred to, namely, the difficulty of explaining,
on any theory of colour vision which receives wide support at the present
day, the existence of the sensations red, yellow, green and blue, if violet
is only seen as grey. If the physiological mechanism required for red
and also that for blue are functioning already it is difficult to see why
violet should not also be sensed.

(b) Colour Preferences. The order of colours as given in Table III
calls for some comment. McDougall suggests^ that the relative attrac-
tiveness of different colours for a child may be determined by two factors,
namely, brightness and novelty. We have already remarked upon the
influence of brightness, revealed by the high position of yellow and pink
as well as of white. That W. was also susceptible to the influence of
novelty, as were the children of McDougall and Myers, was shown by
the following short series of tests carried out shortly after the experiment
just described.

I procured some paper of a red colour, considerably brighter than the
red Holmgren wool, and slightly tinged with orange. A piece of card-
board, 10 cm. square, was covered with this paper, and was affixed to a
large wooden board covered with dark grey cloth. A piece of white
cardboard of the same size as the red square was also affixed to the
board, about one foot away from the red square. The board would
appear thus:

Red White

^ To speak more strictly we should perhaps say that red, yellow, green stimuli give
rise to colour sensations in an infant of that age. We cannot of course prove that the
sensations experienced are exactly similar to those which are caused by these various
lights in adults. ^ Op. cit. 349.

374 The Colour Perception of an Infant

This board was held on my knees in front of W. for 1| minutes, the
cardboard squares being equidistant from his central line of vision, and,
as before, we noted the length of time he looked at either of the cards.
The positions of the cards were then reversed (the previous right hand
card being now on the left and vice versa), and after an interval the
board was held before the child for another 1^ minutes.

Table IV shows the respective scores of red and white on successive
days.

Table IV

Red

White

l8t day (dull) 1 ...
2nd day (bright)
3rd day (dull) ...
4thday(duU) ...
5th day (moderate)
6th day (bright)

121 sees.

81 „

78 „

6 „

18 „

3 „

12 sees
33 „
10 „

13 „
17. „

3 „

These figures show most strikingly the effect of the loss of novelty
upon the attractiveness of the red square.

At the same time, I do not think that the order of relative attractive-
ness of the coloured wools for W. can be explained merely by the two
factors of brightness and novelty. We have already seen that the order
of the colours red, brown, green, blue and violet cannot be explained by
reference to their degrees of brightness, as they were all of equal bright-
ness. As regards novelty, violet was certainly the most novel colour
for W. ; yet it was liked least of all. On the other hand yellow, his
favourite colour, was the colour of his nureery wall-paper, while his
favourite plaything and a constant companion in his cot at this time
was a pink eiderdown quilt. Again, neither blue nor green would be very
familiar to him, for unlike McDougall's child, he had seen practically no
foliage, the experiments taking place in the spring in Scotland. Black
he was very familiar with ; I had previously noted how often he was to
be found gazing at a black piano, black hats, black boots, etc. Yet black
is liked better than the less familiar blue, green and violet.

The novelty that attracts, in such experiments as these, may, it
seems to me, be the novelty of the specific object with its specific
colour, rather than the rarity of the colour in the (^'dinary everyday
experience of the child. If a new wool is introduced suddenly into the
middle of a long series of experiments, I am inclined to think that its
novelty would attract for a few days, however familiar the colour might
be to the child. Where, as was the case in the experiments with W.,

^ The references in brackets are to the light on each pai-ticolar day.

C. W. Valentine 375

the presentations of the various colours were equally distributed over
the whole series, this particular source of novelty-effect would be reduced
to a minimum.

I would suggest in the most tentative manner, that at the early age
of three months, the relative attractiveness of any or all of the colours
is determined by a factor more fundamental than novelty, namely, their
comparative powers as stimuli to the organism.

Myers has already urged that the attractiveness of red for infants
and savages is of a very fundamental nature, and he refers to the
"excitatory action" of red upon organisms lower than man^ Fere
found that the general stimulating effect of colour is greater in the case
of colours at the ' warm ' end of the spectrum than of those at the violet
end. The strength of his subjects, as tested by the handgrip, was at its
highest when the patient was under the influence of red light, at its
next highest with orange and yellow, less with green and blue, and
least with violet^.

Now if it be supposed that yellow and pink owe their high position
partly to their brightness, then the order of the colours as preferred by
W. is remarkably like that given by Fere. It may of course have been
accidental that W. happened to like the colours in the order given, and
it would doubtless be unwise to press this suggestion as to the depend-
ence of the order of preference upon the powers of stimulation possessed
by the colours, unless similar results are obtained from experiments
upon other infants at a similar early age.

Several observers, as we have seen, agree upon a preference for
' warm ' colours among infants of from six to twelve months old. But
I am not aware that anyone has hitherto observed such a marked
indifference (or possibly aversion) to violet as was shown by W.^, or such
a preference for brown and even for black before the spectral neighbours
of violet, viz. blue and gteen.

It is possible that such extreme indifference to violet, blue and

1 Op. cit. 362.

2 See F6r6, Sensation et Mouvement, Paris, 1900, Chap. vi.

^ The records show that violet was looked at almost as often as green, blue, brown and
even white, so that its low score cannot be attributed to its escaping notice. The figures
for all colours were : violet 62, green 70, blue 72, brown 73, white 75, red 87, black 100,
yellow 110, pink 114. In making this calculation each glance given by W. to a wool is
scored as one, whether he looked at it for a minute or only for one second. Evidently
violet did not suffer through lack of opportunity. These figures indeed suggest that the
position of a colour in the order of preferences depended very little upon its power of
causing a reflex movement of the eyes towards it ; see especially the low score of white in
this list, less even than those of red and black.

376 The Colour Perception of an Infant

green, even if found in many children, may not be characteristic of
them beyond the age of three or four months. ( W., as we shall see, had
changed considerably by the age of seven months.) It is even possible
that young infants under, say, four or six months may resemble in this
respect the hysterical subjects of Fere, who proved to be more sensitive
to the different stimulating effects of the various colours than were
normal subjects.

But the preference for red continues beyond the period of infancy.
Thus experiments on children in Antwerp^ showed that red was the
most pleasing colour from four to nine years of age ; and Winch, ex-
perimenting in London schools, found that red was the best-liked colour
among Standard I children, yellow coming generally second or thirds

VI. Experiments by the ' Grasping Method'

When W. was seven months old I began further colour experiments
by means of the grasping method. Experiments were performed
practically every day for a month. The coloured objects used were the
same wools as were used in the previous experiments, except that
brown was omitted and a grey of the same brightness as the red, blue,
green and violet was introduced. Also I used a white wool instead of
white linen.

Two of the wools were placed on a table covered with a dark grey
cloth. W. was comfortably seated on his mother's knee, and one wool
was placed opposite each shoulder, about eight inches from the edge of
the table, just out of the baby's reach. While the wools were being
placed in position they were hidden by a black screen. About three
seconds after this screen was removed, W. was brought slightly forward,
bringing him close up to the table and within reach of the wools. This
plan was followed because at first he was so eager to grasp anything
within reach (including the screen) that he frequently seemed to seize
one colour without having looked even momentarily at the other, so
that there was no real * choice.' In spite of these few seconds' pause,
my impression still was that W. was often so eager to grasp a wool,
that the specific colour of the wool was comparatively unimportant, and
that only a very strong preference for a colour would cause it to be
chosen more often than its partner.

^ By Schuyten, quoted by Meumann in his Experimentelle Padagogik, zweite Auflage,
1911, 244.

- This Journal, iii. 42.

C. W. Valentine

377

A note was taken as to which wool was seized, and as to which hand
was used. When W. grasped both colours simultaneously, each colour
scored a half-mark. The same wools were then arranged in a position
the reverse of the preceding, i.e. the wool that had been on W.'s right
was now on his left and vice versa. The same pair of wools was
presented to W. ten times at one sitting, the position being reversed
each time. On the next day two other colours were similarly presented
ten times, and so on, until each of the wools had been presented with
each of the others. Great care was taken, each day, to choose those
colours which had remained unused for the longest period, thus mini-
mising any effects due to novelty. Table V gives the results of these
36 experiments involving 360 choices.

Table V.

Yellow

Red

Pink

1 Grey

Violet

-White

Green

Black

Blue

Total

Yellow

5A

5

1
! 5

8

7

6

7h

7

51

Red

H

6

' 5

7

6*

6

H

5

45^

Pink

5

4

1 5i

5

7

7

6

5

444

Grey

5

5

4i

i

5

5

5^

6

5

41

Violet

2

3

o

1 5

5

5

5

7

37

White

3

31

3

5

5

1

5

5

7

BH

Green

4

4

3

H

5

5

5

5

35i

Black

n

U

4

1 4

0

5

a

5

35

Blue

3

5

o

1 5

1

3

! ^

5

5

34

It will be seen that, as at three months, yellow still holds first place.
Red and pink are now almost bracketed second.

The most striking difference apparent between the results of the
experiments at seven months and those at three months is the drop in
the comparative attractiveness of brightness. White is now only on a
level with violet, gi-een, blue and black, and at least not more attractive
than a dull grey\ Pink, also, scores only about the same as red, while
at three months it was much preferred to red.

It should however be made clear that W.'s interest in the colours
during these later experiments seemed to be slight compared with his
enormous interest in the game of merely grasping them. Moreover,

^ It is conceivable, though it seems to me very unlikely, that the prolonged series of
experiments with these same wools from the age of 3 months to that of 4 months may
have resulted in these colours being felt vaguely as familiar even three months later
in this second series of experiments. In which case the comparatively high position of
grey (compared e.g. with blue) may be due partly to its novelty, as grey was not used in
the previous experiments.

J. of Psych. VI ■ 25

378 The Colour Perceptimi of an Infant

the scores of the. various colours were levelled to some extent by the
habit of using always the same hand. At first W. used almost
exclusively the right hand, choosing generally the colour on the
right. I attempted to get over this difficulty as Myers did, by placing
the wools in the median plane, one being about four inches nearer
the child than the other\ But I found that W. invariably took the
colour nearer to him, so this method was abandoned and the results
gained by it were not counted. Another variation of arrangement was
attempted. One wool was placed about three inches to the right and
the other about three inches to the left of a point immediately in front
of W.'s right shoulder, so that each wool was about equally well placed
for grasping with the right hand. Strange to say, this method had not
been in use for more than two sittings (experiments nos. 14 and 15)
when suddenly, on the occasion of the 16th sitting, W. began to use his
left hand more than the right. In spite of the unfavourable position of
the wools for the left hand, the left hand was used seven times and
the right only three times at this sitting. I therefore at once
reverted to the original plan of placing one wool immediately opposite
to each shoulder, and this arrangement was used throughout the
remaining experiments. The new preference for the use of the left
hand continued to the end of the experiments.

It seems extremely likely that a habit of using one hand predomi-
nantly may level the scores of the various colours to an extent which
conceals the real preferences of the child for certain of the colours,
though of course one can at least infer that these preferences are not
strong enough to overcome the tendency to use one hand more than
the other. Occasionally, in the midst of a long run chiefly of left-hand
grasps during the latter half of the experiments, the liking for a certain
colour would apparently assert itself and the right hand would be used
in order that this particular wool might be taken; e.g. in experiment
no. 24 (white and red) in ten grasps the right hand was used five times
alone and three times simultaneously with the left, but always in order
to grasp the red.

Occasionally, too, the left (or right) hand would cross over to seize
the wool on W.'s right (or left). The numbers of times the various
colours were seized in this way were as follows :

yellow 8

red 5

pink 4

grey 4

blue 4

white 3

black 3

green 3

violet 3

* Op. cit. 357.

C. W. Valentine 379

However, in spite of these possible variations in the way of grasping,
my impression was that, with so active a child as W., the eagerness to
grasp anything must materially level the scores of the various colours.
Furthermore, as this eagerness to grasp might be exceptionally strong
some days whilst on other days there might be a greater tendency to
attend to the colours as such, I would suggest that the method is far
from being so entirely free from all objection as has been thought.
Thus if W. were fatigued or bored with colouts on a day when yellow
and grey were presented he might very well every time grasp with his
then favourite hand (left); thus the score would be yellow 5, grey 5
(as was the case). He may also have been in a mood comparatively
indifferent to colours when he chose grey five times and violet five
times. These figures would suggest that the yellow and violet were
equally pleasing to him. Yet on another day, when more alive to the
influence of colours, he might choose (as he did) yellow eight times and
violet only twice, which suggests that the previous tests had been
unduly favourable to violet or unfair to yellow.

In view of this difficulty it seems to me that, when the subject is so
eager to grasp as W. was and falls so readily into a habit of using the
right (or left) hand predominantly, the grasping method is not so
satisfactory as the method used with W, at 3 months, if the precaution
is taken to reckon the percentage scores (see page 370) when using the
latter method. It is evident, however, that the difficulty mentioned
does not apply equally in the case of all children. McDougall's child,
for example, " grasped with both hands in turn usually with free alter-
nation, the use of the right hand predominating a little, although she is
by nature left-handed ^"

The only important point in which my results fail to agree with
those obtained by McDougall in his long series of experiments with L.
is the preference of L. for blue, green and white before grey. This result
I should be inclined to take as more reliable than the results given by
W., owing to the difference between the children which I have just
mentioned.

Of course one must guard against the assumption that the same
results ought to be given by different children. Individual differences
may exist between infants of six months in reference to colour prefer-
ences, as they do in reference to numerous other activities in the life of
a baby''.

1 Op. cit. 343.

2 Further, completely satisfactory comparisons are not possible owing to our ignorance
of the comparative brightness and saturation of the colours used by di£Ferent observers.

25—2

880 The Colour Perception of an Infant

Not only may colour preferences vary among diflferent infants, but
the attractiveness of colour in general is also doubtless greater with
some than with others. In the case of W. the desire to look at colours
seemed to be much less at seven months than at three, owing no doubt
to its being ousted by the exceedingly strong desire to handle anything
and everything. When W. was eight months old I made one trial of
the method used at three months, but was soon convinced that he
would not look for more than two or three seconds at a time at any
colour, so eager was he to be doing something.

VII. Experiments by the method of Grasp and Reward.

At the end of the experiments just recorded, when W. was 8^
months old, I attempted a further series by means of the method of
'grasp and reward' suggested by Myers \ No very definite results
were obtained, but I give some account of this set of experiments as an
example of this new method.

So far there had been no proof of the discrimination of blue from
green by W., though we had seen reason to believe that even at three
months they were each discriminated from violet and red. I therefore
presented the green and blue wools, as in the grasping experiments
just described, but with the following addition. Whenever W. took the
blue wool he was at once rewarded with a sip of honey or jam. At
first the pairs of wools were presented twenty times at each sitting.

The association between reward and grasping (something) was soon
established. At the second sitting, I recorded " when W. has taken
green and gets no reward he is very eager to seize the blue^." At the
fourth sitting I recorded " mouth open for jam as soon as wool (blue)
taken," and again " blue taken very deliberately. W. looked round at
once for jam " (the jam being on a plate on my right and W. on my
knee). Thenceforward W. very frequently looked round for jam, but
sometimes when green was taken.

It soon became evident that he tired of the honey or jam (though
only allowed a very tiny sip) before the end of the twenty presentations.
Thus at the seventh sitting, during the first ten presentations W.
turned every time to his right and evidently enjoyed the honey, whereas
in the second ten presentations he only turned once (when blue was

1 Op. cit. 364.

* Previously he had almost invariably been content to play with the one colour seized.
This attempt to seize the second wool may have been only a development of the tendency
to seize both wools, which McDougall observed as his children grew older.

C. W. Valentine 381

chosen). Thereafter I never gave more than ten or twelve presentations
at a sitting.

Results gained by the Method of Grasp and Rewa,rd. It seemed at
first a reasonable plan to divide the twenty choices of each sitting into
two sets of ten. If any association were built up in the course of the
sitting between blue and ' reward,' blue would tend to be chosen more
frequently in the second half than in the first.

Taking thus the seven sittings, of twenty choices each, we get the
following figures :

First half of each Second half of each
of the sittings of the sittings

Blue Green Blue Green

Number of times chosen 33 37 37 33

The increase in the number of times blue was chosen is seen to be
inappreciable. But probably it is unsatisfactory to divide the figures
thus. For as has already been pointed out, twenty choices proved too
many at a sitting, the desire for the reward obviously slackened and
sometimes seemed to change to aversion. More reliable results are
probably obtained by considering only the first ten choices. Comparing
the results given by only the first five choices and the second five choices
in all the eight sittings S one gets the figures

First five choices Second five choices
of each sitting of each sitting

Blue Green Blue Green

Number of times chosen 14 26 21 19

Here there is a distinct suggestion that the association is being
established ; green scores nearly twice as many as blue in the total
based on the first five choices of each sitting, but barely as many as
blue in the total based on the second five choices of each of the sittings.
The numbers are too small perhaps on which to base any confident
assertion.

My own general impression, if that counts for anything, was that
the association seemed to he made fleeting ly ; during some sittings it
would seem even strong, and then again it would entirely vanish. If
there were any association it certainly was of this fleeting and variable
nature, though of course that is just what one would expect, judging
from the beginnings of association in general in the child mind.
Certainly also it was not carried on from day to day. For during the

i One sitting of 10 choices was added after the seven sittings of 20 choices.

382 . The Colour Perception of an Infant

first four sittings blue was chosen twenty times and green twenty,
while during the last four sittings the numbers were, blue fifteen, green
twenty-five times.

I also tried the grasp and reward method with red and blue. I
argued that if a much stronger association between blue and reward
were built up when blue had only to be distinguished from red, the
absence or weakness of association in the blue and gi-een tests would
suggest that blue and green were not discriminated so well as blue and
red. On the other hand if no stronger associations were built up when
red and blue were used, the failure to establish an association when
green and blue were used need not be ascribed to lack of discrimination
between blue and green, but would show that at this age an association
could not be formed between one of two definite colours and a reward,
even when (as in the case of red and blue) the colours certainly gave
rise to different sensations.

In the red-blue experiments W. was rewarded when he chose blue,
as in the blue-green experiments^ The results were as follows :

First half of Second half of

each sitting^ each sitting

Blue Bed Blue Bed

Number of times chosen 18 25 22 21

These results show less evidence of an association between blue and
'reward' than was the case in the blue-green experiments. Nor is
there any indication that the association was established gradually fi-om
day to day. For, in the course of the first four sittings, blue was
chosen nineteen times and red twenty-five times ; during the last four
sittings the figures were, blue twenty-one, red twenty-three times.

In so far as these experiments with red and blue show less evidence
of the association than did the green-blue experiments, I think it may
be ascribed partly to the cause already mentioned (in the first footnote
to this page), and partly to the growing indifference of W. to the rewards,
though these were varied from time to time. I attempted still further
tests with blue and green after the red-blue series but abandoned them

1 Thus the 'reward' had to contend with the natural preference for red. No doubt
this hindered to some extent the formation of an association between blue and reward.
(Cf. Miss E. M. Smith's findings with dogs, "Some Observations concerning Colour Vision
in Dogs," This Journal, v. 176.) Obviously however no satisfactory conclusion could
have been made if the choice of the already preferred red had been rewarded.

2 Each sitting in the Bed-Blue series comprised 10 or 12 choices. As red was preferred
to blue by W. we should expect a larger number of choices of red until some association
between blue and reward was established.

C. W. Valentine 383

as useless. W. was much more interested in playing with the wool
chosen than in the reward ofifered, even though, at another time of the
day, when not busy with the more intellectual work of inspecting wools,
he would evince great delight in the honey or fruit-juice used as a
reward.

Summing up the results of these grasp and reward experiments, we
may say that they give some indication of an association with blue
when it has to be discriminated from green. Failure to obtain a
stronger association cannot be ascribed to the lack of discrimination
between blue and green ; for tests with red and blue, which were
undoubtedly discriminated, produced an even less decided association.

VIII. The use of the hands in grasping.

I have already mentioned the curious change made by W. when,
towards the middle of the first series of grasping experiments, he began
to use the left hand predominantly in grasping, instead of the right, as
at first.

Table VI shows the actual figures. It should be compared with the
table given by Myers showing the growing preference for the right
hand on the part of his child up to the 13th month, in spite of an
earlier preference for the left hand^

Table VI ^

Number of grasps

Bight band Left band Botb bands

1st set of 6 experiments ... ... 45 14 1

2nd „ ,, 58 2 0

3 expts. with both wools placed to the right of W. [19 9 1]

3rd set of 6 experiments ... ... ... ... 3 45 10

4th „ ,, 11 39 10

5th „ „ 10 41 9

Of the thirty-one occasions when W. used both hands, ten times he
grasped one wool in each hand practically simultaneously, and twenty-
one times he grasped one and the same wool with both hands. It will
be seen that the number of grasps with two hands increases in the
latter half of the experiments, which agrees with what McDougall found
with his children. Myers found the opposite, but his child was several
months older than McDougall's and mine.

1 Op. cit. 357.

^ For the sake of convenient division into series of 6 before and after the odd three
experiments, the first experiment (right hand 2, left hand 8) and the last two experiments
(right hand 1, left hand 10, and right 4, left 6) are omitted in these totals.

384 The Colour Perception of an Infant

In the later series of ' grasp and reward ' experiments immediately
following, W. was still more addicted to the use of the left hand, the
figures for the last 120 grasps being: Right hand 15, Left hand 105.

It is interesting to note that this specialisation of the left hand for
the purpose of wool grasping was not accompanied by any left-handed-
ness in other activities. W, still seemed to use his right hand more
than his left in his play, particularly in dealing with heavy objects,
such as books. McDougall also remarked that his child L. was by
nature left-handed, yet used the right hand somewhat more frequently
in the grasping experiments. During the early months of W.'s life, his
right-hand reflex grasp was stronger than his left ; and when voluntary
grasping began, the right hand learned to grasp much more readily
than the left. In the sixth month he was often seated on my knee by
the open piano, and had learned to hit the notes, greatly enjoying the
sounds, or his own production of them. This performance, which was
done at first with right or left hand indifferently, was soon relegated
largely to the right hand, and during the eighth month (in which the
specialisation of the left hand for wool grasping was observed) the right
hand was used almost exclusively by W. when thumping the piano.

Thus there appears to be specialisation of the right hand in one
action, and of the left in another action even at this early age. If this
were shown to be frequent in infants it might throw a new light upon
the problem of right- and left-handedness. Possibly in the case of a
right-handed child, the left hand may tend to specialise in actions which
are very simple, thus setting free the right hand for more serious work.
Baldwin found that more distant objects called forth the right hand
in a right-handed child, but that objects at an easy distance called
forth either hand indifferently. Attractive colours, he believes, act like
objects at a greater distance.

Perhaps, then, one may explain the change observed in W. from the
. predominant use of the right hand to the use of the left somewhat as
follows. At first, the coloured wools, by their strong attractiveness
drew out the right hand of W., naturally right-handed. As, with
practice, grasping the colours became easier, and as with familiarity the
wools became somewhat less attractive, the right hand was called forth
less, and the easy task of grasping relegated largely to the left hand.
But if this is a correct surmise it remains something of a mystery why
the change from right hand to left should begin just during those
experiments in which the wools were most favourably placed for the
right hand.

C. W. Valentine 385

IX. Summary of Results and Conclusions.

I. There is good evidence that at the age of three months an infant
may experience the sensations of red, yellow, brown, green and blue.

II. In the case of W. the order of preference of the colours used

_ „ 1, (white) , (brown) ( blue ] . , ^

was as follows : yellow, | ^.^^ J , red, |^,^_^^ | , |^^^_^j , violet.

III. The order of preference seems to be partly determined by
brightness, but cannot be explained entirely by reference to brightness
or to novelty.

IV. It is suggested that the order of preference is partly deter-
mined by the relative powers of the various colours as stimuli to the
organism.

V. At seven months the same infant still liked yellow best of all
the colours used, and then red and pink. By this time the comparative
attractiveness of white had decreased, being no greater than that of
violet or even grey.

VI. There was some suggestion of an association between the
grasping of the blue wool and the idea of a reward, when blue and
green wools were offered to W. The lack of more definite association
can be ascribed to the difficulty of establishing any association of such
a nature at this age, and need not be attributed to failure to discrimi-
nate blue and green.

VII. In the course of the grasping experiments W. developed a
habit of using the left hand almost entirely in spite of the fact that
he showed a distinct tendency to right-handedness in his ordinary
actions.

Appendix.

Determinations of the brightness values of the colours v^ed.

Test 1. The colours were placed, two at a time, on a dark grey
background, in the extreme periphery of the field of vision of an
observer, who fixated a point immediately in front of him. One colour
was placed some eight inches above the other and the observer was
asked to say whether the upper or lower colour appeared the
brighter. By this method the yellow and pink, naturally the brightest
of the colours, were judged to be about equally bright, and of the
other colours, the violet appeared slightly brighter than the blue

386 The Colour Perception of an Infant

and green, which in their turn were very slightly brighter than the
brown and red. This method, however, was not easy to carry out
satisfactorily owing to the persistence of the appearance of blue colour
in the blue and violet wools. But the very difficulty found in saying
which of the colours was the brighter, which always occurred with any
pair of the darker group (green, blue, violet, red and brown), is itself
an indication of the close resemblance of these colours as regards their
degree of brightness when seen in the peripheral field of vision.

Test 2. Each of the colours was placed in turn in the peripheral
field upon various shades of grey, until a grey was discovered of the
same degree of brightness on which the shape of the wool vanished.
Two subjects were tested. The yellow and pink were judged equal in
brightness to the same grey paper (equivalent to 185° W. + 175°Bk.).
The violet, blue, green, brown and red were also judged approximately
equal in brightness to another grey paper (equivalent to 32° W.
+ 328°Bk.). The blue and green appeared perhaps slightly brighter
than the others.

Test 3. Lastly the ^I'c^er test was used. A long piece of one of the
wools was closely gummed on to a piece of cardboard, the wool being
placed in parallel lines so that no interstices were left\ Similar cards
were made with all the wools.

Two subjects were tested by means of the flicker test. Each of
them found that the flicker disappeared simultaneously in the case of
the pink and yellow wools. Each member of the darker group of
colours (green, blue, red, brown and violet) was tested with each of the
other members of the group. In every pair of colours flicker dis-
appeared practically simultaneously. Thus the brightness of the
colours as seen by direct vision was approximately the same. The
equivalence of the brightness values of the various colours and those
of the greys, the black-white values of which have been given above,
was also confirmed by the flicker test.

^ This method was suggested to me by Dr C. S. Myers. See his Text Book of Experi-
mental Psychology, Part u. Experiment 68, for details of method.

{Manuscript received 6 November, 1913.)

THE TESTIMONY OF NORMAL AND MENTALLY
DEFECTIVE CHILDREN i.

By T. H. pear and STANLEY WYATT.

{From the Psychological Laboratory, University of Manchester.)

I. Introductory.

II. Description of the experiments.

III. Subjects taking jjart in the experiments.

IV. The children's testimony.
V. Treatment of the results.

VI. The results obtained.

A. The ^narrative.'

B. The ^interrogatory.'

The categories in detail; items, colours, sizes, duration,

sequence.
' Reconstruction ' of the event.
VII. Conclusions.

I. Introductory.

Two important questions in connexion with the psychology of
children's testimony still remain unanswered. They are^ :

(1) How far is the testimony of normal children reliable for
ordinary purposes ?

(2) To what degree, and in what directions, does the testimony of
the mentally defective child differ from that of the normal child of the
same physical age ?

Examples of the disagreement in the answers to these questions may
be given here. Babinsky^ declares that children are the most dangerous

* Amplified from two papers read before the Sub-section of Psychology at the Meeting
of the British Association for the Advancement of Science, Birmingham, September, 1913.

^ The questions have been formulated by Whipple, "The Psychology of Testimony,"
Psychol. Bull. 1911, vni. 307.

^ Die Kinderaussage vor Qericht, Berlin, 1910.

388 The Testimony of Children

of all witnesses, and demands that their testimony be excluded from
court record wherever possible. Similar statements are made by
Dupree^ Gross^, however, stakes his thirty years' experience in the
court against the views of these physicians. He declares that a healthy
half-grown boy is the best possible witness for simple events ^

With regard to the testimony of mental defectives, little experi-
mental investigation has been undertaken^ and opinions seem limited
to the statement that their testimony is less valuable than that of
normal persons, or to similar expressions which give us no information
of psychological value.

The present experiments, by comparing the testimony, of the same
event, given by normal and by mentally defective children, attempt to
obtain some facts which may throw light on the two questions mentioned
above. They deal, however, only with children of school age, so that
the related question of the comparison of the testimony of normal and
of mentally defective adults is still untouched. Many other points
towards the elucidation of which the experimental results offer some
evidence will be treated later. Some of the chief are : 1. The
comparative value of the testimony, of the same event, of normal and
defective children. (Very often the comparison has not been made in
this way, but different experiments on the two classes of subjects have
formed the basis of the statements made.) 2. The relative value of
' spontaneous ' and ' interrogated ' evidence. 3. The relative effects of
different kinds of suggestive questions upon the two classes of children.
4. The effect of repeating the testimony after a long time, in which
the subjects have had the opportunity to think over the event and to
discuss it with their friends, including both those who have, and those
who have not witnessed it.

It may be pointed out that in several respects the present experi-
ments differ from those which have been performed elsewhere. In the
first place, there seems to be no published record of any experiments on

1 "Le t^moignage: 6tude psychologique et m^dico-legale," Rev. d. deux Mondes,
1910, Lv. 343-370.

2 "Zur Frage der Zeugenaussage," H. Gross' Archiv, 1910, xxxvi. 372-382 (cited by
Whipple, op. cit.).

* Whipple, op. cit. p. 308.

•• For a general account of the experimental work on the psychology of testimony see
Whipple, Manual of Mental and Physical Tests, Baltimore, 1910 (gives bibliography up
to 1909); Psychol. Bull. 1910, vii. 2; 1911, viii. 9; 1912, ix. 7; 1913, x. 7; Dupr^e, op.
cit. The most important sources are Stern's Beitr. z. Psychol, d. Aussage, Leipzig,
1903-6 ; Erinnerung, Aussage und Liige, Leipzig, 1909, and articles in the Ztsch. f. angew.
Psychol., especially 1911, iv. 378-381 (bibliography from 1908-10).

T. H. Pear and Stanley Wyatt 389

testimony in this country ; most of the work having been done on the
Continent or in America. It is possible that the spontaneous interests
of children in different countries are not identical : so that one cannot
conclude that experiments performed e.g. in Germany give results which
are valid for England. Comparative work of this kind is needed.
Again, the event used as the basis of these experiments was, we believe,
more complex and richer in incident than those previously used. It
was also repeated six times before subjects of different grades of mental
efficiency, and in different environmental conditions. Finally, the
number of subjects (143) was larger than usual.

11. Description of the Experiments.
The ' picture test ' and the ' event test ' ; their relative advantages.

Previous workers in this field have employed, as the object on which
a report has to be given, either a picture or a pre-arranged event.
Both forms of test have their own advantages and disadvantages. The
picture can be made very complex ; it may contain a great number of
items, colours, positions, etc. : it is a constant and invariable stimulus,
and so can be employed many times on different classes of subjects, for
purposes of comparison. Its disadvantages are equally obvious. It
lacks solidity, movement, and temporal sequence — three facts which
detract greatly from its value in an experiment which is intended to
approach the natural conditions of ordinary life. Again, it is exceed-
ingly difficult to present a picture to a class of subjects for a definite
time, and still to keep them in ignorance that they will be required to
report afterwards upon what they have seen. The 'event' as a test,
when carefully arranged, suffers from none of these disadvantages, but
up to the present, it seems to have been comparatively simple, so that
there has been relatively little to report when it was over, and hence the
opportunity of studying individual differences has not been very great.
The repetition of the event, too, in front of other classes of subjects,
does not seem to have been carried out very often, perhaps because of
the difficulties inherent in such a procedure.

The present experiments attempted to combine the richness in
detail and complexity of the ' picture test ' with the naturalness of the
' event test.' One of the items figuring in the event was a picture, so
that we have here one class of test inside the other. The test, too, was
repeated six times before different groups of subjects. The repetition

390 The Testimony of Children

of the test made it possible to use small classes of subjects at any one
time, so that they all could easily see what was happening.

The event chosen. Requirements to he fulfilled. It was desirable to
select an event which, although containing more incident than those
used by previous workers, was still capable of exact repetition. It must
also be striking, but not artificial in appearance, in order to avoid the
arousal of the children's suspicion. Again, the event must not be too
striking, or an unusual amount of subsequent reflexion upon it would
have been provoked.

Since items which are connected with living persons prove to be of
great interest in such circumstances, two persons of different sex were
selected to take part in the event, and moreover, since, under ordinary
conditions, an event does not always show only one well-marked focus
of interest, both persons were active at the same time. The fact that
both in the ' performers ' and in the onlookers in the event the two
sexes were represented makes it possible to ascertain the direction of
the interests of the various subjects in the performers.

It may seem that the event to be described, complex and full of
interest as it was, would have been likely to arouse the suspicion in the
children that some kind of test was being carried out. To find whether
this was the case, at every school the children were asked, after they
had given their testimony, if they had suspected anything of the kind,
or if they had thought that the event was anything but an ordinary
visit to the school. One child only, out of the 143 subjects, had had any
suspicion, and he " thought that it might be a memory test." (This was
at the Fielden Demonstration School, Manchester.) But the boy was
careful to add that on that account he did not tell any of the others
what he had thought. In all cases the children were genuinely
surprised when, the next day, they were asked to write down an account
of what had happened.

The event was enacted on three different days in the six schools
chosen^ which were situated in Manchester, Bolton and Liverpool'
towns which are sufficiently wide apart to make it unlikely that the
news of the experiment would spread. In any one town the event
always took place in the different schools on the same day. The same
two persons (hereafter referred to^ as A and B) always carried out
the event, with none but the very slightest alteration in different

1 For descriptions of schools and subjects see pp. 394, 395.

* A was one of us; B was Miss N. Hilton, a graduate research student in the psycho-
logical laboratory. We offer her oar heartiest thanks for her valuable assistance.

T. H. Pear and Stanley Wyatt 391

schools {e.g. the unavoidable alteration of a colour in the dress or of the
position of an object), and with gratifying success. In any case in
which an alteration in the rehearsed order was made, it was communi-
cated to one of us, and allowed for in scoring the results.

A time of day was chosen at which the children would be engaged
in a singing lesson : this ensured that they had nothing to occupy them
when the two persons entered the room. This precaution was necessary
in order to avoid the possibility that the children might simply
continue their work when the interruption occurred, and take no
notice of the event. We can confidently assert that this did not
happen. The order of proceedings allowed B to observe the children
during the whole time that she was in the room, and she reports that
all the children took great interest in the doings of the visitors.

The classes were always arranged in a compact group when the
interruption occurred, so that the whole event was seen by all, and no
child was very far from the performers. The fact that some children
were nearer than others could not, of course, be avoided, but the objects
figuring in the event were all large, and no questions were asked which
referred to details of small objects.

The teacher in charge of the class had been previously told that
the event would take place, but that she must not prepare the class in
any way, and must act quite naturally. The teachers did not know the
exact time at which A and B would enter; all they knew was that
during a certain lesson there would be an interruption, of the general
nature of which they were made cognisant. The fact that they them-
selves knew only in a general way what was to take place contributed,
no doubt, to the interest which they took in the proceedings. It
also avoided the risk of the hint that an unconsciously hlase expression
on the teacher's face might have given to the children, had the teachers
previously seen the rehearsal of the event.

Description of the dress of A and B.
Dress of ^.

Navy blue suit.

Fawn coat.

Blue and white striped muffler.

Red and blue striped tie.

Black 'bowler' hat.

Black boots.

392 The Testimony of Children

Dress of B.

Navy blue costume, black buttons.

Pink blouse with lace over it.

Brown fur and muff.

Brown fur hat, turned up with white, and trimmed with orange

velvet and grey feather.
Fawn gloves.
Green bag.

Violets pinned on muff.
Red roses.
Black shoes.
Blue stockings.
Coat well open at the front.

Description of the pre-arranged event. The lesson was interrupted
by the entrance of A and B, who knocked at the door of the class-room,
entered, advanced to the teacher in charge of the class, and shook
hands. A introduced B to the teacher in a low tone. The following
performance was carried out :

On entering the room, A carried a brown bag in his right hand, his
black ' bowler ' hat and brown gloves in his left hand, and a yellow cane
walking-stick over his left arm. After putting his bag on the floor, he
shook hands with the teacher and made an introductory remark, " I have
brought those things to show you that I told you about." He put
his hat, gloves and stick on the piano ^ then placed his bag on the
teacher's desk, opened it, and very quickly took out the following
articles in the order named below :

A newspaper {The Daily Citizen), — thrown out, apparently carelessly,
but so that it rested against the bag with its name exposed to the class,
— a bunch of keys, a briar pipe, a box of matches, a pocket-knife and a
book.

These were taken out of the bag very quickly, and put carelessly on
the desk in order to give the impression that A was looking for some
things which they were covering. The following articles were then
taken out more slowly and carefully, and were held up and shown to
the teacher, quite naturally, but in such a position that the whole class
could see them.

A small flag, consisting of a green oblong background on which was

^ In the case of Group Y this was not done, the articles being deposited elsewhere.

T. H. Pear and Stanley Wyatt 393

a yellow cross, a coloured statuette of a shepherd boy holding a lamb, a
bunch of artificial roses (red and white), and a large paper picture of a
cat and a canary, which had the title, "The Cat and the Canary,"
printed on it in large letters.

The picture was unfolded and shewn to the teacher (and thus to
the class, as it was held in such a position that the whole class could
see it) for ten seconds.

(The picture was selected from a stock of posters used to advertise
kinematograph films at the picture theatres, and was one which had
not yet been put into circulation. A picture of this kind was chosen
because the items on it were simple, familiar, boldly drawn and
distinctly coloured. The newspaper was selected in order to see if the
more common names of papers which also have the word ' Daily ' as part
of their title would be reported instead of the real title.)

As the articles were taken from the bag, appropriate remarks were
made in a low tone, such as " How do you like this ? " " Will this do ? "
The articles, after being shown to the teacher, were placed on the desk
at the side of the bag, and were afterwards replaced in the bag in the
reverse order to that in which they had been taken out. In the mean-
time B, after having been introduced to the teacher, stood at one side
of A, and assumed a bored expression. (It may be said that her acting
was excellent.) She took some violets from her muff, rearranged them,
and then replaced them. The two visitors then shook hands with the
teacher, said " Good-afternoon " to the teacher and the class, and left
the room. The lesson was immediately continued.

The performers were naturally careful to leave the school premises
as quickly as possible, and to arrive only a minute before the event
took place, in order to avoid being seen by children other than those
concerned in the experiment. One of them started a concealed stop-
watch as they entered the room, and stopped it as the event finished, so
that the duration of the event was known. It varied between 2 minutes
10 seconds and 2 minutes 35 seconds.

The room was always chosen so that those entering it could not be
seen approaching the glass door, and would not be visible until they
had completely entered the room. In every case the event took place
during the afternoon session of the school.

The presence of B served several purposes. Besides tending to
provide a counter-attraction for the onlookers, she added many details
of dress and appearance which could be used as the basis of testimony,
and, being relatively unoccupied, she was able to watch the children

J. of Psych. VI 26

394 The Testimony of Children

unobtrusively, and to report anything of interest in the behaviour of
the class.

The Subjects' attitude toivards the experiments. From the evidence
of the teachers, quoted on p. 390, we know that the children did not
suspect a test. From our general impression gained from conversation
with the teachers, it does not appear that the event created any
extraordinary amount of interest amongst the children, or that they
spoke about it much amongst themselves when they dispersed or
reassembled the next morning. The fact that the event was followed
by a period during which they were in the presence of the teacher
prevented them from comparing notes, to any great extent, directly
after the event.

As the testimony was taken early in the morning after the event,
we may then conclude (and in this we are supported by the statements
of the children themselves) that very little leakage of the account from
one child to another had taken place. The usual impression of the
occurrence seems to have been that " the gentleman wanted to sell the
things," or (in the defective schools) that he " had made them." After
the children knew that they had been victims of a plot, there was
naturally an exchange of recollections, and this we have investigated by
suddenly requiring the testimony of the children again, seven weeks
later. In the case of any event of which children might be required to
give evidence it is often difficult to avoid the effect of knowledge of the
experience of the same event by others, and the value of an estimate of
the reliability of their own impressions, before and after consultation
with others, is obvious.

III. Subjects of the Experiments. '

The results of this investigation were obtained from two schools for
normal children and four schools for mentally defective children', the
six different groups being constituted as follows :

Group I. 39 normal children, of both sexes, attending the Fielden
Demonstration School, Manchester. Their ages ranged from 11 to 14
years, and in social status they were slightly superior to the average
scholar of the elementary school. At this school modifications in the
curriculum constantly occur, and visitors frequently enter the class-

^ I.e. children of such a degree of mental subnormality that they had been adjudged
to require teaching in 'special' schools (in the sense usually attached to this word in
England).

T. H. Pear and Stanley Wyatt 395

rooms, consequently the occurrence of an event such as that which
formed the material of this investigation would be expected to cause no
undue excitement. As a ' control ' experiment, an elementary school
in which the curriculum and arrangements are more constant was visited,
viz. the school mentioned under II.

Group II. 26 normal girls, ages 11 to 12 years, at the Clarendon
Street Council School, Bolton.

Group III. 28 mentally defective children, 14 girls and 14 boySj
ages 11 to 14 years, at the Chatham Place Special School, Liverpool; f •

Group IVi 20 mentally defective children, 15 boys and 5 girls, ages
11 to 13 years, at the Orwell Road Special School, Liverpool.

Group V. 16 mentally defective boys, ages 11 to 14 years, at the
Flash Street Special School, Bolton.

Group VI. 14 mentally defective boys, ages 10 to 13 years, at the
Kay Street Special School, Bolton ^

Since most of the children of the different groups fall within the
same age limits, it will be possible to compare one group with another
with respect to the efifect of the different conditions of life upon their
powers of giving testimony.

IV. The Children's Testimony.

The ' nan^ative ' and the ' interrogatory.' It may be mentioned that
the above two terms are now frequently used in connexion with this
work. The narrative (corresponding to the Bericht of Stern) is the
account given by the subject when he is allowed to proceed in his own
way, unhampered by questions or by any personal influence. In the
interrogatory (the Verhor of Stern), on the other hand, he is required
to give answers to a set of pre-arranged questions, which are read to him
by the experimenter, and some of which may be suggestive in varying
degrees.

In the present experiments a period of 19| hours (a period found
to be most convenient in the case of the first school, and so kept
constant in the subsequent experiments) was allowed to elapse after
the event, then the teacher, without preliminary warning, said to the
children: .

" I want you to write an account of everything you saw from the

'* We offer our very hearty thanks to the head masters and head mistresses of these
schools for their kind permission to carry out the experiments, and to the clafla teachers
for their valuable co-operation in obtaining the results,

26—2

396 The Testimomj of Children

time the lady and gentleman entered the room to the time they
went out."

When all had completed the narrative, the papers were collected
and the children instructed as follows :

"I am now going to ask you a number of questions. You will
probably not be able to answer all the questions, but some of you will
be able to answer more than others. When you are unable to give an
answer, leave a space where the answer should be. The answers must
be as short as possible."

The words used by every teacher, and the order of proceedings had
been fixed by us, and instructions for carrying on the tests were in the
teacher's hands on the morning after the event, so that the conditions
of the test were the same in each school.

V. Treatment of the Results.

In the evaluation of the results, an ' item ' was taken as the unit of
measurement, and an item was defined as any particular piece of in-
formation about the event. Thus, in the phrase, ' A brown bag on the
table,' there are four items of information. It was thus possible to give
exact numerical expression to the range (the number of items mentioned
whether correct or incorrect) and the accuracy (the number of items
correct) of the narrative and the interrogatory.

In both the narrative and the interrogatory, one mark was given for
each item correctly expressed, and a similar value was awarded to each
incorrect item.

An event may conveniently be divided into categories, and accord-
ingly the following were selected as relevant to the aims of the
investigation : items, colours, shapes, sizes, position, action, sequence,
and number.

VI. The Results obtained.

A. The Narrative.

The mean and standard deviations^ (o-) for each of the groups are
given in the following table. The value for the range and number
correct are given separately. (For the description of Groups I, II, etc.
in this and subsequent tables see pp. 394, 395.)

^ The standard deviation is the square-root of the mean of the squares of the deviations
of the separate values from the mean of all the values, and is a measure of the degree of
scatter of the separate values about their mean.

T. H. Pear and Stanley Wyatt

397

Group

Number
of cases

fiange '

Number correct

Percentage

Mean

1

Mean ! o-

correct

I

n
III

IV
V
VI

39
26
28
20
16
14

47-7
74-8
24-7
16-9
310
18-1

30-2 1
19-6 i

8-6 1

60
16-5

7-9 ,
1

460 27-2
72-0 18-9
23-1 ! 8-5
16-3 5-9
301 15-5
17-7 7-8

96-3
96-1
93-1
96-3
97-0
97-7

It will be seen that in every group the degree of accuracy attained is
remarkably high, and hence the spontaneous account of an event is
exceedingly reliable, even in the case of mental defectives. In many
respects, not a single deviation from the actual situation weis to be
found. Thus, when the testimony of children is unaffected by questions
or suggestions, it is worthy of the utmost consideration.

However, the amount proffered by the different subjects varies con-
siderably, and in no case is the range very extensive. A glance at the
table will show that the girls of Group II give the most detailed account
of the proceedings (mean, 74*8) ; the reports of the children of Group I
are much 'thinner' (mean, 477). The mentally defectives tender still
shorter accounts of the event ; the means of Groups III, IV, V, and VI
being respectively 247, 16*9, 31 '0, and 18*1. In this respect the spon-
taneous account given by the mentally defectives is much less valuable
than that given by the normal children. Such quantitative difference
provides a means whereby the two classes of children may be dis-
tinguished from one another. Individual differences in the range of the
report are generally well marked ; this is especially noticeable in the
children of Group I (<r = 30'2). Although the possibility of individual
variations is greatest in Group II, since this group yields the highest
mean, the standard deviation is relatively small (19*6) ; a value which
shows that the performances of the girls of Group II are steadier and
fluctuate less than the children of Group I. This may be due in part
to the different methods of teaching employed at the two schools. The
more rigid discipline to which the girls of Group II are subjected may
tend to produce a greater degree of uniformity in the results than the
freer discipline which prevails at the Fielden Demonstration School.

Classification into Categories. The following table shows the separa-
tion of the results into the categories mentioned on p. 396.

The upper line in each group represents the mean number correct

308

The Testimony of Childre^i

per subject in each category ; the lower line gives this value as a per-
centage of the total number correct in all the categories.

Group

to

S

CD

o

Si

03

o

'S
o

"3
a

O 00

2 a

a

0)

3

3

o3

u

S

- a

h- (

'^

S

M

(^

<

Ph.13

QQ

S

;z;

^ 1

121

2-7

•9

17

4-5

15^4

32

3-9

•8

•7

26-4

5-9

2^0

3-7

9-8

33^6

6-9

8-5

1-8

15

■ „ 1

10-6

5-3

1^4

1-8

73

25^0

13 9

3^9

23

•5

14-8

l-o

20

2-4

101

34-7

19^2

5 4

31

•7

Mean* -j

11-3

4-0

12

1^8

5-9

202

8^5

3^9

1-6

•6

20-6

6-7

2^0

31

9^9

34 1

13 1

6-9

2^5

11

n. (■

7-0

•18

•11

•14

41

8^6

23

•61

•18

30-2

■7

•5

•6

17-6

370

9^9

2^6

•7

—

IV j

5-4

•2

•05

•15

1-6

6-4

2-0

•3

■1

•05

33- 1

1-2

•3

•9

9^8

393

12-3

1^8

•6

•3

V i

8-8

•62

•12

•18

2^8

11-6

5^0

•75

■24

—

29-1

21

•4

•6

94

38-5

16-6

2 5

•8

—

VI 1

5-6

•07

■

•07

2^6

4^9

37

•29

•29

•07

31-5

•4

.—

•4

14^9

28 ^2.

20-9

1-6

16

•4

Mean* -

6-7

•27

•09

•13

2^8

7^9.

32

•49

•20

•06

310

11

•4

•6

12^9

36^0

14-9

21

•9

•3

* In evalaating the means of the different groups no allowance has been made for the
different numbers of subjects in the groups.

From a consideration of the above table it is at once evident that
certain components of the event appeal to the children more than others.
Over one-third of the items enumerated are included in the category of
action, a fact which tends to show that children are primarily interested
in the moving aspects of the situation. This illustrates very clearly the
defects of a picture as material for a testimony experiment, since it may
fail to stimulate the subject's interest in movement. Another category
which attracted a prominent share of the subject's interest is that of
* items.' For convenience, and also in order to effect a more complete
analysis, this category was divided into items connected with the
persons (these included articles of dress and personal features), and
into items not directly associated with A and B, such as articles taken
from the bag. The only other category which is noticeably prominent
is that of ' position.' The remainder of the narrative is very insignificant;
-there is a remarkable lack of evidence, upon such categories as colour,
shape, and number, particularly in the case of the mentally defectives.
95 7o of the evidence of these mentally defective subjects is made up of

T. H. Pear and Stanley Wyatt 899

a description of actions, items, and positions ; the remaining 5 °l^ is
divided between all the other categories. These three categories con-
stitute over three-fourths of the narrative of the normal children, though
with this class of children the remaining categories assume a relatively
more important position. The fact that in the case of the normal
children, only 6 '7 °l^ of the description deals with colours is remarkable
when one considers that colours formed such an apparently prominent
part of the event. Possibly the name of a familiar object tends to be
given in preference to its colour, shape, or size, and only when the object
is unfamiliar do these categories receive relatively more attention.

There are several points of difference between the narratives of the
normal children and those of the mentally defectives. As we pass from
the nan-atives of the defective to those of the normal children, such
components as colours, sizes, shapes, etc. are found to be mentioned
more frequently. Also, those defectives who produce the most intelligent
narratives begin to note these categories. Further, the mentally defec-
tives usually describe those parts of the event in which they are most
interested, independently of their chronological order. The normal
children, on the other hand, almost invariably describe the event in
the order in which it took place.

The narrative of the defectives is often fragmentary and discon-
nected and repetitions frequently occur. Often the transition from one
sentence to another is very abrupt.

Some of the descriptions given by the mentally defectives throw
much light upon the condition of the minds of these subjects. In one
case, it was said that the bunch of keys taken from the bag was used to
wind up the gentleman's watch. The canary in the cage was often
described as a " poll-parrot." To one, the flag was " a piece of cloth,"
and to another, " a cross on a piece of cloth." One subject even went so
far as to state that she " saw a book with lots of pictures in it," although
the book in question was never opened. The normal children do not
show such wide deviations as these, yet occasionally imaginative
statements are made.

A feature of the papers of the normal children, and especially of
those of the girls of Group II, is the aesthetic appreciation of the
appearance of A and B. It was often said that B's, clothes suited her ;
that her clothes were very nice and she looked nice to speak to. One
girl in Group II was evidently very much impressed, for she states that
"they were exquisitely dressed, and the lady's clothes and colours
harmonized with each other. They were not very old, but were very

400

The Testimony of Children

graceful and polite, and when they stood beside each other they went
very well."

Distribution of Interest between A and B. The results of the narrative
were further arranged to show the distribution of interest between A
and B. The following table shows the extent of this distribution.

Group I

Group II

Group III

Group IV

Group V

Group VI

Number
correct

A
B

14-4
7-2

22-7
25-7

7-4
5-4

4-9
4-0

7-4
5-3

5-4
4 0

It will be seen, that with the exception of Group II, more information
is given about A than about B. The description of A given by the
children of Group I is twice the length of the description of B. In
order to ascertain if the difference between Groups I and II were due to
the presence of boys in Group I, the values for the boys and girls of this
group were evaluated separately and were found to be as follows :

Number correct per girl ...
Number correct per boy . . .

13-0
15-8

B
70

7-4

Thus the description of A given by the girls is shorter than that given
by the boys, but it cannot be said that an appreciable sex difference in
interest exists here.

The girls of Group II give equally good descriptions of A and B. If
the results of the children of Group III, which contains 14 girls and 14
boys (mental defectives), are treated separately, we find that the descrip-
tions of A and B given by the boys are longer than those given by the
girls. The averages for the boys are 8'4 and 6"6 respectively; the
corresponding values for the girls are 6'3 and 4*1.

B. The Interrogatory.

Range and Accuracy. The questions asked were of a most com-
prehensive nature, and dealt with all the aspects of the event. They
were given orally, in order to minimise the temptation which would
have arisen if the questions had been printed, to look back and answer
any question which had been passed over, when the answer to it had
been suggested by a question occurring later in the series.

T. H. Pear and Stanley Wyatt

401

In the following table will be found the average number of correct
and incorrect replies, together with the mean variation (m.v.) for each
of the groups tested'.

No.

of

cases

Number correct

Number incorrect

Group

Mean

M.v.

Mean

M.v.

1st test

2nd test

1st test

2nd test

1st test

2nd test

1st test

2nd test

I
II
III
IV

39
26
24
19

73-2
64-0

48-2
54-8

68-2
63-4
49-5
47-4

7-8

9-8

11-6

5-8

7-0
8-3
7-8
5-5

49 1
31-5
54 1
81-7

55 0
42 1
63-6

80-8

9-8
12-7
111

8-3

11-4

12-4

15-7

5-3

It is at once evident from this table that the children of the different
groups show considerable variation in the number of questions answered.
Some children give an answer to every question ; this may be the out-
come of the desire to excel or to please, but it has, of course, a disastrous
effect upon the reliability of their evidence. Others are more critical
and cautious, and consequently fail to answer many of the questions.

A comparison of the number of correct and incorrect replies gives
an indication of the accuracy and hence of the reliability of the in-
terrogatory in general. It will be remembered that in the narrative
the number of errors was almost negligible. This condition, however,
does not obtain in the interrogatory. Over one-third of the replies of the
normal children and over one-half of the replies of the mentally defec-
tives are incorrect. Thus the interrogatory of the latter group is very
unreliable, and the corresponding testimony of the normal children
must be treated with great reserve.

It is interesting to note the inversion which takes place in passing
from the normal children to the mentally defectives. In the case of
the former group the majority of the replies are correct ; in the latter,
the correct replies are in the minority.

Repetition of the intei'rogatory after an interval of seven weeks. The
interval of seven weeks which elapsed between the two interrogatories
had only a slight effect upon the children's memory of the event. As
might be expected, the accuracy of the second interrogatory was less

1 Groups V and VI are not included in this table as their results in the interrogatory
were rendered useless by a slight misunderstanding of the directions.

402 TJie Testimony of Children

than that of the first, the decrease in the number of correct replies
being accompanied by an increase in the number of incorrect replies,
but not to the extent that might have been anticipated. Still, the
increase in the number of incorrect replies is correspondingly greater
than the decrease in the number of correct replies. Thus, what was
uncertain in the first place becomes still more uncertain later, but what
was distinctly perceived suffers little at the hands of time\

Classification into categories. The table on page 403 shows the
division of the replies into categories.

The upper row of each category represents the mean value per
subject. The lower row gives this value as a percentage of the total
number possible in the columns headed ' Number Answered,' and as
a percentage of the number answered in the columns headed ' Number
Correct.'

Range. The percentage number of questions answered in the different
categories is fairly uniform. In the case of the normal children it ranges
from 60 to 80°/^. The questions which refer to actions are answered
most frequently ; the next in order of fi'equency are those which relate
to personal items and positions. The defective children respond most
frequently to the suggestive questions (85"8 °/q), and to a slightly less
extent to those concerning colours (81 "8 7o)> actions (78'6 °/^), and posi-
tion (77*7 °U. The tendency to answer every question is well illustrated
in the replies of the defectives of Group IV. The mean number of
questions answered reaches the high value of 817 °/g, and in the
categories of suggestion and colour (which contain the greatest number
of questions) the values are 96'1 °l^ and 91*6 7o respectively. Even the
normal children of Group I reply to 77*0 7o of the total questions.
The most cautious children are the girls of Group II, who reply to
only 60*6 7o of the questions.

When the interrogatory was repeated after an interval of seven
weeks, there was a slight increase in the number of questions answered
by the normal children.

Accuracy. It will be remembered that over one-third of the narra-
tive dealt with the active aspect of the event; this category is also
prominent in answers given to the interrogatory. 840 °l^ of the replies

^ K. M. Dallenbach, in an article on "The Eelation of Memory Error to Time
Interval" (Psychol. Rev. 1913, xx. 323-337), concludes, as the results of experiments with
pictures, that the memory error increases with the time interval, very rapidly at first and
then more gradually as the time interval becomes greater. The percentage error after 45
days was found to be 22-4 for one picture and 18*1 for another. His methods differ
considerably from ours.

T. H. Pear and Stanley Wyatt

403

404 The Testimony of Children

of the normal children in this category are correct ; in the case of the
girls of Group II the accuracy reaches the exceedingly high value of
93*8 7o- III the mentally defectives only 61"5°/^ were correct for this
category. Thus the importance of movement in material for investigating
testimony receives additional emphasis.

This resemblance of the interrogatory to the narrative is further
increased in the category of items. Three-quarters of the replies to the
questions relating to 'personal items' are correct; the answers to the
questions dealing with the remaining items are slightly more inaccurate.
These results again show that the attention of the children tends to be
focussed upon the personal factors of the event.

Throughout this investigation, a noticeable feature was the un-
reliability of the evidence in connexion with colours^. Only 56*7 "/^ of
the replies of the normal children are correct ; and the evidence of the
defectives is still more inaccurate (38'4 °j^). Unless the colours are
strikingly prominent, they are seldom noticed.

The evidence relating to the sequence of events was very unreliable ;
only one-quarter of the replies given being correct. It should be stated,
however, that the questions in this category dealt chiefly with the order
in which the articles were taken out of the bag, and as this process
was somewhat rapid, it evidently passed unnoticed by the children.
Still, the duration of this process occupied the greater part of the event ;
and yet the evidence upon it is almost worthless.

In all the categories, the replies of the defective children are less
accurate than those of the normal children ; the respective mean values
for these two classes being 62*0 °/^ and 51"0 7o- ^^ every category, some
components pass quite unperceived, whilst others are observed by
everyone.

The effect of repeating the questions after the interval of seven weeks
is to produce only a slight decrease in the accuracy of the replies. This
decrease amounts to 3 °l^ and 6 "/^ for the normal and defective children
respectively. Thus the increase in the number of questions answered is
accompanied by a decrease in the accuracy of the replies.

The Categories in detail.

A. Items. The items directly connected with A and B are
enumerated with greater accuracy than those of more remote connexion
with them. This tends to show that the children were primarily in-
terested in the pereons and their possessions. It is curious to note the
' Also noted by Dallenbach, op. cit., Whipple, op. eit. p. 308.

T. H. Pear and Stanley Wyatt 405

gradual transition from certainty to uncertainty which occurs as the
object in question becomes less and less prominent. The following
questions and replies illustrate this point :

Questions

Correct

Incori

Had the lady a fur ?

59

8

Had the lady a mufif ?

58

9

Did the lady wear eyeglasses? ...

50

14

Had the lady a bag ?

41

17

Was the lady wearing a rose ? ...

40

22

Further, what is irrelevant to the whole situation, providing it is not
too incongruous, generally passes unnoticed. Thus the children were
shown a picture of a cat and a canary, upon which was an objectively
very prominent trade mark consisting of the letter >S^ on a diamond-
shaped blue background. This was perceived by six children only;
evidently the interest of the children was centred upon those aspects of
the picture which gave it its meaning.

None of the children succeeded in enumerating correctly the names
of all the articles which A took from the bag. With the exception of
the flag, flowers, statuette, and picture, the articles were just taken out
of the bag and immediately placed on the desk, and hence attracted the
attention of the children for a comparatively short time. This may
account for the fact that the articles were occasionally said to include
such objects as a cigar-holder, pouch, tobacco, string, scissors, letters,
kite, balloon, hair-tidies, handkerchief, looking-glass, mouth-organ, and
railway-guide.

It is interesting to note that though the children were generally

aware that there were initials on A's bag, they failed to notice what

the letters actually were.

The various ways in which the statuette was perceived produced

some interesting results. It was a representation of a curly-haired

shepherd boy holding a lamb, but it was described as such by only

eight subjects.

Others perceived it as a girl, a lady, a man, a boy, a doll, a soldier,

a musician, or as Britannia, Justice, Jesus and His Mother, Holy Mary,

an angel. Nelson, or Florence Nightingale. The figure was said to be'

holding a spear, a picture, a cross, a crook, a head, a vessel, a baby, a

jug, flowers, knitting, or a lamp.

Thus, when the objects are inattentively perceived, the individual

differences in interpretation are considerable.

406 The Testimony of Children

B. Colours. Usually a colour had to be intensely prominent if it
were to be accurately perceived. Thus the colour of ^'s muffler was
described by the normal children as white (16), green (10), blue (14),
grey (3), yellow (2), black (2), brown (2), and in Manchester as the
Manchester University colours : blue, green, and silver (3). The actual
colours present were blue and white. His tie was depicted as an object
of still more varied hues. On the other hand an article such as 5's
dress generally evoked unanimous replies ; 49 normal children gave it
correctly as blue and eight others stated that it was black. Whenever
there is any uncertainty as to the existence of an object, this un-
certainty is accentuated in the responses to the questions about its
colour.

It appears evident that the children often receive impressions of
' lightness ' or ' darkness,' and not of the actual colour. For instance,
B'b muff and fur were in reality brown in colour ; in the majority of
cases the colour was given as black, much less frequently as brown.
Only very occasionally was it given as grey or white. Often, though
an object was correctly perceived, its colour passed unnoticed. The
title of the picture, "The Cat and the Canary," was printed in large
red type. Forty-six of the normal children gave the correct title, but
only nine were able to name the colour of the letters. Other colours
given were: black (19), yellow (5), white (5), gold (3), blue (3), brown
(1), and green (1).

People are by no means agreed on the question of the reliability of
evidence upon the colour of hair. In the present investigation there is
much more agreement than difference. J.'s hair was decidedly fair;
the actual colours given were light (28), brown (16), golden (5), white
(3), red (2), grey (1), yellow (1), sandy'(2), silvery (1), and 'ginger' (1).
It should be noted that in this case the colour black was never men-
tioned, which again supports the contention that the normal children
generally receive a distinct impression of ' lightness ' or ' darkness ' as
the case may be. The defective children, however, show some variation
from the above situation. Sixteen of these children state that ^'s hair
was black, in fact a dark colour was given more frequently than a light
colour.

Thus many erroneous statements are made in connexion with the
colours of objects, and a careful consideration of the factors at work is
necessary before such evidence is accepted as correct. As a rule, in the
observation of such events, colours do not seem to be in the focus of
attention.

T. H. Pear and Stanley Wyato

407

C. Sizes. The defective children were entirely unable to estimate
the sizes of the components of the event. Their answers showed that
they had no conception of the dimensions of such objects as A's bag or
the picture, and in this respect they exhibited a remarkable difference
from the normal children. Consequently the following remarks relate
only to the normal group.

The following table gives the results in detail :

Length of bag

Breadth of bag

Height of bag

Height of man

Length of picture
Breadth of picture .
Height of statuette .
Length of book

Width of book

Length of flag

Width of flag

Group I

'

Group II

1 ■

Mean of

Number
of cases

Mean

M.V.

Number

Mean

M.V.

1 1 and II

of cases

i

ins.

ins. 1

ins.

ins.

1 ins.

39

25-2

5-7 '

21

19-2

5-5

1 22 0

36

12-3

2-4

22

10-4

3-8

11-3

35

131

3-6 1

17

10-5

3-7

1 11-8

37

65

5 0 1

13

62

7 0

63-5

39

36-2

7-2

16

311

90

33-6

39

24-3

61 i

13

22-2

51

23-2

38

91

•1-8

21

8-5

2-4

8-8

26

8-4

1-9

11 .

8-9

2 0

8-7

25

5-3

•8

10

5-7

1-9

a "5

33

7-6

2-1

17

7-2

2-2

7-4

33

4-8

1-3

16

1 4-7

1-2

4-7

Correct
value

ins.

18-0

12 0

8-0

68-5

36 0

24 0

9 0

9-0

6 0

9 0

6 0

D. Duration of the Event. The children were asked to estimate
the length of time over which the event lasted, and the question gave
rise to some very interesting results. The actual time occupied by the
event varied from 2 mins. 10 sees, to 2 mins. 35 sees, at the different
schools; the following table gives the results obtained from Groups I
and II.

Group I

Group II

Mean

M.V.

Mean

M.V.

Ist test ......

2nd test

8-1 mins.
9-3 „

3-2
2-5

i 11-0 mins.
10*6 mins.

3-0
2-5

. . As in the previous case, the defective children evinced no idea of
the length of the time-interval; and, in the few instances in which
estimations were expressed, they showed no resemblance whatever to
the correct value. Estimations ranging from half-an-hour to an hour
were very common.

408 The Teatwiony of Children

In every case, the time-interval was enormously over-estimated* ;
an example of the familiar case of the 'filled intervals' After a period
of seven weeks the interval is still further over-estimated by the
children of Group I, but the estimations of the girls of Group II are
slightly less erroneous. This interval between the tests also tends to
produce more agreement between the results, since there is a consider-
able reduction in the mean variation.

The over-estimation by the girls of Group II is greater than that
by the children of Group I. It will be remembered that the narrative
of the former group was longer and included more details than that of
the latter, hence the difference in the over-estimation of the time-
interval may be due to the fact that the interval appeared to be more
' filled ' to Group II than to Group I.

A separate treatment of the time-estimations of the boys and girls

of Group I reveals a slight superiority in accuracy of the boys over the

girls. The actual values are :

Mean m.v.

Boys (21) ... 7*7 mins. 3-1 mins.

Girls (20) ... • 8-5 „ 3-3 „

E. Sequence of the Event. The evidence in this category is most
unreliable ; the replies of the defective children being much less satis-
factory than those of the normal children. The majority of the children
state that the articles were taken out in the following order ^ :

1

2

3

4 5

6

7

8

9

10

Group I.

Pr.

Pe.

A'«.

Fg. Fs.

Ft.

Ft.

Fs.

Ft.

Pt.

,, II.

S.

S.

Fs.

Ft. Ft.

Ft.

Fg.

Fg-

Ft.

Ft.

„ m.

S.

S.

Ft.

Fs. Ks.

Ft.

Ks.

Ft.

Fs.

Fs.

„ IV.

Fg-

S.

Pe.

Ft. Fr.

Ke.

B.

Fe.

Ft.

Pt.

(Incorrect replies are

italicised.)

= Picture ;

Pe.=

Pipe;

Ks.=]

Keys; Fg. = Flag; Fs.= Flowers;

S. = S

tatuel

B.=

= Book;

; Pr. = Paper;

; Ke.=

Knife.

Pr.

Thus the table consists almost entirely of the Statuette, Flowers,
Flag, and Picture. These were the objects which received the longest

* Similar cases of over-estimation have been recorded by J. Dauber: "Die Gleichfor-
migkeit des psychischen Geschehens und die Zeugenaussagen," Fortschritte d. Psychol.
1912, I. 2, S. 102; by W. Stern, Beitr. z. Psychol, d. Attssage, 1904, ii. 1, S. 32 fif. and 57;
by R. Oppenheim, ibid. 1905, ii. 3, 75 ff. ; by 0. Lipmann, Ztsch.f. angew. Psychol. 1911,
IV. S. 312 f. and also by H. Breukink, "Ueber die Erziehbarkeit der Aussage," Ztsch. f.
angew. Psychol. 1909, 32-87.

2 When an interval is occupied by auditory, visual, or tactile stimuli, this filled
interval appears longer than an • empty ' interval of the same length, unless the interval is
very long.

^ The actual order is given on p. 392.

T. H. Pear and Stanley Wyatt

409

exposures, and the children's attention was dominated by them. Hence
as regards the sequence of events, only those objects which receive
additional emphasis are noted with an appreciable degree of accuracy.

Comparison of the replies relating to A and B respectively.

In the following table, the upper row of each category represents
the mean value per subject. The lower row gives this value as a
percentage of the total number possible in the columns headed
' Number answered,' and as a percentage of the number answered in
the columns headed 'Number correct.'

First test

Second test

Grcup

Number answered

Number correct

Number answered

Number correct

A

B

A

B

A

B

A

B

I I

II 1
Mean |

26-7
80-9
22 1
66-9
24-4
73-9

24-4
81-3
210
70-0
22-7
75-6

18-4
68-8
16 1
75-3
17-3
72-0

15-6
63-6
15 0
72 1
15-2
67-8

25-4
76-8
23-2
70-3
24-3
73-6

25 0
83-4
22-9
76-3
24-0
79-8

17-4
68-5
16-9
72-8
17-2
70-6

16-0
63-8
16-5
72 0
16-2
67-9

m 1

IV 1
Mean j

17-2
52 1
22-0
66-6
19-6
59-3

23 1
77-0
26-2
87-3
24-6
82-1

10-1
58-7
12-5
56-9
11-3
57-8

11-9
51-5
11-9
45-4
11-9
48-4

18 0
54-5
21-3
64-5
19-6
69-5

23-8
79-3
26-6
88-7
25-2
84-0

10-6
58-9
11-6
54-4
11-1
56-6

12-7
53-4
12-4
46-6
12-6
50 0

Thus the questions relating to A are answered, on the whole, more
frequently, but less accurately, than those referring to B, and hence we
are justified in assuming that the children were more interested in A
and what he did, than in B and her actions. This is true for both the
normal and defective children. The relative inaccuracy of the defective
children is again prominent in this aspect of the investigation.

The results of the boys and girls of the different groups were also
separately tabulated, but there was no evidence of any appreciable sex
differences. There was a slight tendency for the boys to be more
accurate than the girls of the same group in the replies in connexion
with A.

J. of Psych. VI 27

410

TJie Testimony of Children

Suggestibility of the Subjects.

Of the 150 questions which constituted the interrogatory, 36 were
framed to test the suggestibility of the subjects. The questions dealt
with all the more important aspects of the event, and were of varying
degrees of subtlety. The results obtained are given in the following
table :

First test

Second test

I

II

Mean

III

IV

Mean

I II

Mean

m 1 IV Mean

Number )
answered (A) j

Number )
correct (C) )

28-6
15-4

21-4
16-0

24^9
15 •?

27-3
12-6

34^6
12 3

30^6
12-5

27^9
11^6

23 •?
11^9

25^8
11^8

29-2
10^7

32 3 30 7
10 2 i 10 5

Accuracy = —

•54

•75

•64

1
•45

•35

•40

•42

•50

•46

•37

•31 34

Thus, as a rule, the children are susceptible to suggestions; the
normal children to the extent of 36 "/^ , and the defectives to the extent
of 60 °/o. Hence the latter are much more liable to suggestion than
the former, they exercise very little judgment, and their attitude is
decidedly uncritical. In the second test there is a noticeable increase
in the susceptibility to suggestion, the percentage value increasing from
36 to 54 in the case of the normal children. Consequently the ability
to resist suggestion depends in part upon the freshness of memory of
the material to which the suggestions relate. There is no evidence to
show that this difference in suggestibility between Groups I and II is
due to the presence of the two sexes in Group I. The children of
Group I also give more varied replies than the girls of Group II. Thus,
when asked the question : "On which part of ^'s face was there a cut? "
the girls of Group II, who answered the question, gave : none (7), near
eye (1), and on neck (2) ; but the replies of the children of Group I
were as follow : none (4), chin (3), right cheek (8), left cheek (5), near
eye (1), forehead (2), neck (1), and over eye (1).

T. H. Pear and Stanley Wyatt 411

Correlations between resistance to suggestion, intelligence and age.

Coefficients of correlation were calculated between the accuracy of
the replies to the suggestive questions and general intelligence*
(teacher's estimate). The coefficients for Groups I and II are "03 and
— '13 respectively, showing the absence of relation between intelligence
and suggestibility. However, there is a small positive correlation be-
tween age and ability to resist suggestion. For Group I the coefficient
is "24; hence, so far as the evidence warrants, the older the child the
less suggestible he tends to be.

Effect of previous experience. Knowledge of similar situations
previously encountered has a considerable influence upon the replies
given. Components of the event which are vaguely perceived, or not
perceived at all, are often interpreted or supplied according to the
manner in which they are most usually experienced. Thus, in almost
every case, the colour of the handkerchief in J.'s overcoat pocket was
said to be white (66). Other colours, rarely given, were pink (1),
brown (2), green (1), yellow (1), red (3), and blue (4). As no handker-
chief was to be seen it is evident that the children acted upon the
knowledge that the most usual colour of such an article is white.

Further, the relevancy of a situation determines in part the nature
of the replies given. J.'s walking stick was an ordinary cheap, light
yellow cane. A gold band on such a stick would be decidedly out of
place, hence we find the children stating that a silver band adorned the
stick, although the stick did not possess a band of any description.
Similarly, the general appearance of A suggested to the children that
he was worthy of a gold watch-chain, and not one of silver or of brass,
yet in reality his watch-chain could not be seen.

The effect of the range of the child's experience upon his evidence
is well illustrated in the answera to the question : " What was the name
of the newspaper?" The replies of the Manchester children include
The Guardian, City News, Courier, Evening Chronicle, and Evening
News. The Bolton children gave similar names, since most of the
papers are common to the two towns. In the replies of the Liverpool
children are to be found the names of such local papers as The Echo,
Express, and Mercury.

1 By means of the product-moment formula.

27—2

412 The Testimonij of Children

Comparison of the Suggestiveness of the Questions.

By ascertaining the number of correct and incorrect replies given to
the suggestive questions, it is possible to arrange these questions in the
order of their suggestiveness. Such an arrangement shows that the
most suggestive questions refer to those components of the event which
are not prominent or which are indefinitely perceived, and particularly
to those suggested components which might be expected to exist.
The cases of the silver band on the walking-stick and the gold watch-
chain are examples.

When asked if the figure of the statuette was wearing sandals or
boots, 21 normal children gave sandals, 11 boots, and only one correctly
as nothing at all. Very probably the alternative form of the question
was instrumental in directing the children's thoughts away from the
possibility of the figure being barefoot, and in leading them to pro-
nounce in favour of either sandals or boots.

There was an interesting difference in the replies to the two
consecutive questions, " Was there some blue ribbon round the cat's
neck ? " and " Was there a bell on the ribbon ? " The former was much
more suggestive than the latter, which again may be due in part to the
wording of the question. Since blue ribbon was mentioned, it tended
to suggest that there was actually ribbon present but that the only
doubtful point was its colour. The omission of the word ' blue ' from
the question would probably decrease its suggestiveness.

The least suggestive questions are those which refer to the most
prominent components of the event or to those parts which are un-
common or irrelevant. Hence a question such as : " Had A a white
waistcoat ? " was answered correctly every time by the normal children,
but 11 of the defectives replied in the affirmative. Thus, under normal
conditions, it seems probable that a white waistcoat would have been
noticed by the normal children. Again, when asked if A had a mous-
tache, the normal children answer correctly in the negative on every
occasion, but 15 of the defectives give affirmative replies.

Correlations. Coefficients of correlation^ between the order of
suggestiveness of the questions for the different groups tested were
evaluated, and found to be as follows :

Groups

r

I and II

•48

I and III

•64

I and IV

•87

II and III

•87

II and IV

•73

III and IV

•79

1 Evaluated by means of the product-moment formula.

T. H. Pear and Stanlf.y Wyatt 413

Thus there is a high, positive correlation between the order of sug-
gestibility of the different questions for the different groups, showing
that the suggestiveness of each question remains fairly constant from
group to group.

In connexion with Groups I and II, the children were classified in
the order of their general intelligence, and these intelligence classifica-
tions were correlated with (a) range of narrative, and (6) number correct
in interrogatory. The coefficients of correlation between the order of
intelligence and the range of narrative were '45 and "16 for Groups I
and II respectively. Since the children of Group I varied in age from
11 to 14, whilst those of Group II were of approximately the same age,
it seemed possible that the difference in the values obtained for Groups I
and II may have been due to the influence of this age-factor.

Accordingly, in the case of Group I, coefficients of correlation were
obtained between the age of the children and the range ^ of the narrative,
and also between the former and their order of intelligence. These were
found to be "32 and "46 respectively. By means of the method of partial
correlation it was now possible to find the closeness of the relationship
between the range of the narrative and the intelligence classification
independently of their common relationship to the age of the children.
This value worked out at '36. When * intelligence ' was made constant,
the correlation between age and range of narrative fell fi-om '32 to '14.
Thus there is a tendency for the more intelligent children to give the
longer narrative, and this tendency is slightly more dependent upon the
intelligence than upon the age of the children.

Coefficients of correlation were also worked out between the number

correct and the number incorrect in the interrogatory, and were found

to be as follow :

I II III IV

Isttest -21 -67 -27 -57

2nd test -11 -65 -18 --12

These values show that there is, on the whole, a slight relation
between the number of correct and incorrect replies in the case of the
normal children (Groups I and II). The number of incorrect replies is
roughly proportional to the number of correct replies. This uniformity,
however, does not apply to the answers of the defective children.
Hence it seems that the individual fluctuations are greater in a group

^ In the narrative the accuracy is so great that marks given for it are practically
identical with those given for 'range.' The above correlations therefore may be taken as
representing the relation of either 'range' or 'accuracy' to the age and intelligence of the
children.

414 The Testimony of Children

of defective children than in a group of normal children. It will be
noticed that the steadiness and uniformity of the girls of Group II
furnish further evidence in this connexion.

'Reconstruction' of the Event.

The event may be ' reconstructed ' ' from the replies to the questions
which constitute the interrogatory, since these questions refer to all the
necessary components of the event. The answers to each question are
usually varied, hence the reconstruction is made from the 'majority
replies,' that is, from those replies which occur most frequently in con-
nexion with any particular item. In the present case, reconstructions
have been made from the replies of the children of Group II (normal)
and from those of Group V (defective). The incorrect parts are printed
in italics, and the numbers in brackets under each part represent the
extent of the correctness or incon-ectness as a percentage of the total
number of answers given to that question.

Group II {normal).

A lady and gentleman entered the room. The gentleman was

carrying a hag in his left hand, a hat in his right hand, and a stick
(57) • (31)

over his left arm. He had on a fawn coat, a blue tie, and was wearing
(54) (78) (60)

a black or white muffler. His clothes were blue, and his hat was of the

(66) (36) (89)

bowler type. He placed his hat on the piano and shook hands with

(56) (82)

the teacher. He had light hair, which was parted at the side. He had
(57) _ (90)

a ring on his finger, and his watch-chain was made of gold. The colour
(75) (93)

of his stick was brown, the shape of the handle was crooked, and a silver

(67) (96)

band encircled the stick. A white handkerchief was to be seen in the
(90) (88)

gentleman's over-coat pocket. The length of the bag was 19 ins., the

breadth one foot, and the height 10 ins. This he placed on the desk.

(68)

The lady was wearing a blue dress, a white blouse with lace over it, and
(71) (55) (63)

a black hat, one side of which was turned up and which contained a
(59) (66) (33)

^ Following Dauber's procedure (op. cit.}.

T. H. Pear and Stanlfa^ Wyatt 415

black feather. She carried a fur which was unfastened, a black muff,
(73) (100) (72)

a black bag and brown gloves. On the lady's uiufF were some flowers,
(70) (70) (62) .

she took these off, held them for a short time, and then replaced them.
(85) (82)

The lady had black hair, was wearing a rose, and was a little taller than
(62) (52) (53)

tlie gentleman.

On the bag were some letters, these were M.P.L., and on the left of

(81)

the bag was the lettei' M. The label on the bag luas white, and the bag
(44) (56) (100)

was brown. The first thing that the gentleman pulled out of the bag was

(37)
a statuette, the second a statuette, the third flowers, the fourth a picture,

(53) (18) (30)

the fifth a picture, the sixth a picture, the seventh a picture, the eighth a

(34) (30) (30) (25)

flag, the ninth a picture, and the tenth a picture. They were in the
(80) (82)

room for eleven minutes. The length of the picture was 31 ins., the

width 22 ins., and it had no border round the edge. The picture showed

(67)
a cat sitting watching a canary in a cage. The breast, paws, and face
(93) (75)

of the cat were white, the back was grey, and only two of the paws were

(38) (73)

visible. The title of the picture was " The Cat and the Canary," and

(95)
the letters composing the title ivere yellow. On the picture were also some

(38) (62)

red flowers on a table. The statuette represented a girl holding a basket.

(42)

The colour of the girl's hair was golden, her cloak was white with a
(53) (53)

yellow lining. She was also wearing sandals.

(73) (80)

The gentleman pulled out a bunch of roses, three in number, vrith a
(46) (45)

yellow flower in the middle. The book which he pulled out was red,

(33) (93)

and measured 8*9 ins. by 5*7 ins. The flag was 7'2 ins. by 4*7 ins., and

had on it a yellow cross on a yellow background. The paper was the
(87) (32) (46)

' Citizen,' and was not opened. The pipe was brown, curved, and had
(83) (69) (35)

a black mouthpiece.
(43)

416 The Testimony of Children

Group V {defectives).

A lady and gentleman entered the room. The gentleman was

carrying a bag in his right hand, a hat in his left hand, and a stick
(48) (32)

over his left arm. He had on a black coat, a blue tie, and was wearing
(50) (30) (35) (65)

a white muffler which he kept on whilst in the room. His clothes were

(93) (34)

black and his hat was of the bowler type. He placed his hat on the

(65) (39)

piano and shook hands with the teacher. His hair was black and was
(64) (38)

parted at the side and on his cheek was a cut. He wore eyeglasses and
(93) (87) (38)

had a ring on his finger, and his watch-chain was made of silver. The
(84) (54)

colour of his stick was brown, the shape of the handle was round, and on
(68) (85)

the stick was a silver band. A white handkerchief was to be seen in the
(41) (88)

gentleman's overcoat pocket, and there was a flower in his coat.

(65)
The lady was wearing a blue dress, a white blouse, with lace over it,
(67) (75) (67)

and a blue hat, one side of which was turned up, and which contained a
(50) (55) (71)

black feather. She carried a fur which was unfastened, a muff, gloves,
(82) (64) (83) (95)

but no bag. There was a dog's head on the muff, she took nothing off the

(50) (52) (68)

muff, and she replaced what she took off. The colour of the lady's hair

(57) (48)

was black, her fur and muff were black and her bag was black, her gloves

(66) (65)

were also black. She was wearing a rose, she carried an umbrella, shook
(50) (81) (55)

hands with the teacher, and put some roses on the piano. She afterwards

(52) (40)

picked up the flowers. She also picked up a book and put it on the piano.

(28) (60)

She had a watch on her wrist.
(83)

On the bag, which was brown, were the letters L.M.R. and on the

(93)

right of the bag was the letter M. There was also a white label on the
(53) (45)

T. H. Pear and Stanley Wyatt 417

bag. The gentleman pulled out of the bag first a statuette, second a

(38) (38)

statuette, third flowers, fourth a picture, fifth flowers, sixth a knife, seventh
(24) (36) (33) (20) " (25)

a book, eighth a picture, ninth a picture, and tenth a picture. The
(31) (46) (51)

picture had a black border round the edge, and showed a cat sitting
(30) (89)

looking at a bird. The cat had a white breast, face, and paws, a black

(58) (54) (48) (46)

ba^k, and part of the back was yellow. Only two of the paws were visible.

(27) (55)

The title of the picture was ' The Two Friends,' and the letters which
(19)

composed the title were black. The bird was yellow and was still. At

(44) (98)

the side of the picture were some white flowere, there were also some
(63)

red flowers on the table. There was some blue Hbbon around the cat's
(34) (89)

neck, on which was a bell. The owner of the cat was looking through
(83) (73)

the window. The statuette represented a girl with brown hair, and she

(39) ■

wa,8 wearing a white cloak with a white lining. She also wore boots, and
(50) (69) (70)

held a baby in her arms. The flowers which the gentleman took out of

(33) (44)

the bag were red, 5 in number, and there was also a blue flower in the

(67)

bunch. The book was red and contained a white paper. The flag was

(65) (65) (39)

red and had on it a yellow cross ; the flagpole was made of wood. The
(44) (53)

paper was called the ' Echo,' and was not opened.

(21) (65)

VII. Conclusions.

The results of these experiments, when the findings of previous
investigators are taken into account, support the following conclusions :

1. As a rule, the evidence of children is reliable only when it is
given spontaneously. In such a case it is decidedly valuable and worthy
of the utmost consideration.

The reports of the normal children are distinctly superior, both
qualitatively and quantitatively, to those of the mentally defectives.

418 The Testimony/ of Cht'Idren

2. The narrative is chiefly devoted to the description of actions,
and the enumeration of items and positions of objects. Other categories
are relatively ignored. The interest of the children in movement
illustrates very clearly one of the defects of a picture as material for
a testimony experiment. Colours, sizes, and shapes are mentioned with
greater frequency by normal than by mentally defective children.

3. The interrogated evidence is much less reliable than the
spontaneous report. More than one-third of the replies of the normal
children and more than one-half of those of the mentally defective are
incorrect. Only those questions which relate to the chief and out-
standing features of the event are answered with a high degree of
accuracy. As in the narrative, the most reliable sections of the
interrogatory are those which refer to actions and items.

4. A repetition of the questions after an interval of seven weeks
showed, contrary to expectation, that this period of time had little effect
on the children's memory of the event. The accuracy of the second
series of replies was only slightly less than that of the first.

5. The knowledge of past experiences of similar situations has
a considerable influence upon the replies given. Components of the
event which are vaguely perceived or not perceived at all, are often
interpreted or supplied according to the mode in which they are usually
experienced. The relevancy of the situation and the range of the child's
experience also determine in part the nature of the replies given.

6. Children's evidence bearing on the colours of objects is very
unreliable. Usually colours which are not very prominent are very
inaccurately described. Whenever there is any uncertainty about the
existence of an object, the uncertainty is accentuated in the responses
to the question about its colour. The mentally defective children err
to a greater extent than the normal children in this respect.

7. In every case the duration of the event was enormously over-
estimated ; there seems to be a definite connexion between the number
of details observed and the amount of the over-estimation, corresponding
to the familiar illusion of the estimation of ' filled ' intervals.

8. Most of the children are susceptible to suggestion, the suscepti-
bility increasing as the interval between the event and the process of
recall becomes greater. There is no congelation between general in-
telligence and suggestibility, but there is a small correlation between
resistance to suggestion and the age of the subjects.

The most suggestive questions are those which refer to the less
important or more obscure features of the event, and especially to those

T. H. Pear and Stanley Wyatt 419

suggested components which might be expected to exist. On the other
hand, those questions are least suggestive which refer to the most
prominent, uncommon, or irrelevant components of the event.

The order of suggestiveness of the questions is approximately the
same for each of the groups tested.

9. There is no evidence to show that the testimony of normal
children of the age tested is affected by sex differences, although that
of the mentally defectives shows signs of being influenced by this factor.

10. The methods of teaching adopted in different schools for normal
children may exert a considerable effect on the testimony of the children
in these schools. In the school in which ' free discipline ' was a
prominent feature the individual differences between the children's
performances were more marked than in the school in which the
discipline was more rigid.

11. The testimony of the mentally defective diffei"s from that of the
normal children in the following respects :

(a) The narratives of the mentally defective children are frag-
mentary and disconnected, the event is described independently of its
chronological order. The range is shorter and the deviations from the
actual event are greater than in the case of the normal children. The
evidence upon the categories of colours, shapes and sizes is very meagre.

(b) In the interrogatory, the number of incorrect replies given by
the defective children exceeds the number of correct answers. The
reverse obtains in the case of the normal children. The replies of the
defectives are less accurate and more varied than those of the normal
children ; the level of accuracy is lower throughout all the categories.

(c) The defective children generally have no conception of the
absolute magnitude of objects, and their attempts to estimate the
duration of the event are equally futile.

(d) They are less able to resist suggestion than normal children.

(e) The defective children cannot resist the temptation to answer
every question. It is a source of weakness in their evidence, and in this
respect generally differentiates them from the normal children.

(/) The environment and past experience of the defective
children have a greater influence upon their evidence than is the case
with the normal children.

(^) In all the categories the level of accuracy attained by the
mentally defectives is much lower than that of the normal children.

(Mamiscript received 24 November 1913.)

THE CONDITIONS WHICH AROUSE MENTAL
IMAGES IN THOUGHT\

By CHARLES FOX.

{From the Psychological Laboratory, University of Cambridge.)

I.

§ 1. The nature of the experiments.
§ 2. The m.ethod.

II.

§ 1. Analysis of the mathematical group.

§ 2. The results.

§ 3. Analysis of the historical group.

§ 4. The results.

§ 5. Analysis of the grammatical group.

§ 6. The results.

III.

General conclusions.

I.

§ 1. The original purpose of the following experiments was to
demonstrate to the subjects who took part in them the importance
of imageless thought, and to show them by their own introspection
the distinction between a mental image and the meaning of which
the image is merely the vehicle. The subjects were told that they were
to investigate the existence and importance of thought without images,
and to try to find out the content of such thinking. They were also
told to try to distinguish, as far as they could, between the thinking act
and the thought.

There were fifteen subjects, eleven men and four women, and by a
preliminary experiment.it was found that they were all possessed of
fairly strong mental imagery. Four were strongly visile, two strongly

1 Read before Section I (Subsection of Psychology), at the Meeting of the British
Association for the Advancement of Science, Birmingham, 1913.

Charles Fox 421

audile, one strongly motile. The others belonged to no single pre-
dominant type but had strong imagery of a mixed kind ; three being
auditory-visiles, three auditory-motiles and two motor-visiles. It is
interesting to observe that the four women belonged to the motor or
auditory types and were poor visualisers. One was strongly motile, one
was strongly audile and two were motor-audiles. With four exceptions
the subjects were post-graduates, and all had previous practice in intro-
spection in connexion with some earlier experiments, so that they all
knew what was required of them.

§ 2. Twelve statements were selected, four involving mathematical
conceptions, four historical, three grammatical, and one was a couple of
lines from Milton. With ten of the statements the fifteen subjects were
tested, with the remaining two only thirteen.

The subjects were told to record on a sheet of paper everything they
could discover by introspection after each statement had been read twice
by the experimenter. They were told to put down everything, however
unimportant it appeared to them, e.g. if they were conscious that during
the act of thought they felt muscular strain or tension, or if they were
conscious that their minds took certain ' directions.' In fact all details,
whether mental or physical, were to be noted. They were also told to
put down first of all whether they realised the meaning of the statement
read to them ; and as soon as the meaning was realised the process of
thought which had led to its realisation. If possible they were to state
w^hat the realisation consisted of; and whether it involved mental
images or not. The fact of agreement or disagreement with the
statement was said by the experimenter to be immaterial ; the im-
portant thing was the realisation of the significance of the statement
and the imagery, if any, aroused. In cases where images did arise they
were instructed to state whether the realisation of meaning preceded or
succeeded the occurrence of the mental image.

These instructions were only fully carried out as regards that part of
them which related to the realisation of meaning and to the occurrence
of images. An examination of the records revealed, quite unconsciously
on the part of the subjects, the conditions under which mental imagery
arises. The subjects stated definitely in several instances why the
images came and what part they played in enabling them to realise
the meaning. They had no preliminary theory on the subject, but
simply recorded their introspective observations.

The statements were not dictated in the order in which they are
here analysed ; to secure variety and freedom of thought and to maintain

422 Conditions which arouse Mental Images in Thought

interest in the experiment, the statements belonging to different groups
were interspersed, so that, as a rule, two of a like kind did not come
together.

II.

I 1. The propositions in the first group were :

i. (a). The whole is greater than the part.

i. (6). The whole is equal to the sum of its parts.

i. (c). If equals are added to equals the wholes ar-e equal.

i. (d). If unequals are added to unequals the wholes are unequal.

The following is an analysis of the results obtained after classifying
the introspective records :

i. (a). Four subjects realised the meaning before the phrase was
finished, three without any mental image.

Two recognised the phrase first, and the meaning later. Of these
one had verbal images ; the other no image, but a ' sense of its truth.'

Four realised the meaning first, and images appeared in three cases
after realisation.

Five subjects realised the meaning by the aid of visual images.

Thus five subjects obtained no images at all during the thought
process ; and in three other cases the images which came played no
part in securing assent to the proposition, as they appeared afterwards.
Of those who obtained images some got pictures of geometry books used
at school, or of divided circles, or of solid objects such as spheres divided
up. The phrase had for some subjects obviously acquired a meaning
which was reinstated in consciousness immediately one began to utter it.

i. (6). Five subjects realised the meaning without images, but two
of them obtained images later. One of the five disagreed with the
proposition on the ground that it was not true of some conceptions,
e.g. the soul; another obtained the meaning thus, "I conceived the
whole as having parts, and if so the sum of the parts must give the
whole." Both these subjects are students of moral science.

Seven subjects realised the meaning by the help of visual images of
circles, squares, triangles or solids divided up.

Of the remaining three subjects, one obtained an image of the series
l + i + i+..-3o = 2, then an image of a dissected disc put together ;
another obtained a very clear verbal image which was "so strong at
first that I did not grasp the significance of the statement, then I
assented immediately"; the third realised the significance before the

Charles Fox 423

completion of the sentence and had an image of a solid figure. Another
subject, too, realised the meaning in anticipation at the word ' sum.'

Thus only three subjects failed to receive any image during the
whole process.

i. (c). In this case nine subjects realised the meaning immediately
without the help of mental images. Four others realised it by the help
of verbal images, and the remaining two by the help of images of
geometrical figures. Both these last are very slow in grasping ideas.
Four subjects agreed to the statement in anticipation before it
was finished. Several others said that assent was instantaneous. The
proposition, owing to its occurrence in geometry books, had for them
an acquired meaning which was realised immediately.

i. (d). This statement is obviously not necessarily true, and was
introduced in order to see by what means the falsity was realised.

Only two subjects realised the falsity immediately, and neither of
them used imageiy for the purpose.

Six subjects suspended judgment before realising that it was false ;
the realisation taking place in all cases by means of images.
The remaining seven made use of images.

Of these last, two disagreed with the proposition and used visual
images ; in one case these were very clear and vivid.

Two had doubts about its validity, but on the whole thought it was
true ; these had images.

Two were very hazy about the meaning and obtained vague images ;
one of them said that " the possible meaning was clear without definite
images." Apparently, then, the images that he got later were the result
of deliberation.

One had a visual image of lines in a book. He had agreed to the
proposition at first without using images and then the possibility of its
falsity occurred to him and the visual image came. This case is especially
noteworthy as this subject had not received images in any of the three
former cases.

The important feature of experiment i. (d) lies in the fact that only
two subjects failed to get images, and these were the only cases in which
the falsity of the proposition was realised at once.

§ 2. Several conclusions emerge as the result of the foregoing
analyses of the subjects' introspection. In the first place it is perfectly
clear that a considerable amount of thinking is entirely independent
of mental images. Of the 60 thought-processes of the 15 diflferent
subjects, 24 or 40 °/„ occurred without mental imagery. As the

424 Conditions which arouse Mental Images in Thought

statements selected would involve mental images of a very simple and
definite tjrpe, if they occurred, namely of simple geometrical figures ;
and as the frequent mention of images during the experiment would
in itself act as a suggestion to arouse images which would otherwise
not occur, it seems probable that under normal conditions of thinking
images would not arise in more than 50 "/^ of the cases \

In some cases strong imagery interferes with the act of thinking;
thus one subject in i. (6) said that " the image was so strong at first that
I did not grasp the significance of the statement." But in others,
especially where there is some difficulty, a strong image may aid the
thought-process; in i. {d) one subject had "a picture of clusters of
triangles and other geometrical figures, and decided that nothing
whatever could be stated of them." The image in this case obviously
plays the same part as a diagram does in a geometrical proof.

If we compare the statement i. {a) with i. (6), the former involves a
simpler conception, being easier consequently to grasp, and the experi-
ment shows that images were less fi'equent in i. (a) than in i. (6). This
result seems to be general. Images tend to appear if the realisation of
meaning is not at once clear, or if there is a delay or a struggle in
consciousness. Where the meaning is easily grasped or where assent
has been previously given there seems to be no tendency to embody the
thought in an image. Thought is, in these cases, carried on by meanings.
If the meaning is very clear to the subject the image, as we have seen,
may actually interfere with thought, since the subject tends to dwell on
the image to the exclusion of the meaning. These results are made
clearer by a comparison of i. (c) with i. (d), the former being much easier
to realise. In the case of i. (c) nine subjects grasped the meaning without
the aid of images, whereas in i. {d) only two subjects failed to receive
images. The increased difficulty of realising the meaning resulted in
thirteen subjects receiving images of some kind.

Again, suspension of judgment and doubt, both of which may be
regarded as instances of delay or struggle in consciousness, are conditions
which facilitate the emergence of mental images, as i. (c?) clearly shows.

The following introspection results bear out these conclusions. With
reference to the proposition, " If equals are added to equals the wholes
are equal," one subject said, " Of course they are ; assent came long
before there was any thought of criticising (examining ?) the statement."
Another said, " This was very easy and needed no consideration ; beyond

^ Out of the 176 cases examined 77 obtained the meaning without images.

Charles Fox 425

suggesting thoughts of arithmetic and geometry the statement was
accepted at once." There were no images here according to the subject.
A third said that she assented at once ; and anticipated tKe meaning
before the whole was read. The realisation of meaning was instantane-
ous and there were no images. The proposition i. (d), namely, " If
unequals are added to unequals, etc.," was taken before i. (c), i.e. " If
equals are added, etc.," and the introspective results of one subject are
sufficiently instructive to be quoted in full. With reference to i. (d) he
said, " Doesn't sound probable somehow. Rather a wrench necessary to
enable me to realise that you have two sets of things to work with.
Then I saw that it is a lie (sic) as a picture of heaps of plain wooden
bricks arose, thus :

At the statement i. (c) he said, " Sounds much more reasonable. Of
course it's true. No need to fetch the bricks out to prove that one
(i.e. no image at all)."

§ 3. The propositions in the second group were :

ii. (a). Whilst Britain was prospering under Roman rule the Roman
Empire itself was beginning to show signs of decay.

ii. (b). Every histoHcal event has a political cause.

ii. (c). The whole organization of society was once based upon the
system known as feudalism.

ii. (d). Mechanical inventions have had important effects upon the
social life of England.

An examination of the results of introspection yielded the following
analysis :

ii. (a). Four subjects had either very faint images or none at all.
One of these said that he deliberately concentrated his attention on the
meaning of what was uttered. This latter case is noteworthy, as the
subject in the first series of experiments was very rich in images. Two
of them found the statement easy to follow and assented readily.

Eleven subjects had visual images of maps, soldiers, school pictures,
Caesar's landing in Britain, etc. In seven of these cases there were
distinct indications of a conflict or movement backwards and forwards
in thought.

ii. (b). Seven subjects received no images. Of these, two assented
to the proposition immediately; but three others had difficulty in grasping

J. of Psych. VI 28

426 Conditions which arouse Mental Images in Thought

the meaning. Two of these latter three are moral science students who
are practised in thinking in abstract terms and the third has had some
philosophical training. This perhaps explains the unexpected result, as
these students are accustomed to the terms ' event' and 'cause' and have
acquired the habit of thinking of their meaning apart from images.

Seven subjects had images, some very faint. Four of these found
difficulty in deciding on the meaning ; one said explicitly that the image
was called up for this reason.

The remaining subject felt inclined at first to call up visual images
of historical events to test the truth of the statement, but combated this
by concentrating his attention on the bare meaning instead.

ii. (c). (13 subjects.) Four subjects had no images. One of them
disagreed at once and one decided to accept the statement at once.

Nine subjects received images; three of these were verbal images
and two were images of history books. One subject had to dispel the
image before he could seize the meaning. Five of them agreed with
the statement immediately.

ii. (d). (13 subjects.) In reading this sentence a distinct pause
was made after the words 'mechanical inventions.'

Three subjects had no images and in each case assent to the
proposition was immediate, without effort.

Ten subjects had images, nine visual and one auditory. Two of these
stated that the images were due to the pause in reading ; and in fact
the majority of the images were directly concerned with some form of
mechanical inventions. One of the subjects had a coloured verbal image
of the word ' social ' — light to dark yellow varied with green.

§ 4. Examination of the introspection records of this second group
reveals some further conditions which facilitate the occurrence of mental
imagery, and also confirms the conclusions previously reached.

Immediate or ready assent to a proposition, or ease in understanding,
both of which imply the free flow of thought, usually means that the
subject has no images in the focus of consciousness. But there are
exceptions to this which need further inquiry, for in ii. (c) five subjects,
who agreed to the statement readily, had images. If a subject makes a
deliberate attempt to concentrate his attention on -the meaning of a
statement he may succeed in suppressing images which would otherwise
occur.

In cases of conflict, or disagreement with a suggested statement,
which manifestly implies a conflict of some sort, images tend to appear.
To make use of a bold metaphor, the mental image seems to be the

Charles Fox 427

result of friction in thinking just as a spark may arise from the friction
of two hard bodies. This is shown very clearly in some of the intro-
spection records with reference to ii. (6). One of the subjects was struck
by the difficulty of defining ' historical event ' and he was aware of images
of Pilgrim Fathers, Fire of London, Balkan War, etc. He could not
assent to the proposition after considering these cases. As a rule this
subject does not receive images. Another subject said " Such a state-
ment cannot be accepted without investigation. Therefore one draws
images." He proceeded to get images of the Spanish Armada, the
English Revolution and Voyages of Discovery. Then having fully
realised the meaning of the statement he gave a qualified assent. A
third subject said that she could not understand the meaning of the
proposition although she had images, and then proceeded thus: "Further
attempts to make a meaning brought up a picture of someone speaking
in Parliament." Here the attempt to overcome a difficulty is assigned
as a reason for the development of the image. Another example of the
same process was given by a particularly careful subject who is very
conscientious in his statements, thus : " First impulse to be inclined to
agree. Then it flashed across me that there were cases in which it was
not true. For a moment or two I seemed to have an example of this
near at hand and strove in vain to find it. Then at last it flashed across
my mind as a faint picture." In this case we actually observe the birth
of the image in a presented difficulty that requires to be solved.

A further condition favourable to the arousal of mental images was
revealed by the ii. {d) series. Emphasis or a pause which constitutes
a break in the free flow of thinking is favourable to the production of
imagery. Thus one subject said " The pause after ' mechanical inven-
tions ' gave opportunity for a faint image of an engine." Another stated
that " After ' mechanical inventions ' the hesitation gave time to picture
such." This condition, too, seems to fit in with the general idea of a
conflict or struggle which is again shown to be favourable to the develop-
ment of imagery.

It was shown above that in certain cases a strong image may obstruct
the attempt to understand. Another instance of this is provided by the
experiment ii. (c). The painstaking subject recently referred to said
" Image of society under the shape of a mob. Image of feudalism as a
picture of a man doing allegiance to his liege lord. When I tried to
realise the significance of the statement it was twice obstructed ; at first
by the picture of my old history room at school, then by my history
book open at the page on feudalism. These images being dispelled the

28—2

428 Conditions which arouse Mental Images in Thought

thing was then first clearly taken in, though I think I had grasped it
faintly on hearing it for the first time."

It seems, then, that a mental image is due to an obstruction in the
free flow of ideas, so that for a thinking process to proceed to its proper
conclusion the attention must be concentrated on the meaning to the
exclusion of the image.

§ 5. The dictated statements in the third group were as follow :

iii. (a). All verbs that make a statement must be accompanied by
some noun.

iii. (6). Grammar is useless because we speak well without a know-
ledge of it.

iii. (c). You should never use a preposition to end a sentence with.

iii. (d). Laughing to teach the truth

What hinders? As some teachers give to boys
Junkets and knacks, that they may learn apace.

After classifying the results of introspection the following analysis
was obtained :

iii. (a). Six subjects received no images. Four of them realised
the meaning promptly and another deliberately concentrated his
attention on the meaning, another assented to the statement slowly.

Six subjects had verbal images. In all these cases the realisation
of the meaning was slow and laboured. In one case there was a conflict
in the subject's mind, in another a doubt which was settled by assent
without conviction.

Two subjects obtained visual images. In one case there was a
conflict and in the other a distinct search for examples.

One subject made no statement about images.

iii. (6). Eleven subjects obtained no images during the realisation of
the meaning of the statement, though in some cases images came later.
In all these cases the realisation was easy and in several it prompted a
train of reasoning.

Four subjects received images. One of these did not realise the
meaning of the statement at once, but another did. One subject had
an image of his college where this question was discussed.

iii. (c). Eight subjects realised the meaning without any images at
first, but two obtained images afterwards. Five of them realised it easily
and immediately or by having heard it previously ; in one case the
realisation was slow.

Seven subjects received images. Of these, one saw a sentence having
the preposition ' with ' clearly defined and another heard this word by

Charles Fox 429

an auditory image. Three of these subjects realised the rule first and
saw the absurdity of it later, whilst one apparently did not see the
absurdity of the rule.

iii. (d). The results of this series require a more detailed analysis.

Six subjects obtained no image at all; of these four understood the
meaning of the lines promptly and thoroughly, and two realised the
meaning slowly, in one case not thoroughly.

Three subjects obtained what may be described as an associative
image, namely an image not directly called up by the lines but evoked
by association with their meaning. In these cases the image was that
of a book on education in which a similar doctrine to that expressed in
the lines was discussed. Now those who had these images must have
realised the meaning before the images came, since such images depend
on understanding the meaning. We may therefore say that nine subjects
realised the significance of the passage without the aid of images ; seven
promptly and two slowly \

Six subjects obtained images before realising the meaning. Four of
them were visual images, one was verbal and one was a taste image of
junkets. In two of these subjects the realisation was slow, tv/o did not
understand the passage, but the remaining two realised its meaning
immediately.

§ 6. The results obtained from group iii. thus serve to confirm the
previous conclusions ; e.g. in iii. (d) we see that, on the whole, prompt
and thorough understanding coincides with the absence of images. The
passage was not very easy to follow at first and directly suggested several
images, but apparently those who concentrated their attention on the
meaning failed to get these images. On the other hand those who did
get images, as a rule realised the meaning slowly or not at all. This is
verified also by considering the results obtained from series iii. (a).
Similarly with regard to iii. (c), of the seven subjects who understood
the rule immediately or easily five obtained no images. Series iii. (6) in
this group of cases also shows that where the reasoning to a conclusion
takes place easily images, as a rule, do not occur.

' G. H. Betts iu bis admirable study of The Distribution and Function of Mental
Imagery (New York, 1909) says " our associative machinery may bring before the mind
many elements which have no function in the thought of the moment, but are only
incidents, by-products of the thought process" (p. 49).

430 Conditions which arouse Mental Images in Thought

III.

The question which the experiments have helped us to solve is
concerned with the part played by mental imagery in thought proper
as opposed to reverie or day dreaming. In the latter case images come
and go and we can no more explain the origin of the particular images
that arise than we can account for the things we see when our eyes are
open. Where, however, there is a definite end in view, a specific problem
to be solved an answer seems possible. For- this purpose it is necessary
to draw a distinction between relevant and adventitious images. In a
particular train of thought there may be a number of images present
which neither aid nor hinder its development, — adventitious images.
The conditions favourable to the awakening of relevant images, which
form an integral part of the thought process, appear to be the same in
the majority of cases. And these conditions have in many instances
been unmistakeably indicated by the subjects during the course of the
experiment. Their evidence is made more trustworthy by reason of the
fact that it was given spontaneously without any preconceived notions
on the matter. None of the subjects had any idea before the experiment
started of what the conditions were, nor were they consciously engaged
in discovering the conditions, but in answering quite different questions.

The experiments show that any delay or conflict in consciousness
is a favourable condition for arousing a relevant mental image, that is,
one that will in some way tend to help towards a cessation of the conflict.
All the other conditions which we have found to be suitable for stimu-
lating the production of mental images are reducible to this general
formula. Thus, conflict or disagreement with a suggested statement,
an attempt to overcome the difficulty of understanding a proposition,
suspension of judgment, doubt, emphasis or a pause, all have been shown
to produce mental images abundantly. And all of these are examples
of struggle or delay in thinking^

The experiments also show directly that the contrary set of
conditions are unfavourable to the production of images. Thorough
or immediate understanding, an easily grasped conception, ready assent
to a proposition, straightforward or unimpeded reasoning, are all cases
in which, as a general rule, images play no part. Further, concentration

^ This agrees with the conclusion reached by Betts {op, cit. 94) that images tend to
emerge "at points where our thinking is baffled," but his other conclusion that the
images at these points are mostly irrelevant is doubtful.

Charles Fox 431

of thought on meaning is unfavourable to the stimulation of mental
imagery, but this cannot be brought easily under the above general
formula.

Thus, whatever promotes the easy or unimpeded flow of thought is
unfavourable to the production of mental imagery and vice versa. The
law will perhaps be made clearer by comparing the stream of thought
to the flow of electricity in a conductor ; where the resistance is high,
heat becomes apparent, and where it is sufficiently increased, light
breaks out.

Possibly we have here an explanation of the fact that children have
richer and more vivid imagery than adults. For difficulty in under-
standing abstract ideas and relative inability to fix attention on meanings
are just the best conditions for arousing strong mental images^

'' After this paper was completed my attention was called to a work by Dr Aveling
On the Consciousness of tlie Universal and the Individual (Macmillan, 1912). Aveling
correctly maintains that thinking can take place with concepts alone as contents ; and
on p. 170 he rightly states that "if we stop to ask ourselves what an unfamiliar word
means, we generally discover that another word, or an image, is aroused as exemplifica-
tive of its meaning." This is but a particular case of our general law that a difficulty
will produce an image.

{Manuscript received 1 January 1914.)

ON CHANGES IN THE SPATIAL THRESHOLD
DURING A SITTING.

By GODFREY H. THOMSON,

Lecturer in Education, Armstrong College, Newcastle.

1. Object of the paper.

2. Description of the experiments.

3. The psychophysical methods used.

4. Catch errors.

5. Processes of calculation.

6. The raw results.

7. Corrections for personal differences and diurnal variation.

8. Probable -errors and significance of the results.

9. Influence of the experimenter.
10. Summary.

1. Object of the Paper.

In some previous experiments^ (carried out for another purpose),
the writer was led to suspect that during one sitting there was a
tendency for the spatial threshold at first to sink, and later to rise
again. A sitting in those experiments lasted about twelve minutes,
during which one hundred applications of the aesthesiometer were
made to the right forearm of the subject. At that time six subjects
were examined in all, but only in one case (subject No. 2, a man) could
the suspected tendency be considered to be clearly proved. The present
experiments have been carried out to test the point further, and have
confirmed it to a reasonable degree of probability.

1 G. H. Thomson, "Comparison of Psychophysical Methods," this Journal, 1912,
V. 233.

Godfrey H. Thomson 433

2. Description of thk Experiments.

The experiments were carried out with the same simple apparatus
and with the same precautions as were noted in the earlier series on
subject Q\ Some additional precautions, and certain changes in the
psychophysical methods used, will be noted presently. Subject 6 was
again available, and in addition three new subjects were examined,
numbered 7, 8, and 9, of whom 7 and 9 were women. Subject 8 was
the writer himself, and the experimenter in this case was No. 6.
For subject 7 and also for subject 9 both No. 6 and No. 8 acted as
experimenters, sometimes at alternate sittings, sometimes at alternate
portions of a sitting. For subject 6, No. 8 acted as experimenter. A
third person was always available as clerk.

Subjects 7 and 9 had no knowledge whatever of the purpose of the
experiments, nor had subject 6 in the first series ; but this subject had
some suspicion of it in the second and third series. Of course the
writer when acting as subject had to endeavour to avoid any bias which
his foreknowledge would bring. The data suggest that the endeavour
to be impartial resulted in going to the other extreme, for the expected
result is absent in the case of subject 8. It may be however that
this fact is due, not to the endeavour to avoid bias, but to the differ-
ence in sex; or there may be other reasons which will be suggested
later.

No subject was told during a sitting whether the answers given
were correct or not. At the end of a sitting however each subject
except No. 7 was told how he or she had done, especially how many
catch errors had been committed. No. 7 was given no information
whatever.

During the experiments on this subject, and during most of the
experiments on No. 6, the arrangements were such that not only
the subject but also the experimenter was ignorant whether the answers
were correct or incorrect. This was managed in the following way.
The experimenter, who sat on one side of a screen through which the
subject's arm projected, was provided with a list showing the order in
which the different applications of the aesthesiometer were to be made,-
and after warning the subject that he was going to begin, he pro-
ceeded through this table at a regular rate fixed by the unobtrusive
ticks of a clock. The aesthesiometer was thus always applied at the

' Op. cit. 214 ete.

434 Changes in the Sjyatial Threshold dicing a Sitting

same intervals and for the same length of time. The subject com-
municated her judgments one, two, or doubtful to the clerk by signs
invisible to the experimenter, who was thus kept in entire ignorance of
how the sitting was turning out. The writer has not however detected
any differences between those series of experiments in which this
practice was followed and those in which the judgments were com-
municated by word of mouth and were therefore heard by the
experimenter.

3. The Psychophysical Methods used.

By the ' method ' used is meant the type of sequence according to
which the various stimuli were presented to the subject. The 'process'
of calculation afterwards applied to the data is another and independent
question which is discussed on a later page.

The method used in the case of subject 6 was for the sake of
continuity the same as that employed in the previous experiments^
where it is more fully described, viz. the Method of Non-Consecutive
Groups. The distances used for the aesthesiometer for this subject
varied by half-centimetres from five centimetres downwards, and were
always constant. A group consisted of five applications of say three
centimetres, mixed with five catches in which only one point was
presented. The groups followed one another in a chance order which
was varied in a cyclic manner from day to day and was altered
in toto at the commencement of every new series of ten sittings.
The sitting proper consisted of ten groups, and therefore of a hundred
applications of the instrument. In the case of this subject only, each
sitting proper was preceded by a preliminary group, always at three
centimetres. The other three subjects however were given no such
daily preliminary practice.

With subjects 7 and 9 a method was followed which was more

suitable for the purpose here in view. This method may be defined, in

terms of the nomenclature suggested elsewhere by the writer^ as the

Method of Right and Wrong Cases with Catches. In any one sitting

five stimuli differing from one another by steps of one centimetre were

used, in addition to the catch stimulus consisting of one point only,

or zero distance. For example in a certain sitting the stimuli used

might be

1, 2, 3, 4, 5 cms.

» Op. cit. 204, 214 fif. '^ Op. cit. 204.

Godfrey H. Thomson

435

Table 1. Method of Non-Consecutive Groups. Subject 6.
February 2ith, 1913, 8.21 a.m.

Pre-
liminary

3 cms.

i
1 cm.

ii
4^ cms.

iii
2^ cms.

iv
^ cm.

V

1^ cms.

vi
4 cms.

vii
5 cms.

viii
2 cms.

ix
'6^ cms.

x
3 cms.

Q

A

Q

A

Q

^

Q

i4

Q

A

Q

A

Q

^

Q

^

Q

A

Q

A

Q

^

2
1
2
2
1
1
2
1
1
2

w

w
w

w

w

2
1
2

1

I

2
1
2
1

w
w

w

w

w

1

2

2

1
2
2
1

1
1
2

1

1

2
2
2
1
2
1

w
w

2

1
2
1
1
2
2
2
1
1

w
w

w
w
w

1
2
2
1
1
2
2
1
2
1

w
w

w
w

1

1
1
1

2
1
2
2
2
2

1
2
2
2

1
2
1
1
1
2

1

2
2
2
2
1
2
1
1
1

w
w

w

1
1
1
2
2
1
2
2
2
1

w

1

2
2
2
2

1
1
1
1
2

The Q (question) columns indicate whether a double or single touch was presented. The A (answer) columns
give the answers ; blank means correct, w means wrong. No answers douhtful were given at this sitting.
There is one catch error, in period ix.

Table 2. Method of Right and Wrong Gases with Catches. Subject 7.
February 2Uh, 1913, 8.32 a.m.

1

ii

iii

iv

V

vi

vii

viii

ix

X

Q

A

Q

A

Q

A

Q

A

Q

A

Q

A

Q

A

Q

A

Q

A

Q

A

a

w

1

e

w

d

b

w

1

a

w

b

w

e

b

w

e

w

h

w

1

c

1

e

e

w

1

1

1

1

e

1

a

w

a

w

1

1

c

w

d

d

1

1

c

w

1

1

d

d

w

1

1

1

1

a

w

a

w

e

e

e

1

d

c

1

h

w

1

1

1

1

b

w

1

a

IP

1

1

c

w

1

1

1

d

w

1

1

e

1

a

w

1

c

w

b

w

1

e

to

a

w

c

w

1

1

c

w

1

1

1

b

w

1

1

b

w

1

&

w

1

d

d

d

1

1

1

1

1

a

w

e

Here a, h, r, d, e mean double touches of 1, 2, 3, 4, 5 cms. respectively.

436 Changes in the Spatial Threshold during a Sitting

At the next sitting the values of these stimuli would be changed each
by the same number of millimetres, say to

1-2, 2-2, .3-2, 4-2, 5-2 cms.

and in the course of ten sittings every millimetre would be used'. The
lower limit was not necessarily one centimetre but whatever it was the
stimuli in ten sittings covered five centimetres, millimetre by millimetre.
This was done for a purpose not germane to the present issue: nor need
it have been mentioned here except for completeness and because other-
wise some of the decimals in the results might have appeared to be
arithmetically impossible. The hundred applications in a sitting were
divided into ten periods which are designated throughout this paper by
the Roman numerals i to x. In each period there were five catches and
five double touches, namely one application of each of the five distances
used at that sitting. The order in which the stimuli were presented
was determined before the sitting by drawing cards. The periods, in
the cases of subjects 7, 8 and 9, are not groups, for in a group the
stimulus must be the same throughout : but in the case of subject 6
the periods happen to be also groups. The accompanying Tables 1 and
2 are examples of these two methods. They were filled up column by
column, beginning at the top of the left-hand column.

With subject 8 the method followed was in most respects the same
as that just described, but the stimuli, which in this case were not
changed from sitting to sitting, but were the same throughout, were in
steps of one and a half centimetres.

4. Catch Errors.

Before describing the processes of calculation used, it should be
explained that the catch applications of only one point were not used in
any way in the calculations. They were there simply as a check, and
the rule followed was that a sitting was rejected in which more than five
catch errors (out of a possible 50) occurred. This number was chosen
because when the catch errors did become more numerous than this
they usually became very numerous. When a series of experiments is
begun on a new subject catch errors are usually very frequent in the

' See F. M. Urban, "Die psychophysischen Massmethoden," Archiv f. d. ges. Psychol.
1909, XV. 295—6, or The Application of Statistical Methods to Psychophysics, Philadelphia,
1908, 54—5.

Godfrey H. Thomson 437

first few sittings and the subject is apt to be incredulous that he can so
often mistake one point for two.

In the case of subject 6 three series of experiments, each of ten
satisfactory sittings, were carried "out. The first series took place in
March, 1912, and in fourteen consecutive sittings the number of catch
errors never once rose above five. The first four sittings were treated
as practice sittings; the remaining ten sittings form the first series. In
these ten sittings there were only fourteen catch errors.

The second series was begun on Jan. 20, 1913. Between that date
and Feb. 6th seventeen sittings were held and seven were rejected for
catch errors : namely on Jan. 20th with thirteen, on Jan. 21st with six,
and on Jan. 23rd with seven (which dates come at the beginning) and
on Jan. 30th with twelve, on Feb. 1st with nine, on Feb. 2nd with
twenty, and on Feb. 3rd with ten (a period when the subject was
indisposed). The ten satisfactory sittings contained, in all, seventeen
catch errors and form the second series for this subject.

The third series for this subject began on Feb. 21st, 1913, and ended
on March 3rd. No sittings were rejected, and the total number of catch
errors was eighteen.

With subject 7 two series each of ten sittings and one half-series of
five sittings were carried out. The sittings began on Jan. 20th, 1913,
when there were ten catch errors. On Jan. 21st there were four, but
this sitting also was counted as preliminary practice. After that, not
a single sitting was rejected, and the total number of catch errors was
only fifteen in twenty-five sittings.

With subject 8 at the first sitting on March 26th, 1913, there
occurred fifteen and a half catch errors (the answer * doubtful ' is
counted one-half), on March 27th there were ten, after which all sittings
were satisfactory. The total number of catch errors in these was ten in
ten sittings.

With subject 9 the number of catch errors on the first day,
March 30th, 1913, was nine, next day nineteen and a half, next day
twelve and a half, after which only one sitting was rejected, on
May 12th, when there were five catch errors and a half. The number
of catch errors in the ten satisfactory sittings was sixteen and a half.

In short, the only rejected sittings under the rule were practice
sittings at the beginning of a series, four sittings when subject 6 was
not well, and one sitting with subject 9.

From the records no connexion between catch errors and period of
sitting could be proved.

438 Changes in the S^jatial Threshold during a Sitting

5. Processes of Calculation.

The process used was the Limiting Process^ In the case of
subjects 7, 8, and 9 this process could be applied simply and directly
to the data as follows. Take for example period i in Table 2, and con-
sider the stimuli a, b, c, d, e in this order, that is from the smallest to
the largest. The first correct answer^ is met at d, therefore d (here
4 cms.) is taken as the just perceptible stimulus. Next consider the
same five stimuli in the order e, d, c, b, a, that is from the largest to the
smallest. The first incorrect answer^ is at e or 5 cms. and this is taken
as the just imperceptible stimulus. The average of these two, i.e.
4"5 cms. is taken as the threshold in period i. The same proceeding in
period ii gives 4 cms. for the just perceptible stimulus and 3 cms. for
the just imperceptible stimulus, and the threshold is accordingly taken
as 3'5 cms.

This simple process could not be applied in the case of subject 6
for here the stiaiulus in a period was the same in one sitting. For
example in Table 1 period iii is exclusively devoted to the distance
2^ cm. and although the answers suggest that the threshold is here a
little below 2^ cms. they do not enable us to calculate its value. At
the next sitting, however, period iii was devoted to 4| cms., and in turn
each stimulus was presented, day by day, at this particular period of the
sitting. The results for period iii for the ten sittings from Feb. 21st to
Mar. 3rd, 1913, were for example as follow :

r in cms

h

1

H

2

2i

3

H

4
5

H

5

Correct answers

0

0

0

0

3

0

1

5

5

To these Urban's formula^ is applied. This, which may be described
as a theoretical instead of a direct application of the Limiting Process,
takes the threshold as the mean of T and T' where T is the just
perceptible stimulus and is given by

T=lPr,

while T' is the just imperceptible stimulus and is given by

r = tP'r.

1 See Thomson, op. cit. 210.

* Or the answer doubtful, if it is first met.

^ V. op. cit.

Godfrey H. Thomson

439

Here the r's are the centimetres of the stimulus values. The P's and
P"s can be most easily explained by working out the threshold for the
numbers quoted above. This is done in Table 3, where the first column
gives the r's in centimetres. The second, or p column, gives the propor-
tion of correct answers at each r, and is obtained by dividing the above
numbers by five, p is the probability of a correct answer being given
at the corresponding r. On the other hand, q is the probability of an
incorrect answer, and is obtained by subtracting p from unity: this
gives the third column in the table. The next column contains the P's.
Take the P for 3"5 cms., namely 0"08. This is obtained by multiplying
the q's together from the top of the table down to 3*0 cms. (this gives
0*4) and then multiplying by the p at 3"5 cms. (0'4 x 0'2 gives 0-08).
All the P's are obtained in a similar way. They are the probabilities
of the just perceptible stimulus occurring at the corresponding points.
The P"s are obtained by a very similar process, but beginning at the
bottom of the table and multiplying a number of ^'s and one q. They
are the probabilities of the just imperceptible stimulus occurring there.
The products Pr and P'r are then formed. Then

T=.%Pr =306,
r = SPV = 3-4,
mean threshold = 3'23 cms.

In this way the thresholds for each period could also be found for this
subject.

Table 3. Example of use of Urban s Formula,. Subject 6.
Third Series. Period iii.

r cms.

P

9

p

P'

Pr

P'r

0-6

0

1

0

0

10

0

1

0

0

1-5

0

1

0

0

2-0

0

1

0

0

2-6

0-6

0-4

0-6

0

1-5

8 0

0

1

0

0-2

0-6

3-5

0-2

0-8

0-08

0-8

0-28

2-8

4-0

1

0

0-.32

0

1-28

4-5

1

0

0

0

5 0

1

0

0

0

—

—

—

—

—

—

—

3-06

3-4

Threshold =i (3 '06 + 3-4) =3-23 cms.

440 Changes in the Siyatlal Threshold daring a Sitting

6. The Raw Results.

The above direct process of calculation gives for subjects 7, 8, and 9
a value of the threshold for each period of each sitting. At the end' of
a series of ten sittings these were averaged ; that is — all the period i
thresholds of No. 7 were averaged, all the period ii thresholds, and so
on. This average is obviously uncorrected for diurnal variations in the
threshold, and gives greater weight to the results obtained on days
when the limen was high. The indirect process (Urban's formula) used
for subject 6 is open to the same objection. Nevertheless it seems
advisable to give these uncorrected results in the first place, and this is
done in Table 4. Each column of this table is therefore based on ten

Table 4. Average Changes in the Spatial Threshold on right
forearm during the progress of a sitting.

Subject no.

6

Subject no

7 i

Subject Subject

Weighted

X

no. 8

no. 9

cms.

cms.

cms.

cms.

cms.

cms.
3-99

cms.

cms.

cms.

i

2-98

2 15

3-25

3-30

3-28

4-50

3-50

3-38

6C

ii

2-66

1-95

3-11

2-95

2-28

3-29

4-57

3-20

3 05

■*3

ill

2-40

1-85

3-23

3-25

2-68

3-34

4-42

2-60 1

2-99

"tE

iv

216

2-19

2-58

3-05

3-28

2-99

4-17

2-45 ;

2-83

**•*

V

1-84

1-59

316

3-40

3-28

2-84

4-80

2-85

2-96

o

vi

2-01

1-72

2-29

2-95

3-08

3-09

3-97

2-95

2-74

m

vn

1-95

2-05

2-61

3-35

8-68

3-59

4-95

3-20

3-14

fti

viu

2 06

2-05

3-06

2-45

3-88

3-34

412

3-05

2-94

2

ix

2-52

1-92

2-68

3-05

3-68

3-44

3-75

315

2-98

fr.

X

2-06

1-85

2-15

2-65
3 040

3-28

3-44

4-45

3 05

2-84

Me

ans ...

2-264

1-932

2-812

3-240

3-335

4-370

3-000

sittings each of a hundred judgments, except column 5, based on five
sittings only. Each column is divided into ten periods shown by
Roman numerals. The first figure, for example, 2*98 cms., is calculated
from a hundred^ judgments, namely the first ten judgments at each of
ten sittings. For subject 6 only, the sitting proper was preceded by a
preliminary group of ten judgments. This group is not used here or
elsewhere in this paper.

^ But half of these were catches and are not actually used in the calculations.

Godfrey H. Thomson 441

The last column in Table 4 gives the weighted means of each row.
The weighting is necessary because column 5 is based on five sittings
only instead of ten. The^^first figure, 3'38, in the last column is, for
example, obtained by adding together 2-98, 215, 3-25, 3-30, 1^6I^, 3*99,
4"5(), 3"50, and dividing by 7^. Obviously such a process is uncorrected
for the personal differences between the subjects and gives undue
influence to subject 8.

This last column is therefore the raw result of the experiments,
uncorrected for diurnal or personal variations. Each of the ten figures
in the column is based upon 750^ separate judgments, or 7500 judg-
ments in all.

It will be seen that the threshold is highest in period i and lowest
in period vi. It falls sharply at first, then more slowly, then rises again
and becomes irregular, but shows a sharp fall at the very end. If the
individual columns are examined it will be seen that period i is with
one exception always bigger than periods ii and iii, and with few ex-
ceptions is bigger than periods ii to vi inclusive.

It remains to be seen, in the following paragraphs, whether the
correction for diurnal and personal variation makes any important
change in this result. It will be found that this is not the case.
Further it has yet to be shown that these results are significant, that is
that the number of experiments is sufficiently large for the law of
averages to eliminate ' chance ' variations, i.e. those not due merely to
the progress of the sitting.

7. Corrections for Personal Differences and
Diurnal Variation.

A correction for personal variation from subject to subject can be
immediately and simply applied, by dividing each number in Table 4
by the mean value of its column, and then averaging the rows afresh to
get the weighted means of the ratios thus formed. This gives column A
in Table 5. In variation with period it does not differ in any of its
characteristics from the last column in Table 4.

In the case of subjects 7, 8 and 9 a more thoroughgoing correction
is to divide each individual threshold for each period of each sitting by
the average limen of that sitting. This eliminates both diurnal and
personal variation at once. There are, it is true, certain mathematical
objections to such a method, which are connected with the fact that the

^ But half of these were catches and are not actually used in the calculations.
J. of Psych. VI 29

442 Changes in the Si)attal Threshold dnring a Sitting

stimuli are necessarily in steps ; but these objections are minimised by
the daily change of stimuli in the case of subjects 7 and 9, and are not
important. In the case of subject 6, however, where Urban's formula
is used, it is difficult, if not impossible, to eliminate diurnal variation,
and therefore only the above three subjects are referred to in columns
C and D of Table 5. In column G the diurnal correction has not been
applied ; it is formed in the same way as column A. The differences
between A and C are due only to the exclusion of subject 6. In
column D, however, the diurnal correction has been applied, and com-
parison with column G shows that this correction has not been very
important.

Table 5. Result of correcting for Personal Differences and
Diurnal Variation.

A.

B.

C.

D.

^

B

C

D

i

1143

0-022 1

1-108

1-115

be

ii

1-029

0-028

0-983

1-008

s

iii

1-003

0-022

0-969

0-958

5»

iv

0-953

0-023

0-928

0-932

o

V

0-972

0-033

1-005

1-010

CO

vi

0-915

0-013

0-947

0-955

CO

9

vii

1040

0-022

1-092

1-094

^

viii

0-990

0-026

0-970

0-970

g

ix

1-010

0-020

1-002

0-990

X

0-944

0-022 j

0-987

0-965

This column gives the weighted means of the ratios obtained by dividing each

threshold in Table 4 by the average threshold of its series, given at the bottom

of its column in Table 4.

The probable errors of the numbere in column A.

This column is formed in the same way as column A but is confined to subjects

7, 8 and 9 only.

This column refers also to 7, 8 and 9 only, and shows the result of correcting for

diurnal variation.

8. Probable Errors and Significance of the Results.

In Table 5, column B gives the probable errors of the numbers in
column A. These probable errors are calculated in the following way.

It will be remembered that in order to eliminate personal differences
each threshold in Table 4 was divided by the average value of its
column. For example, the top row of Table 4, referring to period i of
the sitting, then becomes

1-313, 1-113, 1-162, 1-086, 1-012, 1195, 1030, 1167.

Godfrey H. Thomson

443

Of these numbers, 1-012 has only half weight. If this be borne in
mind, their weighted mean will be found to be 1-143, the top number
in column A of Table 5. If we now form the deviations of the above
numbers from 1143, square them, and perform the usual calculation for
the probable error (keeping in mind the half weight of one member of
the series), we obtain 0-022, the top number in column B of Table 5.
The other numbers have been similarly obtained.

These probable errors enable us to say whether the differences
between the periods are significant or accidental. They will be sig-
nificant if the difference between two periods is of the order of six
times the probable error of either, say about 0-14. Thus the difference
in column A between period i and period iii is just significant, the
difference between period i and period vi is more certainly so^

The above probable errors are calculated each from eight numbers
which are themselves averages. It is perhaps more satisfactory to
calculate them from the numerous individual thresholds, and this we
can readily do for subjects 7, 8 and 9, With these three subjects
there were altogether forty-five sittings, and therefore that number of
measurements of the threshold for each period. Each of these is, as
above, divided by the mean threshold for the day, to eliminate diurnal
variation, and thus for each period there are forty-five ratios clustered
round unity. From the distribution of these ratios probable errors can
of course be calculated. For example, two of these distributions are :

Value of ratio ...

•3

•5

•7

•9

1-1

1-3

1-5

1-7

Period i

0

0

6

8

17

8

4

2

Period iv

2

0

8

18

11

5-5

0-5

0

By 17 values at I'l is meant that there are that number of values
firom 1-0 to 1-2. A value exactly at either of these points is split
between the neighbouring compartments. The upper distribution cor-
responds to the average 1-115 at the head of column D in Table 5.

1 This simple consideration is possible because the probable errors are all about the
same size. It is visualised in a figure at the very end of this article, where the shaded area
shows three times the probable error above and below the line representing the changing
threshold. To be more accurate one ought to compare the differences with the probable
error of the differences. For example, the difference between i and iii is 0-140 ±0-031,
that between i and vi is 0-228 ±0-026. The latter is probably significant even if the most
liberal allowance is made for the inaccuracy of probable error formulae for small numbers.

29—2

444 Changes in the Spatial Threshold during a Sitting

The semi-interquartile range is 0"13, and the probable error therefore
approximately 0*02, agreeing well with the value in column B (which
latter, however, includes subject 6 and refers to values uncorrected for
diurnal variation).

The lower distribution corresponds to the average 0*932 in column D.
Its seuii-interquartile range is 0"125, and its probable error therefore
again about 0'02.

The difference of the means of the two distributions is 0*183, and
the probable error of the difference about 0*03. It is, therefore, signifi-
cant. These last considerations have not included the results obtained
with subject 6, because in this case a different process of calculation
was used, viz. Urban's formula. The probable eiTor for this formula
has, however, been worked out^ and can be calculated for the thresholds
found for this subject. They show again that the results are significant
when all three series are put together. Even each series by itself is
almost sufficient. For example, the third series, given in the third
column of Table 4, gives the following results- :

Period i 3-25 cms. ± -218,

Period vi 2'29 „ ± '228,

Difference 0-96 „ ± -316.

In his extended experiments on lifted weights Urban found that
there was no variation in the threshold (other than variation explicable
as we say by chance) to be detected in a sitting of thirty-five judgments^
The criterion which he employed was to perform certain calculations (on
a table resembling our Table 4, but divided into five periods only) which
gave him the measures of the physical and accidental variati£»n^ Similar
calculations can be applied to our results, and it seems advisable to the
writer to do so in an elementaiy way needing but little mathematical

^ Urban's own calculation of the probable error needs correction. See G. H. Thomson,
"The Probable Error of Urban's Formula," this Journal, 1913, vi. 217 — 222, especially
equation (7). I have to thank Professor Urban for communicating to me privately his
approval of these corrections and regret that his letter did not reach me in time to be
mentioned in the article quoted.

- Note that these numbers confirm the probable errors given in column B of Table 5.
For if 3'25 cms. be reduced to nearly unity, the p.e. -218 will be reduced to about "07 :
and as there are eight series the final p.e. should be of the order -07 -r >JQ. It is actually
•022, which agrees very well.

* F. M. Urban, The Application of Statistical Methods to Psychophysics, Philadelphia,
1908, p. 39.

^ Urban, op. cit. Table 25, p. 185, and pp. 34 ff. References are given to Lexis,
Czuber, and Bortkewitsch.

Godfrey H. Thomson 445

manipulation, and calculated to make this criterion of variation clear to
non-mathematical readers.

Column A in Table 5 gives the final results of all the experiments,
and the point at issue is whether the numbers in this column actually
vary with the periods, or whether the differences seen are due to chance
alone and would disappear if larger averages could be taken. The
probable error of each number is given, and of these the largest is "OSS,
while the average is '023 + '001.

Now there are ten values in column A, with an average value of
unity, and if these ten values are all measurements of the same quantity
(that is, if there is no true variation with the time), then they ought
to be so grouped about unity that the semi-interquartile range of this
grouping is equal to the probable error of each. This serai-interquartile
range, calculated by the formation and quadrature of the deviations,
proves to be about "043 + "007, and is therefore considerably bigger
than the probable error of any one of the numbers from which it is
calculated, which lends support to the belief that these numbers do not
differ by chance sampling only, but show also a fundamental difference
not to be obliterated by averaging more and more readings. The
difference between "043 and '023, however, is only barely significant
in the light of the probable errors of these numbers, and these probable
errors themselves are only vague in cases like these where the number
of readings is only ten.

Lastly, if there is such a variation with time as the writer suggests,
there ought to be some correlation between the different columns of
Table 4, especially between those columns referring to the same subject.
This is actually the case ; for example, the correlation between the first
two columns is "465 + '167. The correlation between some columns, and
especially between the upper halves, is a good deal greater.

9. The Influence of the Experimenter.

The writer believes that these experiments make it reasonably
probable that, at least in the case of some subjects and experimenters,
the spatial threshold during the course of a sitting at first sinks on the
average and then later becomes erratic, but on the whole rises, except
for an ultimate drop at the very end of the experiment due in all
probability to 'end spurt,' for it was easy for the subject to judge that
the sitting must be nearing its close, since all sittings were of the same
length. It does not of course follow that these changes would occur

446 Changes in the Spatial Threshold during a Sitting

with every experimenter, and even if it proved to be so, it would still
not be shown that the change is one which takes place in the subject :
for it may be that the change is due to the conduct of the experimenter.
Almost all psychological tests consist in the response by a subject to
stimuli arranged by an experimenter, and in many cases the result may
well be a function of the mental and physical condition not only of the
former, but also of the latter : and especially will this be the case where,
as here, a considerable degree of dexterity is required. To test this point
some experiments have been planned in which the writer's skill in using
the aesthesiometer would be tested by mechanical means, but circum-
stances have not yet allowed these tests to be carried out. Against such
an explanation of the results may be brought, firstly, the follomng
general consideration. A change in the experimenter's dexterity during
a sitting would probably consist of an increase of skill in the earlier part
of the sitting. But such an increase of skill would probably make it
more difficult for the subject to distinguish two points. For example,
increase in the accuracy of applying the two points simultaneously
would presumably make it more difficult for the subject to distinguish
them. But this effect would be in exactly the opposite direction from
that revealed by the present research.

There are a few further points which have bearing on this question.
It has been already said that in some cases subject No. 6 acted as
experimenter instead of the writer. The experiments on subject No. 8
(the writer himself) were thus carried out, and as they show little or no
trace of the variation mentioned this might be taken as evidence that
this variation takes place only when the writer himself acts as
experimenter.

Another point which might be taken as further evidence in the
same direction is the following. A number of the measurements on
Nos. 6 and 7 were made on the same days and at the same hour, and
the subjects were taken first on alternate days, now No. 6 first, next day
No. 7 first. On these days the writer was always experimenter. There
was considerable correlation (almost '67 ± -12) between the daily mean
thresholds of the two subjects, and this might be due to changes not in
them but in the common experimenter.

On the other hand, the correlation might be due to similar weather
conditions or some other common factor. Further, the absence of the
sought-for variation in No. 8 might be due to an unconscious resist-
ance owing to a strong desire to be impartial, for this subject alone
knew exactly what the purpose of the experiments was. Finally, in

Godfrey H. Thomson

447

some experiments on subjects 7 and 9, the writer and subject 6 acted
alternately as experimenter; and, as far as the necessarily smaller
quantity of data enable a conclusion to be drawn, the suspected
variation was as evident with the one experimenter as with the other —
if anything perhaps it was more evident when No. 6 experimented.
Consider, for example, the following numbers, which are proportional
to the threshold for No. 9, and show results obtained by the writer and
also by No. 6 as experimenter :

Subject No. 9. Avei^ages of seven sittings. The experimenter was
changed after each period.

Period

i ii

iii

iv

Experimenter: writer ...
No. 6 ...

1103
1-295

1184
1-060

0-840
0-802

0-877
0-757

10.

Sl

IMMARY.

These experiments therefore appear to make the following results
reasonably probable on the average.

(1) During a sitting there is a change in the spatial threshold.

(2) The threshold falls sharply at first, then slowly and steadily
until about fifty judgments have been given. This improvement is
probably due to the subject's 'finding himself Stimuli are compared
with former stimuli, especially with those about which the subject felt
certain. All the subjects made remarks during and after the sittings
which suggested that this was so.

(3) It then rises or at any rate becomes irregular.

(4) There is a final drop just at the end of the sitting, due probably
to ' end spurt.'

These changes are best shown in the figure (page 448), where the
shaded area represents three times the probable error, on both sides
of the central line.

(5) These changes seem to have their origin in the subject, not
in the experimenter.

(6) They appeared in greater or less degree in all the subjects
except one, the Avriter himself

448 Changes in the Spatial Threshold dnring a Sitting

Variation in the Spatial Threshold during one sitting.

Average of seventy -five sittings on four subjects. The Mean
Threshold is taken as unity.

1-2

■

^

1-1

<

<

'\ Hill,,.

■

1-0

'1 '^

I''
(

4

«< ■

0-9

-

11

' '

1

i ii ill jv V VI vii viii ix x

Periods of sitting (approximately minutes).
The shaded area shows three times the Probable Error, both above and below the average.

{Manuscript received 7 December 1913.)

REVIEW

Development and Purpose : an essay towards a philosophy of evolution.
By L. T, Hob HOUSE. London : Macmillan and Co. 1913. Pp. xxix
+ 383. 10s. nett.

Development and Purpose is the logical outcome and completion of
the lines of thought set forth in Prof Hobhouse's three previous works,
Mind in Evolution, Morals in Evolution, and The Theory of Knowledge.
In an interesting and important preface the author sketches in brief
outline the development of his philosophical views. Rejecting Material-
ism and unable to accept Idealistic theories Prof Hobhouse arrived at
the conclusion "that a philosophy that was to possess more than a
speculative interest must rest on a synthesis of experience as interpreted
by science, and that to such a synthesis the general conception of
evolution offered a key " (xviii). Still later it was seen " that the
evolutionary process can be best understood as the effect of a purpose
slowly working itself out under limiting conditions which it brings
successively under control" (xxvi); and "that there is a spiritual
element integral to the structure and movement of Reality" which,
while not the whole of Reality, nevertheless becomes dominant by a
process of evolution (xxvii).

Turning to the argument. Part I is, we find, chiefly devoted to an
empirical survey. Its keynotes are ' correlation ' and ' harmony.' All
progress may be measured in terms of a more intimate, or a wider-
embracing, correlation : while in the sphere of the Practical Reason,
harmony is shown to be all-important, and a striving towards a more
complete harmony the aim. In Part II the deductive argument is set
forth and the validity of Experimental Reconstruction examined. This
necessitates a review of the logic of induction in the course of which
an important principle emerges, viz. that " the assumption of our
scientific reasoning is that the variable relation is ultimately traceable
to uniform relations, and that is to relations dependent on the intrinsic
character of the terms ' as such ' " (344).

450 Revieio

As regards the author's conception of validity, we find that it too is
saturated with the idea of development. For instance, " the validity of
thought is not that of finality or achievement but of growth " (276),
and again, a method may be valid even " though its immediate result
does not possess final truth," it is an impulse in the right direction
(339). The validity of such a method lies in the fact "that it is
essential to the movement towards truth " (340) ; whence it follows that
"validity is a wider conception than truth" (footnote to 128). The
ultimate test of truth and validity alike seems to be that of coherence
or consilience. Even axioms and first principles are not exempted it
appeal's, since we are told that knowledge does not depend on first
principles which are superior to criticism but on a body of mutually
supporting judgments (265) which we build up as we go along (268).
Nevertheless "What appear as 'first' principles are... not mere assump-
tions [but] they express the pervading unity in a system of judgments "
(273) ; " Truth is objective " (274).

Empirical survey and deductive argument alike agree in demonstra-
ting the presence of two processes in Reality. Mechanism exists
alongside of Purpose ; indeed, " the whole process of Reality is mechan-
ically conditioned" (357). Hence Mechanism is not excluded from
Reality (it is in fact one of the " modes in which Reality operates "
footnote to 329), but "it must be conditioned by relation to an ultimate
harmony, while harmony is equally conditioned by mechanism " (350).
In the case of a psycho-physical organism "mechanical relations are
qualified by teleological relations " thereby forming a new system
(footnote to 329). The two processes, however, must not be falsely
hypostatised as substances and a dualism of Body and Mind set up.
The psycho-physical organism must rather be conceived as a unit
system in which hypotheses of Interaction and of Psycho-physical
Parallelism are alike equally out of place {ibid.). A criticism may
here be raised and it may be asked whether the two processes, the
mechanical and the teleological, are in very truth co-ordinate and
mutually determining ? For it would seem that since the mechanical
order is ultimately imbued and transfused by the teleological order it
is rather to be looked on as a condition of the latter, and in so far
secondaiy and subordinate to it.

Fully understood the central essence of the Real is best expressed
by organicity (338). Moreover, there is every reason to believe in
the steady development of the organic principle. But " organic action
is of mental character " (316). Therefore over and beyond the individual

Review 451

minds familiar to the plain man Hobhouse postulates the permanent
activity of a central Mind " not limited to a single physical organism "
(365). This central Mind may be conceived as not fully defined at the
beginning but as undergoing development (footnote to 370) commen-
surate with the increasing manifestation of Purpose (365). In the final
stages its growth becomes largely self-consciously determined as the
'mother-sense' (comprising massed feeling and other underlying forces),
together with the conditions determining the operation of its activity, are
gradually brought within the focus of consciousness (38, 247) : at the
same time Mind is enabled to attain to a just orientation through the
appreciation of its own development and limitations (280). It is, too,
this very inclusion of the non-rational elements within the sphere of
Reason which constitutes their redemption (249). Reality as a whole
is rational; irrationalism is strongly denounced (145 seq.).

Again, the organic as such is harmonious; th^efore the develop-
ment of the organic principle implies a growth of harmony, which,
eventually at least, is " identical with the growth of Mind " (364). It
follows that until the whole of Reality is harmonised there is necessarily
a certain degree of clash and discord, which we know as evil. This,
unfortunately, is unavoidable since " Harmony alone does not explain
existence " (350). The operation of mind is not absolute but con-
ditioned. True, there is an increasing purpose in the universe, but
purpose rightly understood is " a cause conditioned in its operation by
its own tendency" (319). From this is deduced that though not
every event is necessarily good per se, yet " every event proceeds from
some combination of forces, each of which is somewhere or sometime
necessary for the fulfilment of the world-purpose " (367) ; nevertheless,
until this purpose is completely realised such elements "may remain
in greater or less disharmony with other elements " (363-4). It might
seem that overwhelming importance is here attached to the future.
Further, since transitory evil, positive though it may be, is merely
a condition (ultimately overcome) in the attaining of a completely
established harmony, it seems open to question whether evil, according
to Prof Hobhouse's account, is so ultimate and real as he would
postulate ?

It will be clear that in a work of so comprehensive a scope many
questions of the utmost interest come under consideration : of these
space forbids one to treat in any detail. Suffice it to say that matter
interesting to metaphysician, psychologist, biologist, aesthetician, theo-
logian, moralist, logician and citizen alike is discussed.

452 Review

" Time " we are told " depends on the function of change as necessary
to development" (351). Beauty stands almost alone as "insusceptible
of progress"; new art does not give greater beauty "but a fuller
interpretation of experience with a deeper and more truthful expression
of feeling" (218). In regard to the pre-formation theory, it is urged
that "the germ need not be in the least like the matured order. It
must only have a mode of operation, which is determined by the needs
of that order " (360). Suggestions are thrown out for the origin of the
animate from the inanimate (358) as well as for the rise of consciousness
(287, 364). But' it is in Part I that the matter chiefly interesting to
the psychologist as such is to be found, where inter alia we are given a
new criterion of consciousness, viz. sensori-motor activity (56 seq.).

It need scarcely be said in conclusion that the work is one of con-
siderable interest and importance and demands individual study. It is
especially valuable .for its richness of suggestion and for the emphasis
laid on empirical progress.

E. M. SMITH.

PUBLICATIONS RECENTLY RECEIVED

A Manual of Psychology. By Professor G. F. Stout. Third Edition, revised
and enlarged. London: University Tutorial Press, Ltd. 1913. Pp.
xvii + 769. 8s. 6d

In preparing this edition of his well-known work, Professor Stout has rewrittei)
the greater part of it and has subjected the rest to careful revision. Two new
chapters, on Instinct and Attention, have been added ; and extensive alterations
have been introduced in the account of the development of the Perception of
External Objects, in the discussion of the connexion of Mind and Body, and in the
treatment of the Perception of Spatial Relations. The effect of these changas is jiot
only to widen very materially the scope of the book but also to render it fai- more
easily intelligible to the general body of students.

Psychology applied to Legal Evidence 'and otlier Constructions of Law. By
G. F. Arnold. Second Edition. Calcutta : Thacker, Spink & Co.
1913. Pp. 607. Rs. 12.

In this edition, various additions have been made and Miinsterterg's book on
Psychology and Crime has been freely drawn upon. But "as the author is
unacquainted with German, he has been compelled to relinquish the idea of
consulting" Stern's Beitrdge ziir Psychologie der Aussage, the Zeitschrift fur
angewandte Psychologie, and other foreign periodicals. Despite these defects and
others arising from the training of the writer, the book is valuable as the first
seiious attempt of an English lawyer to consider the data and conditions of legal
evidence from the standpoint of psychology.

Die agrammatischen Sprachst'&rungen : Studien zur psychologischen Grund-
legung der Aphasielehre. Von Professor Arnold Pick. Berlin : Julius
Springer. 1913. S. viii + 291. M. 14.

Nearly half of this book is devoted to a diffuse Vorrede und Einleitung in which the
author lays stress, inter alia, on the value of psychological investigations, especially
those of the Wiirzburg School, for the proper understanding of the pathology of
speech defects. The subsequent six chapters deal with the definition of agrammatism,
the definition of the sentence, the means of expression in speech, the path from
thinking to speaking, Wundt's theory of Gesammtvorstellung, and internal speech.
Thus the author treats his subject essentially from the standpoint of psychological
theory and experiment. He dedicates the book "to the memory of Hughhngs
Jackson, the most profoimd thinker in neuropathology during the last century."

Objektive Psychologie oder Psychorejlexologie : die Lehre von deii Assoziations-
reflexen. Von Professor W. von Bechterew. Translated from the
Russian. Leipzig and Berlin: B. G. Teubner. 1913. S. viii + 468.
M. 16; geb. M. 18.

This book is an attempt to found a psychology on reflexes; the following
sentences will suflice to show its standpoint. Psychoreflexology, which treats of
reactions in and for themselves, apart from the subjective experiences preceding or
accompanying them, gives us, says the author " prose in place of poetry and treats
the neuro-psychic functions exclusively from their external aspect." The author is
convinced that as soon as exact experiment hcxs determined the interrelation between
the objective data of psychoreflexology and the data of subjective experience,
psychology will become as exact a science as physics or chemistry. Only when
psychoreflexology has reached the dignity of an exact science will suojective
psychology rid itself of its present nebulous and metaphysical hypotheses.

454 Publications Recently Received

Disturbances of tlie Visual Inunctions. By Professor W. Lohmann. Translated

by Angus Macnab. London : John Bale, Sons & Danielsson, Ltd.

1913. Pp. 185. 15.*. net.

This work contains a vast amount of useful material, but it can only be

recommended to those who are in a position to realise its shortcomings. Many of

the chapters are exceedingly sketchy ; the paragraphs bearing on psychology are

often inaccurate or insufficient ; and the translation leaves much to be desired.

The Experimental Psychology of Beauty. By Dr C. W. Valentine. London :
T. C. <fe E. C. Jack. (The People's Books Series.) n.d. Pp. 94. 6d

The seven chapters of this little book are concerned with the beauty of colour,
the l^eauty of form, balance and symmetry, and experiments with pictures. The
author has included a number of his own experiments, many of which have not been
previously published. A useful bibliography to each chapter is a]>pended.

A Syllabus for tlie Clinical Examination of Children. By Edmund B. Huey.
Baltimore : Warwick & York. Pp. 34.
A very useful series of blank forms is here given which have proved of value in
recording and tabulating the conditions of the inmates of the Illin^)is State Institute
for the Feeble-Minded. Four forms are recommended (1) for the home record
(heredity, past history and environment), (2) for the teacher's or attendant's record
(habits, capacities, morals, etc.), (3) for the physical examination, and (4) for the
mental examination. The Binet-Simon tests are described and the special pictures
originally used in connexion with these tests are reproduced.

Backward and Feeble-Minded Children. By Dr E. B. Huey. Baltimore :

Warwick & York. (Educational Psychology Monographs.) 1912.

Pp. xii + 22L $1.40.

The author gives a series of brief clinical studies of thirty-five ' border ' cases of

backward or feeble-minded children and feeble-minded adults. A still more intensive

examination of mentally abnormal individuals on these useful lines is required to

further our knowledge of mental deficiency.

Inductive versus Deductive Methods. By W. H. Winch. Baltimore :

Warwick & York. (Educational Psychology Monographs.) 1913.

Pp. 146. $1.25.

The writer describes a series of experiments planned to discover whether the

inductive or deductive method gives the better result, (a) when the children are

tested on precisely the same material as that which they had learnt or been taught,

(6) when the children were examined by means of new material. The answer to the

former of these problems varied in the five schools tested. But in all five cases "the

children who were taught inductively did better work than those taught deductively,"

when they had to apply themselves to new material.

The Marking System in Theory and Practice. By T. E. Finkelstein.

Baltimore : Warwick & York. (Educational Psychology Monographs.)

1913. Pp. 88. $1.

In this monograph the author sets out to determine the distribution of the

marks awarded by different teachers, and the degree of unreliability which

characterizes the percentage system at present in force in most American schools

and colleges.

Les Origines de la Connaissance. Par Professor R. Tubro. Paris : Felix
Alcan. 1914. Pp. 274. 5 fr.
The first two chapters of this book deal with the origin and nature of our
experience of hunger. Trophic sensations and trophic perception, according to the
author, are the forerunners of external perception ; and the trophic efifects of stimuli
are ultimately responsible for our experiences of reality and causality.

Anglo-Indian Studies. By S. M. MiTRA. London : Longmans, Green *fe Co.
1913. Pp. 525. 10s. net.

V5'i'

PROCEEDINGS OF TBE BRITISH PSYCHOLOGICAL

SOCIETY.

June 7, 1913'. Memory and Cousciou.siiess, by A. Robinson.

Are Intensity Difterences of Sensation Quantitative ?, by C. S.
Myers, G. Dawes Hicks, H. J. Watt, and W. Brown.

June 8, 1913'. Can there be Anything Obscure or Implicit in a Mental State ?,

by H. Barker, G. F. Stout, and R. P. Hoernl^.

November 8, 1913. A (Comparative Study of Normal and Sub-normal Children by
means of Mental Tests, by A. R. Abelson.

A Reaction Pendulum and a Disc, illustrative of Weber's Law,
for use in Class Teaching {demonstration), by J. Brodgh.

Observations on the Process of Learning and Re-learning in
Mice and Rats, by Mary E. Macgregor (introduced by
Dr Edgell).

An a priori Argument for the Existence of a Cerebral Centre
for Affection, by A. Wohlgemuth.

January 24, 1914. A Note on the Correlation of Ability and Variability, by
William Brown.

General Ability and the Subjects of the School Curriculum, by
Miss N. Carey (introduced by Prof. Spearman).

An Attempt at an Exact Study of Character, by Edward Webb
(introduced by Prof Spearman).

' In conjunction with the Aristotelian Society and the Mind Association.

cambridok: printed by john clay, m.a. at the uniyersity press

781

AUG 0 5 1987

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY