FT-
n
PSYCHOLOGICAL REV1E W PUBLICA TIONS
n
THE
P sychological Review
EDITED BY
JOHN B. WATSON, JOHNS HOPKINS UNIVERSITY
HOWARD C. WARREN, PRINCETON UNIVERSITY (Index"}
JAMES R. ANGELL, UNIVERSITY OF CHICAGO (Monographs) AND
SHEPHERD I. FRANZ, GOVT. HOSP. FOR INSANE (Bulletin)
ADVISORY EDITORS
R. P. ANGIER, YALE UNIVERSITY; MARY W. CALKINS, WELLESLEY COLLEGE; RAY-
MOND DODGE, WESLEYAN UNIVERSITY; H. N. GARDINER, SMITH COLLEGE; JOSEPH
JASTROW, UNIVERSITY OF WISCONSIN; C. H. JUDD, UNIVERSITY OF CHICAGO; ADOLF
MEYER, JOHNS HOPKINS UNIVERSITY ; HUGO M0NSTERBERG, HARVARD UNIVERSITY ;
W. B. PILLSBURY, UNIVERSITY OF MICHIGAN ; C. E. SEASHORE, UNIVERSITY OF IOWA ;
G. M. STRATTON, UNIVERSITY OF CALIFORNIA ; E. L. THORNDIKE, COLUMBIA UNIVERSITY
VOLUME XXII, 1915
1
PUBLISHED BI-MONTHLY BY
PSYCHOLOGICAL REVIEW COMPANY
41 NORTH QUEEN ST., LANCASTER, PA.
AND PRINCETON, N. J,
Entered as second-class matter July 13, 1897, at the post-office at Lancaster, Pa., under
Act of Congress of March 3, 1879.
PRESS OF
THE NEW ERA PRINTING COMPANY
LANCASTER. PA.
CONTENTS OF VOLUME XXII
January.
A Revision of Imageless Thought. R. S. WOODWORTH, i.
A New Measure of Visual Discrimination. KNIGHT DUNLAP, 28.
An Electro-Mechanical Chronoscope. JOHN W. TOOD, 36.
From the University of California Psychological Laboratory:
XVIII. Practice in Associating Color-Names with Colors. WARNER BROWN, 45.
XIX. The Apparent Rate of Light Succession as Compared with Sound Suc-
cession. BERTHA VON DER NIENBURG, 56.
XX. A Memory Test with School Children. ARTHUR H. CHAMBERLAIN, 71.
XXI. Practice in Associating Number-Names with Number-Symbols. WAR-
NER BROWN, 77.
XXII. Incidental Memory in a Group of Persons. WARNER BROWN, 81.
March.
A Proposed Classification of Mental Functions. GEORGE A. COE, 87.
Color Theory and Realism. KNIGHT DUNLAP, 99.
Point Scale Ratings of Delinquent Boys and Girls. THOMAS H. RAINES, 104.
A Preliminary Study of the Deficiencies of the Method of Flicker for the Photometry
of Lights of Different Colors, Part I. C. E. FERREE and GERTRUDE RAND, no.
Discussion :
The Functions of Incipient Motor Processes. S. BENT RUSSELL, 163.
May.
The Theory and Practice of the Artificial Pupil. LEONARD T. TROLAND, 167.
The Temporal Relations of Meaning and Imagery. T. V. MOORE, 177.
The Shortest Perceptible Time-Interval Between Two Flashes of Light. KNIGHT
DUNLAP, 226.
July.
An Experimental Contribution to the Investigation of the Subconscious. LILLIEN
J. MARTIN* 251.
Emotional Poetry and the Preference Judgment. JUNE E. DOWNEY, 259.
An Experiment in Association. C. G. BRADFORD, 279.
A Note on the Effect of Rhythm on Memory. H. F. ADAMS, 289.
Diagnostic Values of Some Performance Tests. THOMAS H. RAINES, 299.
Processes Referred to the Alimentary and Urinary Tracts: A Qualitative Analysis
E. G. BORING, 306.
September."
The Father of Modern Psychology FOSTER WATSON, 333.
An Investigation of the Law of Eye-Movements. MILDRED LORING, 354.
Variability in Performance During Brief Periods of Work. A. T. POFFENBERGER, JR.,
and GLADYS G. TALLMAN, 371.
The Standardization of Knox's Cube Test. RUDOLF PINTNER, 377.
The Adequacy of the Laboratory Test in Advertising. H. F. ADAMS, 402.
iii
iv CONTENTS
November.
Reaction to the Cessation of Stimuli and Their Nervous Mechanism. HERBERT
WOODROW, 423.
A Study in Simultaneous and Alternating Finger Movements. H. S. LANGFELD, 453.
Retinal Factors in Visual After-Movement. WALTER S. HUNTER, 479.
Experimental Data on Errors of Judgment in the Estimation of the Number of Objects
in Moderately Large Samples, with Special Reference to Personal Equation.
J. ARTHUR HARRIS, 490.
Origin of Higher Orders of Combination Tones. JOSEPH PETERSON, 512.
From the University of California Psychological Laboratory:
XXIII. Practice in Grading and Identifying Shades of Gray. WARNER BROWN,
Correction: May, 1915, p. 216, line 2, delete " in." The sentence when corrected
should read: " One mental process is the meaning of another mental process if it is
that other's context."
VOL. XXII. No. i January, 1915
THE PSYCHOLOGICAL REVIEW
A REVISION OF IMAGELESS THOUGHT1
BY R. S. WOODWORTH
Columbia University
Several years ago I was led by some experiments on
voluntary movement to conclude that an act might be
thought of without any representative or symbolic image,
and further study led me to extend this conclusion to other
thoughts. My attention was soon called, in a review of
this work by Angell, to previous discussions of the same
question, connected with Stout's assertion that there was
nothing psychologically absurd in the conception of image-
less thought. Looking into the contemporary experi-
mental literature, I then made the acquaintance of Binet and
of Watt, Biihler and others of the Kiilpe school, and my
own work soon fell into insignificance beside these extensive
and many-sided contributions. Even the merit of inde-
pendent confirmation was not specially important in this
case, since such confirmation was forthcoming even from those
who, like Wundt, were not at all in sympathy with the con-
clusions of the imageless thought party. It appeared that
imageless thought, the mere gross fact of observation, had
come to stay, and that the only question was what to do with
it. Some psychologists have assigned great importance to
this fact as a demonstration of non-sensory content, while
others have avoided so revolutionary a conclusion by ex-
plaining the fact away through one interpretation or another;
others again have accepted the fact but minimized its im-
portance, treating it as a mere limiting case; and some, while
Address of the President before the American Psychological Association, at
Philadelphia, December, 1914.
I
2 R. S. WOODWORTH
accepting the gross fact, have doubted that it would stand the
test of more refined introspection. Meanwhile, my own
views have been maturing as the result of continued thought
and experiment, and the time is perhaps favorable for resum-
ing the offensive, and endeavoring to uncover the weaknesses
of the negative interpretations, and for offering a conception
of the matter which may possibly appear superior to those
hitherto presented, or at least worthy of some consideration.
Of the interpretations of imageless thought which explain
the fact away without allowing it to modify existing systems
of psychology, the most important is that of Wundt. It will
be recalled that the method employed by the Kiilpe school in
studying the thought processes was drastically criticized by
Wundt, who objected to their experiments as being experi-
ments in appearance only, and held that real thinking could
not be done to order in the laboratory. He himself preferred
to rely on incidental introspections during spontaneous
thought, and in fact reports such observations of his own.1
"In such self-observations," he writes, "it became perfectly
clear to me that the thought was not formed during the
process of its verbal expression, but was present as a whole
in consciousness before the first word was reached. At first
none of the verbal or other images, which subsequently
appeared in running through the thought and giving it ex-
pression, was present in the focus of consciousness, but these
parts of the thought appeared successively as the thought was
allowed to develop." With only this fact in mind, he admits,
one might easily be led to regard the thought as a unit with
a distinctive elementary character. But quite a different
conclusion is reached when other facts are also taken into
account, that of the narrowness of the field of attention, that
of the existence of dim content in the background of con-
sciousness, and that of the "total feeling," itself a unit,
though generated by a complex of images. A thought, in
Wundt' s view, is essentially a complex of images, but these
parts of the thought are too numerous to be present together
in the field of attention. They are present at first only in
1 Psychologische Studien, 1907, 3, 349.
A REVISION OF IMAGELESS THOUGHT 3
the background and are not introspectively visible; but as
the thought is dwelt upon and expressed, its constituent
images come successively into view. What then was the
apparently unitary thought with which the process started?
This, explains Wundt, was a " total feeling," generated by
the complex of images in the background, and itself occupying
for an instant the center of the stage.
It is obvious that such a position is almost inexpugnably
entrenched. The extremely hypothetical nature of the
ground renders a direct attack hopeless. So much as this
may be ventured, that, if the words expressing a thought are
really its constituent parts, it is curious that the same thought
can be thought in different words, and even in different
languages, and still more curious that the words to fit the
thought are not always at hand. Apparently, the same
complex may be composed of different elements, and may
exist with some of its elements lacking. Further, it is curious
to reflect that these verbal images in the background must
somehow be present simultaneously and yet in proper
sequence, since otherwise they might compose quite a different
thought or no thought at all.
But the principal doubt to be raised concerns the " total
feeling." This unitary feeling, present without observable
images, and "adequate to the thought/' would almost meet
the demands of the opposing party, except for Wundt's
insistence on its being a feeling, to the neglect of its noetic
character. Certainly it is not a feeling, in any strict sense,
that straightway finds expression in a statement of fact.
Wundt's analysis leaves out of account the core of the whole
experience, namely, the fact or supposition which was sub-
sequently expressed in a sentence, but which was definitely
and clearly present in mind in advance of the words.
Several writers have called attention to the presence of
vague or apparently irrelevant imagery in moments that
would otherwise appear imageless. The presence of kines-
thetic sensations, habitually unattended to, has also been
shown in many cases, and thus we have become wary of
asserting that a given moment is really devoid of sensory
4 R. S. WOODWORTH
content. Of course, no one has ever supposed that bodily
sensation could be absent from the background of any con-
scious state, but it has been thought possible to distinguish
between irrelevant content and content related to the topic
of thought. We must, however, recognize the probability
that apparently irrelevant sensations and images sometimes
enter into the web of thinking. Especially has the attempt
been made with some success to extend the James-Lange
theory of emotions to cover the so-called " conscious atti-
tudes"; and some would even extend it to cover the imageless
awareness of definite facts, contending that every thought
has its own peculiar motor expression, and that the sensations
generated by the movement furnish the conscious content
of the thought; but no one, as far as I know, has found em-
pirical support for this extreme view.
It is worth remarking that the presence of images and
sensations in many or most moments of thinking does not
disconcert the supporter of imageless thought. He is per-
fectly willing to admit that such content is often or even
usually present; and the only real importance of a few well-
attested instances of thought without such content is that
they furnish him his most direct evidence of the existence of
other content. His main contention is that other content
exists, and that it is the most essential and characteristic of all.
But some psychologists, while admitting the occasional
occurrence of imageless thought, deny its evidential im-
portance. It is merely the limiting case, they say, in a
continuous gradation from thought in clear images, down
through thought in medium and dim images, to thought in
images at or near the zero mark. The most attractive form
of this interpretation is that which sees in the graded series
the progressive automatization of a thought through practise.
When the thought is novel, it comes with abundant sensory
content, but as it grows familiar and habitual it becomes less
sensuous, that is to say, less conscious, until, just as it is
about to become automatic and unconscious, it still shows a
feeble spark of conscious life; and this feeble spark is pounced
upon by the imageless thoughter and rashly heralded forth
A REVISION OF IMAGELESS THOUGHT 5
as proof of some unrecognized species of conscious experience.
In reality, imageless thought is imageless because it is all but
unconscious. This genetic interpretation has been presented
with most force by Titchener1 and by Book.2
The undoubted attractiveness of this conception comes
from its following so neatly from the law of practise, and its
deficiencies arise from its taking account of only one side of
the practise effect. There is much in practise besides the
tendency toward automatism. Seldom does the ' course of
training consist of repeating time after time the same per-
formance, only with increasing smoothness and speed.
Usually the process begins with varied and tentative reactions,
and advances by selection and elimination. Moreover new
forms of reaction, made possible by the progress in facility,
make their appearance in the course of training. Thus the
perfected act omits elements present at the start and contains
elements not present at the start, and may be an entirely
different means of reaching the same result. If therefore
the first thinking on a given topic is fraught with imagery,
while the practised thought on the same topic is bare of im-
ages, it does not in the least follow that the imageless thought
is a condensation of the imaginal. It may be a more econ-
omical substitute. The imagery present at the start may
have been due to a diffusion of excess energy such as is
common in unpractised acts, or it may have furnished a
round-about way of dealing with the problem and have given
place with practise to the more direct attack represented by
the imageless thought.
Practise experiments give little ground for believing that
a series of part acts, by simply becoming very easy and swift,
blend together into a total act in which the parts are lost to
sight. Rather has it been found true that the more inclusive
acts, such as dealing with words and phrases as units, in
typewriting and telegraphy, arise suddenly as new forms of
action, in the progress of training, and themselves make
possible a great increase in the speed of the partial or lower-
1 "Experimental Psychology of the Thought Processes," 1909, pp. 173, 183, 187.
2 PSYCHOLOGICAL REVIEW, 1910, 17, 381.
6 R. S. WOODWORTH
order acts. The partial acts do not blend to produce the
inclusive act, but the latter is hit upon and causes the former
to blend. Attention deserts the parts, which thus become
automatic; but attention still remains keenly alive, being
directed to the more inclusive acts. These higher acts are
real units, and not mere blends; they are clearly conscious
and yet not in imaginal form; indeed, they seem the very
type of an imageless thought.
Observations of new ideas, at their first appearance in an
individual, would be of interest in relation to the interpreta-
tion of imageless thought as exclusively old and well-drilled
thought. In the hope of gathering such observations, I have
sought to catch myself at moments when some new idea
germinated in my mind. Unfortunately, opportunities have
not presented themselves with the frequency that could be
desired; but, in the few instances that I have collected the
experience could be described as the dawning of some new
meaning in things, sometimes with scrappy verbal and visual
images, sometimes with none that were observable. When
they occurred, the images were promptly forgotten, though
the thought was firmly impressed on memory. So far from
accepting the view that imageless thought is automatized
thought, I should be inclined to believe that a new thought is
characteristically imageless, and that it attaches itself second-
arily to a word or other convenient symbol, and is more apt
to occur with an image when it is somewhat familiar than
when it is new.
Still another interpretation of imageless thought, or of the
observations that purport to reveal it, presents a serious
obstacle to our progress. Frequently such statements as
these are contained in the subject's retrospective report: "I
thought of such and such an object," or, "I thought that such
and such was the case," this being the extent of the subject's
description of his experience, except for the purely negative
statement that no images were present. The objection has
been raised by Diirr,1 von Aster,2 and Titchener,3 that in
1 Zeitschrift f. Psychol., 1908, 49, 313-340.
2 Ibid., 56-107.
3 Op. cit., p. 147.
A REFISION OF IMAGELESS THOUGHT 7
such reports the subject is not playing the game. He has
fallen from psychological description into the commonsense
habit of telling what he has been thinking about. He has
committed the Kundgabe or expression error: instead of
describing his thoughts, he is expressing them. He has com-
mitted the stimulus or object error, and, instead of describing
consciousness, is mentioning the objects with which con-
sciousness was concerned. Confronted with this objection,
the subject is apt to reply that he has done his best, that what
was present in his mind was precisely the fact or object men-
tioned, and that if he is forbidden to refer to the object, all
he can do is to hold his peace. Though this reply fails to
satisfy the critic, there is something to say in the subject's
behalf. Suppose, for the sake of argument, that the specific
thought content exists: how would you propose to describe it?
You offer the subject his choice of sensory terms, but these
he rejects as not fitting the case. If then you exclude refer-
ence to objects, you have nothing further to offer him beyond
a few vague and negative terms, such as "imageless," "pe-
culiar, unanalyzable state," etc. In fine, the objection has
force only on the assumption that the state should be de-
scribed in sensory terms, and that non-sensory content is
non-existent. It prejudges the case.
It is curious that the presence of the stimulus error in
reports of images is not treated with a similar seriousness.
Seldom in the literature will you find an image really described.
Instead of an analysis of the visual picture as composed of
colors and shadings in a certain spatial arrangement, instead
of an analysis of the auditory image as consisting of a sequence
of elementary sounds, you read of "a visual image of a
Massachusetts town," or of "an auditory image of the ex-
perimenter saying 'subordinate concept." If it is com-
mitting the stimulus error to report a "thought of" such and
such an object, it is equally committing it to report an
"image of" the object. A strictly descriptive regimen would
require the subject, one would think, to exclude all reference
to the object in the one case as in the other.
Yet consider the situation of an observer who is forbidden
8 R. S. WOOD WORTH
to refer to the object in describing his images. He would
have to confine his report to such statements as "a bright,
somewhat variegated spot against a dark ground," omitting
to state that this was an image of his friend's face. Yet, if
the image, whether faint or vivid, schematic or detailed, was
for him, at the moment, an image of his friend's face, can he
properly describe the consciousness of that moment without
reference to his friend? No question of the logic of meaning
is here involved, but a mere question of fact: Was or was not
a reference to the object present in the momentary conscious-
ness; and, if so, can the state be described without reference
to the object?
The same question arises when we have a presented object
instead of an image. I hear a noise from the street and say,
"There is a horse galloping past." This is a commonsense
reaction which makes no pretense of describing consciousness.
But suppose I do attempt to describe consciousness. It is
then, perhaps, in order for me to tell exactly what auditory
sensations I had. If I do this as well as possible, and find
nothing further, such as an image, to report — have I then,
with my inventory of auditory sensations, fully accomplished
my task of describing consciousness? It would seem not, if
I actually was conscious of a galloping horse, while my report
makes no mention of this object. It is all very well to warn
me of the stimulus error if I show a tendency to go beyond my
momentary experience and tell something about the horse
which may be objectively a fact but was not present in my
mind at the moment; but if I stick closely to the momentary
experience, reference to the object is quite in order and in fact
indispensable; for, as a matter of fact, reference to the object
was probably the most prominent part of the experience. This
is equally true in the case of an image, and I must conclude
that an observer is perfectly justified in reporting an "image
of his friend's face," and that he could not omit this reference
to the object without badly mutilating the experience. If so,
the observer who reports the "thought of such and such
an object" is equally within his rights. He may have omitted
something which a complete description should include, but
A REVISION OF IMAGELESS THOUGHT 9
he has, in all probability, reported the most prominent datum
of his momentary consciousness.
One further 'important objection to the doctrine of image-
less thought is contained in the teaching of such men as
James, Ebbinghaus and Dewey. In speaking of non-sensory
content, we have neglected to define sensation, or, worse yet,
we have, according to these authors, fallen into the error of
excluding relations, forms, patterns, meanings from our
concept of sensation, and then being badly put to it to explain
how they get into perception and thought. It is impossible,
we are told, to draw a line in sense perception between what is
sensation and what is perception; and there is therefore no
excuse for speaking of non-sensory content in sense percep-
tion, nor for speaking of such content as present in thinking,
unless we are ready to make the improbable assertion that
positive content is vouchsafed us when withdrawn from the
world of sense that can never be experienced in the presence
of physical objects.
Instead of attempting to meet this- objection directly,
I propose to go on with a positive interpretation of imageless
thought, in the hope that it may avoid the difficulty, and
ultimately find a legitimate ground for the distinction between
sensory and non-sensory.
To reach a positive interpretation that shall have any
real significance, it is essential to turn away from the isolated
fact thus far considered, and seek other facts which may be
brought into relation to it. A hint as to the most profitable
direction in which to seek for related facts is afforded by the
following consideration. Thought deals largely with data
derived from past experience. New ideas may certainly
be generated in the process of thinking, but in very large
measure the content of thought is provided by memory; and
it is usually this memory content which appears in the image-
less form. It may then be profitable to bring our rather
extensive knowledge of memory into relation with the phe-
nomenon of imageless thought; and it is in that direction that
I propose to search.
On examining the way in which recalled facts present
io R. S. WOODWORTH
themselves, we are at once struck by something that broadens
the outlook considerably. It is not only in thinking, properly
so called, that facts come to mind without images, but in
the most commonplace acts of memory. I recall, without
visual, verbal or other observable images, what I have in my
pockets, where I left my umbrella, whether my neighbor is at
home today. This imageless recall is with some individuals
quite the rule. The facts are clearly enough present in mind,
but if there be any image it is so excessively dim as to elude
detection. Such imageless recall is indicated though perhaps
not fully demonstrated by some of Galton's results; and Miss
Martin has recently1 given a clear demonstration of the
existence of memory content that is "unanschaulich."
In imageless thought, then, the imagelessness has nothing
particular to do with the thinking process; and we are per-
mitted to drop, with some relief, the elevated tone that has
sometimes seemed appropriate to the topic. Thought is
imageless because its data are recalled in an imageless form,
and not because it does not thrive in a sensory atmosphere.
Much effective thinking occurs in the physical presence of
its object. The use of the word " thoughts" to denote non-
sensory content is unfortunate, for the words "thought" and
"thinking" customarily denote a certain mental function or
group of functions, and cannot easily be restricted to any
particular sort of content. The best word would be one that
suggested recall rather than thinking; but I am not at present
prepared to suggest a suitable nomenclature.2
1 Ze itschrift f. PsychoL, 1912, 65, 417-490.
2 Unless the following suggestion can be seriously entertained. It has long ap-
peared to me that we psychologists were on the wrong track in our selection of technical
terms. Our custom is to choose some term of common usage that may convey to the
uninitiated a suggestion of the technical meaning newly attached to it. The trouble
is that the untechnical usage continues alongside of the technical and tends to cause
confusion; until finally psychologists are driven to exclude the untechnical use from
their discourse, and thus lose a very convenient tool of expression. It is nothing less
than a scandal, for example, that the word "feeling" should have been so refined
in usage that the psychologist can no longer speak of a "feeling of hesitation," and
scarcely of a "feeling of familiarity," without an apology and the dread of being mis-
understood by his colleagues. The older sciences, with their greater need for an
extensive technical vocabulary, have gone to work in quite a different way. They
either take unfamiliar Greek and Latin words and derivatives, or they set apart
A REVISION OF IM AGELESS THOUGHT n
What, then, is it, in general, that is recalled? An old
standard answer is that we recall our past experiences. Ob-
jection has several times been raised to this answer within
the last two decades; but the following line of criticism is
perhaps new. In experiments on testimony, or on "incidental
memory," the subject is found to be incapable of recalling
much that has been before his eyes, and even within the
general scope of his attention. If he could call back his
original experience, it would seem that he could give the
testimony required of him. A specially instructive experi-
ment, for our present purpose, is that of Thorndike,1 who
asked his subjects to call up an image of a certain scene, as
of the front of a familiar building, and then, after they had
estimated the vividness of their images, asked them specific
questions, as to the number of pillars in the facade and similar
details. He found a marked inability to answer the specific
questions, even on the part of individuals with very lifelike
images; and, in fact, there was little or no correspondence
between vividness of image and correctness of report on
details. I have frequently repeated this experiment with the
same results. I have never found an individual able to read
off the number of pillars from his image. Only those could
tell the number who had at some time counted them; and
other subjects protested that it was not fair to expect them to
find the number of pillars in the image, when they had never
counted them in the original. All this seemed highly sug-
gestive. It suggested that only that was recalled which had
been noted in the original experience; and that even vivid
some proper name to serve the special purpose. Thus they have their watts and
volts and ohms and amperes, terms regarding the meaning of which no one need
ever be in doubt. Such terms are much better than "thoughts," or than "Bewusst-
seinslagen," with its doubtful translation of "conscious attitudes." I would propose,
accordingly, to follow the lead of physics and chemistry; and since Bewusstseinslagen
were first reported and defined in the work of Marbe and his associates, I would suggest
calling them "marbs," the term to be defined for all time by reference to the original
description by Marbe. Similarly, since the "thoughts" were gradually brought to
light by the school of which Kiilpe was the guiding spirit, I would suggest calling them
<*kulps," defining this term similarly by reference to the original works. These terms
are certainly beautifully compact and euphonious, and those who can bring themselves
to use them will find them very convenient.
1 /. of Philos., 1907, 4, 324.
12 R. S. WOODWORTH
images, described as being fully equal to the actual experience,
were in fact something quite different.
I was thus prompted to undertake an examination of
images and other content of recall, in order to see how far
they could be described as revivals of past experiences, and
how far they consisted of facts noted in the past. I set
myself to recall events from my past life, and in other cases
to recall persons, buildings, towns, and such specific facts as
the exact colors of postage stamps, the quality of a friend's
voice, the shapes, tastes, odors, etc., of a great variety of
objects. What I got was sometimes to be called an image and
sometimes not; but in all cases, with a few doubtful exceptions,
it consisted of facts previously noted. When I say "facts,"
I do not mean verbal statements of facts, but a direct con-
sciousness of some thing, quality, relation, action — of some-
thing which I had observed in the original experience. I did
not get back experiences as concrete totals, but only facts
which I had discriminated out of those totals. In the original
experiences, those facts had had a concrete setting or back-
ground; but this setting was not recalled. The facts were
recalled in isolation.
Often, indeed, a rudimentary setting was present, con-
sisting of either a personal reference, or a spatial reference,
or both. By "personal reference" is meant that the fact
was recalled as my own experience, or that the relation of the
fact to me, or my attitude to it, was recalled along with the
fact. By "spatial reference" is meant that an object was
recalled as being to the right or left, or in a certain town, or in
a certain direction from my position at the time of recall.
Spatial reference was more frequently present than personal.
Neither was universally present; and, aside from them, no
setting was recalled. It frequently happened that several
facts derived from the same experience, or from different
experiences, were recalled almost or quite simultaneously, so
that the recall was richer than would be suggested by the
expression, "isolated fact." Nevertheless all of these facts
had been previously noted, and they did not bring their con-
crete setting back with them.
A REVISION OF IM AGELESS THOUGHT 13
As an example of my results, I will cite the recall of a
colleague speaking in faculty meeting. What I got was a
certain quality of voice and precise manner of enunciating,
rather different from the conversational tone of this individual.
There were no words nor particular vowel or consonantal
sounds present in recall, but simply the quality of the voice
and enunciation. I got also the fact that the speaker was
speaking as chairman of a committee, and something of the
rather critical attitude of the faculty towards him, these facts
being recalled in the "imageless" way. Besides, I got a
spatial reference, in that the speaker was located in a certain
position with respect to my position in the meeting; and a
vague personal reference amounting to an attitude of support
or well-wishing. Beyond this, nothing. No visual back-
ground of faces or furniture, no auditory background of
words spoken, no somesthetic background of myself sitting.
Among the facts thus recalled in relative isolation and
without concrete setting were the following:
Of persons: shape of head or of nose, breadth of face, color
of eye, curliness of hair, blotchiness of complexion, facial
expression, tone of voice, trick of gesture, " smoothness" of
manner, social position, ability, industry, relation to myself,
as being friendly or unfriendly, a superior or dependent,
agreeable, a bore, etc., or as having been seen recently or long
ago.
Of buildings: location, size, color, material, architectural
style.
Of towns: location, general topography, old or new style,
abundance of shade, holiday atmosphere, quietness, associ-
ation with certain events.
These facts run the gamut from simple to complex, and
from sensory to abstrusely relational. They are so varied as
to indicate that any observed fact can be recalled in isolation.
Among the striking instances of isolation were recall of the
color of an object without its shape, of its shape without its
color, of its gloss or shading without either color or shape.
The following interpretation seems scarcely more than a
restatement of these results. An actual situation presents an
1 4 R. S. WOOD WORTH
almost unlimited variety of facts or features, of which an
observer notes a few, the rest remaining undiscriminated in
the background and giving the concrete setting of the features
noted. Later, he may "remember" the situation, but this
is not to reinstate it in its original multiplicity and continuity.
He recalls the features which he observed, or some of them,
but not the great mass of material which remained in the
background. Lacking this setting or background, he is not
in a position to make any fresh observations in recall, and
thus arises the weakness of incidental memory.
If generalized to cover all cases in all individuals, this
statement does indeed go beyond the evidence at hand.
But if the possibility of an occasional recall of the concrete
setting is left open, and the assertion simply made that an
observed fact is often recalled without its original setting,
this conclusion, though modest, is sufficient to furnish a
positive interpretation of imageless recall.
Were it true that a recalled fact always brought with it
its original setting, then, indeed, all recall would involve
sensory imagery. But if a fact is recalled in isolation, it
depends on the nature of the fact whether the recall would be
called imaginal or imageless. If the fact lay as it were on
the sensory surface of things, such as color or tone, its recall
would usually be spoken of as an image. If the fact lay below
the sensory surface, as the fact that a speaker was exagger-
ating, or speaking as chairman of a committee, an isolated
recall of this fact would be unhesitatingly pronounced image-
less, unless, to be sure, it were accompanied by a verbal or
symbolic image derived perhaps from another source than
the original setting of the fact. The definitely imaginal and
the definitely imageless are the extremes of a series, between
which lie many intermediate facts difficult to place in either
class. The expression of a face, the composition of a painting,
the style of a building or piece of music, recalled in an isolated
way, are difficult to classify.
If you set yourself to discover what are the objects of your
attention in a sensory experience, you will usually find that the
actual sensations are less prominent than the things signified
A REVISION OF IM AGELESS THOUGHT *5
by them. You are more conscious of the horse galloping
past than of the actual noises that you hear. When, there-
fore, you later recall hearing a horse gallop past, it is not
surprising that the thing signified should be recalled more
distinctly than the noises; and you are left in doubt whether
to class the recall as an image or not. This is a type of
numerous cases. An observed feature of a situation often
lies partly "on the sensory surface" and partly below, and
the observer does not take separate note of the sign and of
the thing signified, but perceives them together as a single
fact. His recall of the fact may then partake both of the
sign and of the thing signified, though the sensory flavor is
usually weakened in recall. The distinction between im-
aginal and imageless, between sensory and non-sensory, is
not perfectly sharp, and appears, from our present point of
view, to be of minor significance, the main principle being
the isolated recall of observed facts.
I ought really to rest content with the conservative state-
ments that precede, and leave imageless recall as an incident
to the occasional, or frequent, recall in isolation of previously
noted facts. But in the interests of a more clean-cut theory,
I am tempted to more radical and general statements. I
propose to strike out boldly and formulate a theory, hoping
that, whether acceptable or not, it may prove a stimulus to
thought and perhaps to experiment.
The first step towards this theory is to generalize the
conclusion derived from observations already cited, and to
offer the hypothesis that all recall is of facts previously noted,
freed from the concrete setting in which they occurred when
noted. This generalization I hold to be correct for my own
case, and, though the testimony of many individuals regarding
their imagery is on its face in flat contradiction with mine,
the objective test of incidental memory seems to show that
there is something radically wrong with their testimony.
My generalization has the advantage of squaring with the
facts of recall as objectively tested, and the only difficulty is
to explain away the introspective reports of images "fully
equivalent to actual experience," and of "living over the past
as if it were present."
1 6 R. S. WOODWORTH
Without pretending to do full justice to this testimony, 1
must for the present content myself with a few remarks.
Undoubtedly a person may become deeply absorbed in a
remembered experience, because of its great interest for him.
Now his present interest is probably the same as that which
dominated him in the original experience and led him to
observe and react to certain features. If, his interest reviving,
he gets back these features and reactions, he has the essentials
of the original experience from his own point of view, and
satisfactorily lives it over again, even without the concrete
background, the absence of which, in his absorption, he would
not notice, any more than he noted its presence in the original
experience.
As to the vivid image, said to be "in all respects equivalent
to the actual scene," we undoubtedly have, in such a case, a
revival of personal attitude and emotional value, which alone
are enough to create a strong atmosphere of reality. We
must also recognize that what an artist might call the general
effect of a scene is as much a fact to be observed as any
other. The features which can be analyzed out of a situation
are not exclusively details, but include broad effects and syn-
theses and anything that can be the object of attention.
If now you recall the emotional value and general effect of a
scene, along with some of the colors and other previously
noted details, you perhaps have enough to make you testify,
rashly, that your image is in all respects equivalent to the
actual scene. A test of incidental memory would soon
convince you that the "equivalence" is an illusion.
It is also true that a person may observe a scene in such
detail as to recall a great number of its features; and he might
express the wealth of his recollection by asserting that he
revived the entire experience; but, so long as what he recalls
is what -he previously observed, he offers no exception to the
rule that has been formulated.
We have not yet by any means exhausted the relevant
information to be derived from studies of memory. Evi-
dently we should be much helped in any study of recall by
having at hand a report of the process by which what is now
A REVISION OF IM AGELESS THOUGHT 17
recalled was originally learned. We should be helped in
our present inquiry by knowing whether "impressing a thing
on the memory" consists in simply standing before the thing
and letting it "soak in," or whether it consists in reacting to
the thing by observing its characteristic features. It may
be said at once that studies of memorizing give little sign of a
purely receptive attitude on the part of the learner, and much
evidence of a reactive and analytical attitude. Meumann
emphasized the importance of the "will to learn." A subject
might attentively examine a list of nonsense syllables, and
yet make little progress in memorizing it unless his will to
learn were excited. Now the " will " can scarcely be conceived
as acting without means or tools; and its tools consist of
various specific reactions to the matter set for memorizing,
the reactions varying with the material and with the test of
memory that is to be met. Some of these reactions may
properly be called motor; here would be classed the rhythm,
accents, pauses and vocal inflections that are read into the
list by the learner. But in large measure the reactions are
of the perceptual sort, and consist in observing positions,
relations, patterns, meanings, in the matter to be learned.
The recent studies of Miiller throw all these factors into clear
relief. Memorizing is very largely a process of observation,
of noting those features of the material that will serve to
hold it together in the desired way. Some of these features,
such as patterns and relations and the nearer-lying meanings,
are, as it were, found in the material itself; while other
features, the more far-fetched meanings and associative aids,
are imported from without; but this distinction is only one
of degree.
The reactions made in learning, it should once more be
said, are specific, and adapted not only to the material learned
but also to the kind of memory test that is anticipated. If the
subject expects to recite a list of words or syllables throughout,
he observes positions, sequences, patterns and relations that
will serve to bind the whole list together. If he expects simply
to respond to each of the odd-numbered words in the list by
giving the following word, as in the method of paired associ-
1 8 R. s. WOOD WORTH
ates, he takes each pair as a unit, and observes characteristics
of the pair that bind it together, but neglects the sequence of
pairs. If he expects to be called upon to recognize the in-
dividual words of the list, he fixes his attention on them singly,
observing in each, as far as possible, some character that may
serve to impress it. There is no one uniform process of
learning, and the will to learn cannot be conceived as a general
force or agency. What we find in memorizing is a host of
specific reactions, largely of the perceptual sort.
I may be permitted to cite the results of a little experiment
designed to test this matter. I read a list of twenty pairs of
unrelated words to a group of 16 adult subjects, instructing
them beforehand to learn the pairs so as to be able to respond
with the second of each pair when the first should be given as
stimulus. But, after reading the list three times, I told them
that they should, if possible, give also the first word of the
following pair on getting the second word of the preceding
pair as stimulus. I then read the first word of the list, waited
5 seconds for the subject to recall and write the second word;
then read this second word, and waited the same time for them
to recall and write the the third word, namely, the first word of
the second pair; and so on through the list. The results were
most definite: the second members of the pairs were correctly
recalled in 74% of all the cases, but the first members were
recalled in only 7% of the cases. The subjects reported
that this great difference was apparently due to the fact
that they had examined each pair with the object of finding
some character or meaning in it; whereas they had neglected
the sequence of pairs as being of no moment.
This result is instructive in several ways. It indicates,
first, that the will to learn operates not by favoring a general
receptive or memorizing attitude, but by leading to specific
reactions of the observational type. It serves, next, to fortify
the results of other experiments on "incidental memory."
Here the objection cannot be raised that the incidental matter
that is not recalled was never attended to; for the first words
of the pairs were attended to as well as the second. The
experiment also shows the unsatisfactory character of Ward's
A REVISION OF IMAGELESS THOUGHT 19
conception of the process of learning. He has said that
associations are formed by the movement of attention from
one to the other of the terms associated. But here attention
moved from the first to the second member of a pair, and
thence to the first member of the next pair; yet the first move-
ment seems to have established a strong association, and the
second, comparatively speaking, none. Evidently something
much more specific than a mere movement of attention has
been in play. The members of a pair are associated by the
sequence, connection or meaning that is found in the pair.
Finally, this experiment serves to strengthen doubts that have
often been raised, especially by the work on incidental
memory, regarding the adequacy of contiguity in experience
as an associating force. Here the contiguity between the
members of a pair was scarcely greater, in matter of time,
than that between successive pairs; yet the association within
pairs was strong, and that between successive pairs almost
negligible. Since the associations within pairs gave 10 times
as good a score as those between pairs, we may perhaps say
that mere contiguity does not contribute more than one tenth
of the whole associating force, the remaining nine tenths being
contributed by the noting of suitable features in the material.
Even the small fraction thus left to contiguity does not neces-
sarily belong to it; for it is not improbable that the sequence
and relation of successive pairs were sometimes observed.
In fact, of the few correct recalls of first members, practically
all occurred at the beginning or end of the list of twenty pairs;
and it is quite likely that, in these favored positions, attention
was occasionally directed to such incidental matters as the
sequence of pairs or their positions in the list. Except at
the ends of the list, the score for first members was only 1/85
as good as that for second members of the pairs; and this
fraction, rather than i/io, probably represents the proportion
of the total associative force that should be assigned to mere
contiguity; though even this is a doubtful concession.
It may be considered superfluous bravery in me to chal-
lenge the doctrine of association by contiguity, in addition
to all the other enemies already on my hands; but, in reality,
20 R. S. WOODWORTH
I have this doctrine on my hands at any rate. For if contigu-
ity in a momentary experience is a strong and sufficient asso-
ciative force, then any item that is later recalled will in turn
recall its contiguous items and redintegrate the whole expe-
rience or a large part of it, and my hypothesis that what is
recalled is observed facts without their setting would become
untenable.
Now association by contiguity has played a worthy and
important part in the development of psychology, and its
attempt to absorb into itself all other laws of association has,
in my opinion, been a success. Things become associated
only when they are contiguous in experience. That is to say
that contiguity is a necessary condition of association. But
is it a sufficient condition? There is little in the experimental
work on memory to indicate that it is sufficient, and much
to indicate that it is not usually depended on to accomplish
results. The things to be connected must be together, in
order to arouse the reaction connecting them; but, unless
they arouse some such reaction, they do not become con-
nected, except it be very weakly. The reaction may be
described in a general way as a reaction to the two things
together; it is perhaps sometimes a purely motor reaction,
but most often, I believe, is rather to be called a perceptual
reaction, consisting in the observation of some relation
between the two things, or some character of the whole
composed of the two taken together. In any case, the
reaction is specific; and it is this specific reaction, rather than
any general factor like contiguity, or the movement of at-
tention, or the will to learn, that does the work of association.
To judge from the memory experiments, then, what is recalled
is what has been noted — not past experiences in their totality,
but definite reactions which occurred in those experiences.
This conclusion is perhaps even more clearly indicated by
experiments in the learning of nonsense drawings than in the
more usual work with linguistic materials. An instructive
experiment is that of Judd and Cowling,1 who exposed a
rather simple drawing for successive periods of 10 seconds,
l" Studies in Perceptual Development," PSYCHOL. REV. MONOGRAPH 34, 1907,
349-369.
A REVISION OF IM AGELESS THOUGHT 2*
requiring the subject to reproduce it as well as possible after
each exposure. ' The results, both objective and introspective,
showed that the subject usually got first the general character
and shape of the figure, and, continuing his analysis, noted
one fact after another, until a sufficient number of facts was
known to make a satisfactory reproduction possible. There
was no evidence of an inner reproduction of the entire sensory
experience, from which the subject might read off such in-
formation as he required. In a somewhat similar experiment,
T. V. Moore1 called for the learning of a series of simple
drawings. He supposed at the outset that a group of figures
would be memorized by visual imagery, but experience taught
him that there was another factor that was a powerful aid
to memory. This was "a more or less complete analysis of
the figures, an analysis which it is utterly unneccessary for
the subject to put into words." It consisted in noting the
parts and composition of the figures and their resemblances
to familiar objects. He then undertook to compare the
efficiency of memorizing by visualization with analysis
excluded, and by analysis without visualization; and found a
uniform superiority of the analytic method over the visual-
izing. But he also found that it was impossible to exclude
analysis altogether. " Associations crop up spontaneously,"
he writes, "and one simply cannot exclude all analysis of the
figure. ... It is much easier to memorize by analysis to the
exclusion of imagery than vice versa" He believed, however,
that learning by visualization, i. e., by forming an image
which should be a "more or less perfect replica" of the visual
sensation, was a real process. Under the circumstances, it
was evidently impossible for him to prove this; for if analysis
occurred spontaneously — and one has only to look at a
drawing to realize how inevitable it is to note either details
or broader characteristics — and if also analysis was a more
powerful memorizing agency than visualization, it remains
possible that all the learning was accomplished by analysis.
The reality of the strictly visualizing or photographic process
of learning is, I believe, still open to doubt. It is certainly
1 "The Process of Abstraction," Univ. of California Publications in Psychology,
1910, i, 139-153.
22 R. S. WOODWORTH
impossible to avoid perceptual reactions, and to assume the
purely receptive attitude of a photographic plate.
Miss FernakTs data on the memorization of pictures1
show that even good visualizers depend largely, at least, on
specific observations of the features which were later re-
membered; and her results on the recitation of letter-squares
in changed orders2 showed that even the best visualizers among
her subjects were unable to do what it had been supposed was
the prerogative of a visualizer to do, namely, "see the whole
set of letters at once and simply read them off" in the changed
order. She does not doubt the existence of persons able to
accomplish this feat, but believes that they must be rare.
This matter of visualization evidently requires further study,
but the possibility is still open that even the best visualizer
does not carry away a photograph of the scene, or replica of
his visual sensation, but an image which amounts to a syn-
thesis of specific observations, including observations of
broad effects and observations of parts and their relations.
But it is time that I brought my theory out of hiding and
placed it squarely before you. I call it, for lack of a better
name, the mental reaction theory, or perhaps the perceptual
reaction theory. Its basic idea is that a percept is an inner
reaction to sensation. I call it a mental reaction to dis-
tinguish it from the motor reaction which several psychologists
have put forward as being important in attention, perception,
association and the like; for it appears to me that these sug-
gestions, while on the right track in insisting that reaction
is dynamically important, have mistaken the locus of the
reaction, and so are unable to account for the conscious
content that appears in these mental activities. This mental
reaction is not, however, of the nature of an associated sen-
sation, appearing as an image, as if the visual sensation of an
orange, to give the percept orange, must reproduce the
sensations of handling or tasting the orange. Nor, on the
other hand, is the perceptual reaction an emphasis or pattern
or meaning residing in the given sensations. It is something
new, not present in the sensations, b ut, theoretically,as
1PsYCHOL. REV. MONOGRAPH 58, 1912, 8iff.
2 Ibid., p. 71.
A REVISION OF IMAGELESS THOUGHT 23
distinct from them as the motor reaction is. It adds new
content which cannot be analyzed into elementary sensations;
so that the sensory elements, which are often held to supply,
along with the feelings, all the substance of consciousness, in
reality furnish but a fraction of it, and probably a small
fraction. Each perceptual reaction is specific, and con-
tributes specific content. In recall, it is these perceptual
reactions that are revived, and not sensation; and therefore
the content of recall is never, in the strictest sense, sensory.
Nevertheless, as was said before, some percepts lie, as it
were, nearer to sensation than others, so that the distinction
between an image and an imageless recall, while not perfectly
sharp, is still legitimate.
It is possible that this theory may appear not so radical
after all, and not worth the expenditure of so much breath;
for all will perhaps admit that a percept is, in some sense, a
reaction. It is therefore my duty to show that the theory is
worse than it seems, and this I shall attempt to do in the case
of patterns or Gestaltqualitaten. It has long been known
that the same pattern (for example, a melody) can sometimes
be found in different sensory complexes, and it is also true
that different patterns can be found in the same sensory
complex, as in the case of the dot figure. A rather difficult
problem is thus raised, for one would think that the compound
would be determined by the elements. But the real crux of
the difficulty is to get some conception of a pattern or of a
compound, to show what is meant by the togetherness or
grouping of the elements. There are three theories that
attempt to solve this puzzle, that of synthesis, that of systasis
or mere togetherness, and that of synergy, which is none other
than the mental reaction theory. The synthesis theory brings
in the subject or ego to put the elements together; the systasis
theory rejects this deus ex machina, and says that the ele-
ments merely are together, or get together and so constitute
the compound or pattern; the synergy theory holds that the
elements act together, as stimuli, to arouse a further reaction
which is the pattern. The synthetic theory occupies a weak
position, since, unless the systatic theory succeeds in showing
24 R. S. WOOD WORTH
what is meant by the elements being together, there is no
advantage in saying that something puts them so.
Now it is difficult to understand what can be meant by the
elements being together or getting together so as to produce
the group and pattern. If the group included the whole
momentary content of consciousness, we could say that
being together meant simply being simultaneously present,
and speak of the pattern as a character of the whole conscious
moment. But the group does not include the whole of con-
sciousness, but — as in the case of three dots among a larger
number, seen for an instant as a triangle — may occupy but a
small part of the conscious field. The pattern is not the
pattern of consciousness, but a pattern within consciousness.
Nor will it help matters much to substitute for consciousness
the field of attention; for the extent of a group may be either
greater or smaller than that of this field; and, besides, a
familiar pattern, such as a melody or arrangement of lines or
dots, may come to consciousness quite outside the field of
attention. Apperception, then, in the Wundtian sense, does
not explain groups and patterns nor give them any intelligible
meaning. But if we lay aside apperception and try to describe
groups and patterns in terms of their constituent elements,
we are in no better case. What is it that changes when the
pattern changes, the elements remaining constant in quality,
intensity and spatial position? This question is as serious
for the synthetic theory as for the systatic. The synergy
theory cuts the Gordian knot by admitting at once that there
is no change in the elements. In fact, there is no real grouping
or pattern of the elements; they neither get together nor are
put together by some higher agency; but some of them simply
act together, as a complex of stimuli, to arouse a perceptual
reaction which constitutes the grouping and pattern. The
pattern is numerically distinct from the elements, as a motor
reaction is distinct from the complex of stimuli that arouses
it. What pattern shall be aroused at any moment depends
on the readiness of different perceptual reactions to be aroused,
and thus on such factors as frequency and recency of past
exercise, fatigue and present interest and control. In short,
A REVISION OF IMAGELESS THOUGHT 25
the synergy theory proposes to extend to patterns, and to all
percepts, the same explanation that is accepted for such ad-
mittedly mental reactions as the sequence of one idea after
another. No one doubts that one idea may represent a
stimulus for the arousal of another idea, nor denies that the
aroused idea is numerically distinct from the stimulus idea
and adds new content to it. It is the same with sensation
and perception, except that the reaction is usually very
prompt and the perceptual content intimately fused with the
sensational. The fusion is so complete that the pattern
seems to lie right in or among the dots, as the galloping horse
of an earlier illustration seemed to be actually heard in the
series of noises.
But now, finally, I suspect that the party, which allowed
me to proceed some time ago without coming to terms with
their demand for a definition of sensation, will no longer be
restrained. They will insist on taking the floor and address-
ing you as follows: "The speaker is certainly right in calling
a percept a reaction; that is too obvious a fact to need dis-
cussion. But we ask, A reaction to what? And our answer
is, To the physical stimulus. This ' sensation' that the
speaker has interpolated between the physical stimulus and
the percept is pure gratuitous assumption. There is no
warrant for it in introspection, for he himself admits that the
sensation and the percept content. are intimately fused. We
regret that he has fallen into this obsolescent way of speaking,
and would suggest that, in reviewing his remarks, you use the
blue pencil of the censor wherever the word 'sensation'
occurs."
This objection is almost too serious to be dealt with in
brief. I should freely admit that sensation and percept
cannot be distinguished by direct introspection. Yet there
are introspective facts that make the distinction appear
legitimate. When we hear the galloping horse, we are not
only aware of the horse, but we are able to state that we hear
him. It is not quite correct to say that we get only the
meaning, for we know also the sense by which we get the
meaning. So, again, when we have changing percepts of the
26 R. S. WOODWORTH
same stimulus, as in the case of the dot figure, the change of
pattern does not amount to a complete change of the figure,
but there is a constant substratum underlying the changes;
and it seems appropriate to speak of this as sensation. In
recall, even the best images lack something when compared
with actual sensory experience. They lack body and in-
cisiveness; and it appears probable that this lack is nothing
more nor less than a lack of sensation, or, in other words, that
the real sensory process is not resuscitated in the image.
But the concept of sensation might never have arisen in a
purely introspective psychology. At bottom it is a physio-
logical or psychophysical concept. Sensation is that con-
scious content which is in closest relation to the physical
stimulus. It is the primary response to the stimulus, and
may be followed by secondary responses. Neurology gives
good ground for such a distinction, in tracing the sensory
nerves to certain limited areas of the cortex, and finding the
rest of the cortex to be only indirectly connected with the
sense organs. Destruction of the cortical receiving station
for any sense abolishes all conscious use of that sense, while
destruction of neighboring areas, without making a person
blind, for example, abolishes his power of reading, or his
power of recognizing seen objects, or his power of orienting
himself in visual space. Such perceptions are apparently
secondary reactions, while the primary reaction, correspond-
ing to the activity of the receiving station, is precisely that
which distinguishes a person who is word-blind and object
blind, from one who is totally blind. Here is a person who
sees without perceiving, and here is one who does not see at
all. The difference I would like to call sensation. Sensation,
accordingly, would be the consciousness attending the activity
of the sensory receiving stations of the brain, while percept-
content would be the consciousness attending the activity of
neighboring areas. Besides these secondary reactions, there
are undoubtedly tertiary and further reactions, less and less
directly connected with the incoming sensory impulses. They
need not have a sharply limited localization in the cortex,
yet they must be neurologically distinct, and it may well be
A REVISION OF IMAGELESS THOUGHT 27
that every distinct cerebral reaction is attended by its
peculiar conscious content. I know of no reason in neurology
or psychology for supposing that the elements of conscious
content are contributed solely by the sensory receiving centers.
According to this theory, the sensation aroused by a
physical stimulus must precede the secondary or perceptual
reaction; but the interval need not be supposed to exceed a
hundredth of a second, and could not be introspectively
detected. The fusion of the primary and secondary reactions
in consciousness is a fact which I cannot attempt to explain,
since fusion is one of the fundamental peculiarities of con-
sciousness as contrasted with its cerebral correlates. But
I may perhaps make the whole conception a little more
tangible by reverting to the similitude of photography.
A certain photographer found himself without sensitive
plates, though with his camera, in the presence of a scene
which he much desired to preserve. He therefore focused on
the ground glass at the back of his instrument, and, stretching
transparent paper over the glass, traced some of the outlines
of the optical image. He thus created patterns, which lay
really in his drawing and not in the optical image, but which
were blended with the image as long as the image remained.
He preserved his tracing, and found it to differ from a photo-
graph in containing only the facts to which he had definitely
reacted.
In this parable, the optical image is sensation, which is
gone forever when the physical stimulus ceases. The
tracing is perception, which may be preserved, though subject
to decay. But the fusion of the two, depending in the case
of the camera on the presence of the photographer's eye, is
in the case of sensation and perception more deep-seated
and inexplicable. Finally, the photographer was more
restricted than is the process of perception, since he could
only trace outlines and shadings and perhaps colors, and
could not commit to his drawing the more remote relations
and meanings which can be perceived, and, being later
recalled, furnish the content of "imageless thought."
A NEW MEASURE OF VISUAL DISCRIMINATION1
BY KNIGHT DUNLAP
The tests on which this report is based were carried out
in the Nela Research Laboratory in August and September
of 1914, as an adjunct to other experimental work which will
be reported later. The instrument used was constructed at
The Johns Hopkins University several years ago, and as a
result of work with it since that time has been modified into
the present form, which seems good in principle although
the mechanical operation may still be improved.
The instrument which, for convenience, may be called a
duoscope, consists essentially of a polished crystal of Iceland
spar mounted in a telescopic brass tube, which has an eye-
aperture at one end, and a rotatable ring at the other. The
ring is rotated by a worm-screw with a knurled head, and is
provided with a vernier scale, so that the angle of rotation
may be read to one fifth of a degree. The ring is arranged
to hold a disc fitting within it, so that various forms of objects
may be viewed at a fixed distance from the eye.
The first objects tried were as near linear as possible: a
diamond-scratch on a clear glass disc; a fine glass filament
crossing a circular aperture in a metal disc. These were
tried against various backgrounds, and were not satisfactory
because of the difficulty of securing a line of sufficient uni-
formity and without sheen. A narrow slit was equally un-
satisfactory. Finally a rectangular aperture of appreciable
width was found workable when used against a bright back-
ground.
The Iceland spar crystal gives a double image of the line
(or rectangle) used as an object, and the relation of these
two images may be altered by rotating the ring which carries
the line with it. If the line stands exactly in the refracting
1 From the Nela Research Laboratory, National Lamp Works of the General
Electric Company.
28
VISUAL DISCRIMINATION 29
plane of the crystal, the two images are superposed over their
greater length, their displacement being longitudinal only.
If the line be rotated 90° from this position, the two images are
displaced laterally to the maximal distance possible from the
prism (1.089 mm- witn tne crystal used in the present work),
and if this distance be greater than the width of the line (or
rectangle) the two images are separated.
In this way, with a proper linear object, it would be easily
possible to measure (in terms of the visual angle) the displace-
ment of the images giving just perceptible doubleness; i. e.^
the minimum visibile. The advantage of the device lies in
the exactness of measurement and ease of manipulation, with
the possibility of accommodation for relatively short or 'read-
ing distance.' It is possible to obtain a suitable linear object,
but before I had found one I discovered that observation on a
rectangle of appreciable width (i. e., a relatively wide slit)
is much easier, and have therefore adopted such an object
for the present. The fineness of measurement possible with
this instrument is indicated by the fact that the reading
unit in the vernier scale (one fifth of a degree) corresponds, in
the middle of the scale used in the present work, to 1.6" of
visual angle, or .0029 mm. lateral displacement of the image.
The discrimination of doubleness in lines is of course a
matter primarily of difference-sensibility for brightness. If
(in the case of two bright lines, or two images of a single
objective bright line), there is a perceptible dark stripe down
the middle of the combined lines, they are seen as two; if
there is no dark stripe, as one. The three factors involved in
'visual acuity' as tested by the linear method1 are therefore:
(i) The physical distribution of the light-flux on the retina,
determined by the 'resolving power' of the eye; (2) The dis-
tribution of energy or activity in the physiological image of
the retina, determined by the distribution, in the physical
image, and by irradiation, etc.; (3) The difference-sensibility
for brightness differences. Anything which changes any of
these factors, as a change in the lens system, or in the irradi-
1 In the case of determinations by means of two points instead of two lines, the
situation is different, and the histological texture of the retina may be a factor.
30 KNIGHT DUNLAP
ation, will therefore change the 'acuity,' although the dif-
ference-sensibility remains the same. Thus the practical
usefulness of the acuity-determination depends on thorough
(and perhaps unattainable) control of the conditions of ob-
servation. These matters are so obvious that no further
discussion is needed here.
The method of testing 'acuity' by the double images of a
bright rectangle is now apparent. If the two images (phys-
ical) overlap sufficiently, there is a brighter line in the middle
of the combined images; if they are sufficiently separated there
is a darker line in the middle. It is a relatively simple matter
to determine the points at which the dark line and the bright
line are just perceptible.
The crystal used has a maximal image-separation of
1.089 mm., and the slit was fixed at a distance of approxi-
mately 36 cm. from the eye. The slit was 3 cm. long and
0.77 mm. wide.
The most difficult part of the adjustment of the instru-
ment is the determining of the position in which there is no
lateral, but only longitudinal, displacement of the images.
This determination was made by long series of observations of
the bright line obtained by moving in each direction alter-
nately, determining the middle point from these. The zero
point thus obtained is sufficiently exact for practical purposes;
besides, no more exact method is available.
The background against which the slit was viewed was a
plane disc of plaster surfaced with magnesia, at a distance of
97 cms. from the rectangle.1 This was used first in a darkened
room and illuminated with a beam of light from a nitrogen
tungsten lamp. The lamp being enclosed, the only illumin-
ation of the room was the light diffused from the disc. From
the direct radiation of this light the observer was protected
by a black screen, through which the instrument protruded.
The observer was therefore not in complete darkness, but the
illumination in his direction was low (1/27 to 2 c.p.). Ten
minutes or more was allowed for adaptation, so that the sub-
ject was really in a fair state of darkness adaptation.
1 Tests were made at other distances, but as was expected, the distance proved
not to be a factor in the results.
VISUAL DISCRIMINATION 31
Subsequently the instrument and plaster disc were moved
into a room we'll lighted with daylight, so that measurements
were obtained with daylight adaptation.
In the darkened room five illuminations of the plaster
background were used, giving brightnesses of 3, 10, 36, 82
and 168 candles per square meter.
The observers were: laboratory helper Mr. Eric Martiens-
sen, a high school graduate; Dr. P. W. Cobb; Dr. H. M.
Johnson; and myself. The readings on me were taken by
Martienssen; the readings on the other observers were taken
by me. Usually twenty-five determinations were made in
one sitting. Thus, in the work with five intensities, five
determinations were made on each intensity at a sitting, with
only one sitting a day. Each subject had preliminary practice
in observing. These intensities were taken in a different
order on different days. No practice effect is noticeable in
the measurements of any of the subjects.
The observer started with the bright line plainly visible
and rotated the slit until the dark line was just visible:
then he rotated the slit in the other direction until the
bright line was just visible: or vice versa. The observers
found it easier to make the changes rather quickly. Long
looking caused the difference in brightness to disappear.
It is unfortunate that the instrument was made with the
crystal fixed, and the object rotating. It was so made because
this form allows easier construction, and has advantages in
adjustment of the object for the zero point which was de-
sirable while the instrument was in the provisional stage.
The next instrument will be made with fixed object-holder,
so that the axis of the slit will not change during observations.
The rotation of the axis in these experiments was small,
however, and does not vitiate the results. In the table below,
where the axis is not specified, it was 90°, that is the slit was
vertical when in the medial position, i. e., in the position in
which the images were not displaced laterally.
The readings given in the table are the displacements
from the zero position in visual angle computed from the
averages of the designated number of readings on the scale
KNIGHT DUNLAP
of the instrument. Theoretically, the visual angle should
have been computed for each instrument-reading, and then
the averages of the computed values taken; practically, the
computation for the average of the instrument-readings is
sufficiently accurate. The mean variations are not given,
because I have so far not been able to discover what the true
mean variations are. It is evident that the variations cannot
be referred to the averages, because these vary with the width
of the slit employed, regardless of the uniformity of observa-
tions; nor to the average of the range from dark line to bright
line, because there may be variations in the readings which
do not affect this.
TABLE I
THE INFLUENCE OF BRIGHTNESS AND OF ADAPTATION
i. Martienssen. Left Eye 2. Johnson. Right Eye
A. Dark Adaptation. Av. of 25 A. Dark Adaptation. Av. of 25
Brightness
Dk. Line
Br. Line
Range
Dk. Line
Br. Line
Range
3
7' 10"
6' 44"
26"
6' 59"
6' 46"
13'
10
7' 10"
6' 49"
21"
/ I"
6' 49"
12'
36
7' 10"
6' 50"
20"
6' 59"
6' 49"
10'
82
7' 10"
6' 53"
17"
7' oo"
6' 49"
II'
168
7' o"
6' 54"
13"
7' 2"
6' 50"
II'
B. Daylight Adap. Av. of 20
B. Daylight Adap. Av. of 20
/
6' 51"
ii'
6' 59"
47'
12'
3. Cobb. Right Eye
A. Dark Adap. Av. of 10
4. Dunlap. Left Eye
A. Dark Adap. Av. of 25
3
6' 50"
6' 27'
23"
7' 10"
6' 39"
3.1'
10
6' 53"
6' 31'
22"
7' n"
6' 48"
23'
36
6' 49"
6' 26'
23"
7' n"
6' 49"
22'
82
1 68
6' 51"
6' 52"
6' 34'
6' 33'
17"
19"
7' 11"
/ 12"
6' 50"
6' 50"
21'
22'
Daylight Adap. Av. of 75
7' 4"
6' 49"
15'
In the A parts of Table I are given the general results of
the tests with different brightness of slit under darkness
adaptation, and in the B parts, the corresponding results with
daylight adaptation. Two points are clear. First, that in
general the daylight adaptation gives greater acuity; and
VISUAL DISCRIMINATION
33
TABLE II
INFLUENCE OF ANGLE OF Axis OF RECTANGLE
I. Dunlap. Right Eye
A. Daylight Adap. With lens correction. Av.of 20
Axis
Dk. Line
Br. Line
Range
Dk. Line
Br. Line
Range
80
90
100
125
170
f 2"
7' 2"
7' 2"
7' 2"
7' 6"
6' 46"
6' 45"
6' 42"
6' 42"
6' 50"
16"
17"
20"
20"
16"
B. Daylight Adap. Without lens
Av. of 10
C. Dark Adap. Without lens
Av. of 40
80
7' 4"
6' 43"
21"
7' 4"
6' 40"
24"
I25
7' 10"
6' 32"
38"
7' 17"
6' 22"
55"
170
7' 9"
6' 55"
14"
7' 13"
6' 46"
27"
2. Martienssen. Daylight Adap. Av. of 20
A. Right Eye B. Left Eye
90
7' 9"
6' 53"
16"
7' 2"
6' 51"
n"
135
7' oo"
6' 48"
12"
1 80
Unable to
see lines.
7' 8"
6' 54"
14"
67.5
Unable to
see lines.
7' 43"
7' 22"
21"
3. Johnson. Daylight Adap. Av. of 10
A. Right Eye B. Left Eye
90
(6' 59")
(6' 47")
da")
6' 59"
6' 45"
H"
135
6' 55"
6' 46"
9"
6' 59"
6' 47"
12"
1 80
7' oo"
6' 48"
12"
6' 58"
6' 45"
13"
67.5
/oo"
6' 48"
12"
6' 59"
6' 46"
H"
second, that there is no uniform influence of brightness within
the limits of the conditions obtaining.
The results with the lowest brightness differ appreciably
from the results with the higher brightness, but this is due,
in part at least, to the difficulty in judging with this illumin-
ation when near the line-threshold. This is a condition
which must be distinguished carefully from the raising of the
threshold as such, and was clearly a factor in my own case.
Leaving out the dimmest light, the influence of the brightness
is negligible for Cobb, Johnson, and Dunlap.
The change to daylight adaptation is, however, influential
except in the case of Johnson, whose acuity seems to be
34
KNIGHT DUNLAP
exceptional. Tests with the Cobb acuity-object also have
shown Johnson to have unusual acuity with darkness adap-
tation. It is quite probable that the slight effect of the
increasing illumination in the darkened room was due to the
lessening of adaptation.
The most striking result is the uniform lowness of the
threshold. The average range from dark line to bright line
lies for the most part near 20", and is lower in some cases. In
the ordinary test object, using object lines, the measurement
from fusion to dark line is from 30" to 60". The correspond-
ing measurement in the present case is less than 20"; how much
less cannot be determined, as it cannot be assumed that either
the points of geometrical contiguity, or of physical uniformity
of the images, lie half way between the points at which the
dark line and the bright line respectively appear. Schuster
states that the intensity at the edge of the geometrical image
of a uniformly bright surface must be "half the intensity
observed at some distance inside the edge," because when
two surfaces are placed with edges in contact a uniformly
illuminated surface is obtained.1 Assuming this to be true
of the physical image, it is not necessarily true of the psycho-
logical image, as irradiation and contrast (physiological)
effects occur at the margins of the images. The fact that
both light line and dark line thresholds tend to shift with
darkness adaptation indicates influences of this sort, and
we should accordingly expect the light line threshold in gen-
eral to be nearer the point of geometric image contiguity
than is the dark line threshold: an expectation that is justified
by the facts.
Table II gives the results of tests to find the effects of
lenticular aberrations. My right eye is corrected with a lens
of 0.50 C., axis 80°, prism J<° B.D. Tests were accordingly
made on the eye with and without the correcting lens, in the
astigmatic axis and at 45° on either side. The results (II, I,
A, By and C) show that even a low degree of astigmatism is
detectible by this means and also that my eye is slightly under-
corrected by the lens.
1 Schuster, 'Theory of Optics,' page 151.
VISUAL DISCRIMINATION 35
Tests were carried out on both of Martienssen's eyes
(II, 2). He preferred to use his left eye in any sort of monoc-
ular observation. On being questioned about this he said
he had always used that eye because it seemed more natural.
The tests seem to indicate a slight degree of astigmatism in
the right eye with less in the left. Martienssen's eyes had
never been refracted.1 The degree of astigmatism is not
great, for with uncorrected eyes requiring from one to two
diopters of cylindrical correction, the instrument cannot be
used at all.
Whether the instruments and methods for acuity test
above described will be practically useful remains to be seen.
In the matter of precision and convenience the apparatus
seems superior to devices hitherto in use. The fact that the
results differ from those obtained by means of the several
other devices is immaterial. Measurements of this kind give
comparable results only when the same instruments and
methods are used. Since the duoscope method seems sensitive
to adaptation changes, it may be possible to use it as an adap-
tometer; since it seems not sensitive to brightness changes
over a considerable range, it may be a useful instrument for
practical testing of eyes. It seems especially suitable for
detecting slight degrees of astigmatism, and for detecting
the accuracy with which lens corrections for astigmatism
are made, in experimental work where accurate control of
the observer's eye is required.
1 Since the above was written, Dr. Cobb has refracted Martienssen's eyes, with
the following results (sine midriatic):
O.D. — . 37 cyl. axis 175°
O.S. — . 25 sph. — . 37 cyl. axis 155°.
It is probable that the fact that the duoscope readings for the Right Eye are not
harmonious (e. g., the inefficiency at 67.5), is due to over accomodation.
AN ELECTRO-MECHANICAL CHRONOSCOPE
BY JOHN W. TODD
University of North Dakota
Provided that certain changes are made in its electro-
magnets and that it is skilfully handled the most reliable
chronoscope known is the Hipp chronoscope. But for want
of skilful handling it is not unusual to see a dust-coated copy
of the instrument stored away in a museum for apparatus that
looks nice but is rarely used. Nevertheless the Hipp deserves
more respect. After having invested in the costly piece with
its control-hammer additional, the experimenter should fit it
for service by rewinding its electro-magnets with coarser
wire, by insuring a steady current with a good gravity battery
of 12 cells, by discarding the control-hammer and employing
some type of gravity chronometer for control tests.1 It is
the fineness with which the instrument is designed to record
times that makes it unsatisfactory in inexperienced hands but
that insures reliability when correctly operated.
Even after the corrections indicated are made the instru-
ment must be constantly watched and tested, as a fluctuation
of the current or a slight change in the adjustment of the
delicate parts of the apparatus may produce a chronoscope
variation that will entirely obscure the variations in reaction
time. The instrument responds to all irregularities and is
never popular when operated in a hit and miss manner. Many
attempts have been made to devise a simpler chronoscope
than the Hipp. The special aim has been to construct one
that will eliminate the constant care of control and minimize
1 The manner of making these changes and the reasons for them may be found in
the National Academy of Sciences, 7, 397 ff., 1893 (Cattell and Dolley). After cor-
recting the instrument as indicated these writers found average variable errors for
seven series of ten single tests of the chronoscope as follows: 0.96, 0.8, 0.42, 0.4, 0.64,
0.64, and 0.56". Using the same instrument corrected by Cattell the present writer
in making several thousand reaction tests found an average variation of the chrono-
scope in control tests of about I* ("Reaction to Multiple Stimuli," Archives of Psy-
chology, No. 25, 8, 1912).
36
A NEW CIIRONOSCOPE 37
the possibility of getting out of adjustment. The simplest
chronoscope is 'one so designed as to harness the force of
gravity for marking off units of space that may be given time
values. This arrangement eliminates delicate clockwork
propelling devices and reduces the number of adjustable
parts three fourths. Even after the chronoscope is reduced
to its simplest terms three difficult problems remain, First, to
devise a reliable chronoscope release; Second, to construct a
sufficiently accurate reaction recorder, and, Third, the
greatest problem of all, to put down a chro no metric scale that is
trustworthy.
The chronoscope described in this article consists of a
disc compounded of two adjustable parts (Z)i and Z)2, Fig. i),
which are two circular planes of 1/16 in. brass, 11.5 in. in
diameter, with a concentric semicircle of each having a radius
of 4.75 in. cut away, leaving in each case a marginal area
I in. in width. One of the discs is constant with respect to a
pendulum attached to their common axis while the other is
adjustable to allow the various apertures of a tachistoscopic
attachment described later in this article. These discs are
held rigidly together by a set-screw (S. sc., Fig. 2) and rock
with the vibrating pendulum. In making a chronometric
reading with the electrical arrangement the initial position
of the pendulum, i. e., horizontal, is maintained by the force
3$ JOHN W. TODD
of an electro-magnet (EM, Fig. 2) upon the armature (A^
carried near the base of the pendulum whose socket (P.S.) is
shown. To minimize friction the shaft carrying the discs
and the pendulum has cone bearings (Fig. 2). The pendulum
weights are two cylinders of lead set in brass. The gross
relations of the various parts of the apparatus are shown in
the upper left corner of Fig. I, a lateral photograph of the
apparatus.
Fig. 2 is a diagram showing the details of the chronoscope
P
FIG. 2.
arranged for both mechanical and electrical stimulus key and
reaction key. The figure shows the apparatus set for
mechanical operation. SAi is a copper shoe held firmly
between the contact points of a stimulus key by means of the
insertion of the inclined surface of the lever, Z,, and the re-
siliency of the key shaft, supporting by means of a cord over a
A NEW CHRONOSCOPE 39
pulley the armature, A 2, which holds the pendulum in the
initial horizontal position. SA2 is likewise a copper shoe
between the contact points of the reaction key held firmly
by the reagent's finger upon the button, and holding the arma-
ture (Ai) of the reaction index to its initial position against
the arm of the post, P. The armature, AI, to which the
reaction index (7) is attached moves freely upon the spindle
bearing the chronoscope disc, and by means of a coiled spring
flies against the disc when release is made and is carried with
it allowing the index pointer to escape the arm of the post,
P, by which it is held during the reaction interval.1 When the
pendulum is in the initial position the index point rests upon
the zero point of the chronometric scale. The index may be
thrown back to the zero point from any position on the scale
by pulling upon the cord to which Shz is fastened.
When the stimulus lever (L) is suddenly pulled the stim-
ulus hammer (S.Ss) by virtue of the resilient shaft which
carries it and the rebound of the strong coiled spring near its
fulcrum strikes the solid metallic base a blow emitting a
sound which serves as a stimulus, and whose intensity is
variable by means of set screws. With the drop of the stim-
ulus hammer upon the metallic base and the emission of the
sound the shoe actuated by the strong spring of the release
armature flies from between the points, while the armature
flies back to position As. This releases the pendulum and
carries the chronometric scale in the negative direction count-
ing against the reagent until the index flies upon the disc
marking the close of the stimulus-reaction period.
The difficulty in attempting to devise a mechanical chro-
noscope is to give it versatility. It is not hard to provide a
release that at the same time serves as a sound stimulus, and
that offers no resources for the presentation of touch and
light stimuli. With the mechanical device described above,
however, it is possible to give all three stimuli. This is
shown by Fig. 3, a diagram of the three-stimulus key. The
method of giving the sound stimulus is described above.
1 This type of armature although employed in a somewhat different manner was
first used by Bergstrom in a pendulum chronoscope figured and described in the
PSYCHOLOGICAL REVIEW, VII., 1900, 438 ff.
4o
JOHN W. TODD
When it is desired to present a light stimulus with the me-
chanical chronoscope a light wooden arm long enough to
extend beyond the base of the chronoscope is attached to the
shaft of the reaction key. This wooden arm carries a small
black square (A) operating as a shutter to a I cm. aperture
(Ap) in a black screen large enough to conceal the movements
of the experimenter. When it is desired to use tactual stimuli
the stimulus key, by means of the binding posts, is inserted
into the primary circuit of an induction coil preferably with
condenser. The reaction key is introduced in the secondary
circuit in such manner that the cathode is in the button of
the key and the anode in a shallow vessel of salt water. If
the vibrating armature of the current interrupter of the
induction coil is tied back securely against its adjustable
contact and the reacting finger rests upon the cathode and
the fingers of the other hand rest in the vessel, a shock is felt
in the reacting finger when the primary circuit is broken at
Shi. As the key is figured above it is set for all three stimuli
which would be administered simultaneously with the sudden
pull toward the experimenter of lever L.
By removing the arrangement for light stimulus the pair
of stimuli, sound and shock, may be given together; or by
getting out of series with the induction coil, the paired stimuli,
sound and light, may be given. Likewise it is possible to
eliminate any two stimuli and administer a third singly.
When light is presented singly a small rubber plate is attached
to the posterior side of the stimulus lever (L) to eliminate
the sound. In all these cases it is seen that the chronoscope
release is mechanical, by means of Shi. Fig. 2 by means of
dotted lines from the binding posts suggests the wiring ar-
rangement for electrical operation of the chronoscope. When
A NEW CHRONOSCOPY
41
the current is used the two cords are serviceable to set the
pendulum armature and to throw the reaction index back to
zero.
It is seen that the mechanical device can be operated only
when the experimenter and reagent are at close range. When,
however, it is desired to give stimuli from a distance or to
have the reaction from another room the electrical arrange-
ment mentioned above must be employed. The reaction
movement called for by the apparatus is in all cases of the
break-circuit variety. The chronoscope may be put to the
same tasks that are attempted by any type of electrical
chronoscope, at the same time affording a mechanical arrange-
ment that would seem to meet the objections of those opposed
to the electric chronoscope.
METHOD OF LAYING THE CHRONOMETRIC SCALE
A chronoscope is a device for visualizing intervals of time
by freely initiating or terminating the regular movement of
either a point along a graduated scale or a graduated scale
past a point, each division of the scale being the space tra-
versed in a given unit of time. Many a chronoscope has been
devised with perfect balance and bearings but failed because
its scale was too largely a matter of speculation. The
chronometric scale is the real chronoscope — propelling the scale
or moving an index uniformly along it are comparatively
easy accessories.
The possibility of laying a definite chronometric scale was
one of the factors that prompted the present device. Fig. I
shows the method of deriving the scale and of placing it upon
the chronoscope disc. One of a pair of synchronized differ-
ential tuning forks of 256 v.d. frequency is loaded with a small
aluminum feather (A\) by means of a stiff wax. It is then
sounded with its companion, the number of beats per second
counted and its vibration-rate calculated. Then by means of
wax another aluminum feather (A^ is attached to the second
fork, and small increments are made to the wax until the beats
disappear. The forks are again synchronical and their vi-
bration-rate is that formerly calculated. They are mounted
42 JOHN W. TODD
as shown in Fig. I in such manner that feather AI touches
the smoked drum of a kymograph. and feather A\ rests upon
the smoked edge of the chronoscope disc. Into this arrange-
ment a third member is brought, a Morse key rearranged for
break-circuit contact, and bearing two aluminum feathers at
the end of its shaft beneath the button (P). One of these
feathers is adjusted to lie within the same radius of the disc
as the tuning fork feather, A\, and the other is squared with
the point of the feather upon the kymograph.
After the disc is brought to its initial position and the
circuit to the electro-magnet (EM) is closed, it is seen that a
pressure upon the button (P), which short-circuits the current
to EM, will release the disc. At the same time a mark will
be recorded upon the kymograph and another upon the
smoked edge of the chronoscope disc. The entire procedure
is as follows: Start the forks, and with the hand suddenly
revolve the drum of the kymograph, having disconnected it
from the clock-work mechanism, and with the other hand
press upon the button (P) at least twice in quick succession.
FIG. 4.
Lower Fig. 4 is the .kymograph record of the pressures,
and upper Fig. 4 is a diagrammatic view of the disc record
straightened and brought into relationship with it. The time
value of the distance o-y on the chronoscope scale is readily
counted off from the distance on the kymograph scale, A-B.
In laying the chronoscope scale it is necessary only to read
toward o from the point y to establish the first distinguishable
chronometric value from o. There will always be the space
o-x whose time value in toto is known but whose individual
waves are too close together to be distinguished.
After setting the chronographic records with shellac the
scale of time values is engraved upon the chronoscope disc
A NEW CHRONOSCOPE 43
at grade points located by producing a radius of the disc
through the crest of each tuning-fork wave. The time value
of each wave-length was 4* which save in the case of waves
near x is graded in four equal parts, or to the ia. This
chronographic method of laying the scale is superior to the
method of employing sparks produced by an induction coil
with a tuning-fork interrupter because the sparks deviate
considerably in making the aerial gap, and fail to indicate the
true location of the scale.
In order to test the reliability of the chronoscope a control
FIG. 5.
apparatus (Fig. 5) is used consisting of a shaft with a polished
metal point so poised that its axis if produced is a secant of
the pendular arc. This shaft is carried by a comparatively
facile spring in a rigid base. When the pendulum strikes the
shaft it is pushed back breaking the circuit at x, thus releasing
the chronoscope index and recording upon the scale the
interval between the release of the pendulum and its striking
the shaft. The values given below are the averages of 12
groups of control tests of 10 trials each, or 120 tests. The
mechanical release was used in the tests.
Average Interval Average Variation
519.4°" 0.84°"
520.2 1.08
520.4 0.64
520.1 0.18
520.4 0.64
521.4 0.80
5I9-S i-oo
519.0 i.oo
520.3 0.42
519-8 0.48
519-9 0.72
522.0°" Gross Av. 520.20" 0.42°" Gross Av. 0.68°"
44 JOHN W. TODD
The present device by means of its adjustable discs (Di
and Z)2, Fig. i) affords a tachistoscope that is fairly service-
able.1 The point-exposure time may be read off from the
chronometric scale, and affords a maximum point exposure
of 420'. The stimuli are held by a clip behind the discs and
are seen through the sector of the compound disc. B, in
Fig. 3, is a diagram showing a tachistoscopic attachment for
the stimulus key (Stim. K) making it possible to expose
words, colors, etc., in a rectangular aperture for reaction
experiments in discrimination, cognition, choice and asso-
ciation. The screen is large enough to conceal the operations
of the experimenter from the reagent. By using the rubber
plate under the sound hammer the exposures are made almost
noiselessly.
SUMMARY OF THE SPECIAL FEATURES OF THE CHRONOSCOPE
1. It allows either mechanical or electrical release of the
time scale, involving in each case the same parts, and making
it possible to work where a steady electric current is not
available.
2. The reaction mechanism may be operated either
mechanically or electrically.
3. The device makes it possible to lay a chronometric
scale whose units can be exactly placed to within a short dis-
tance of the zero point, and whose total value is exactly known.
4. By means of an attachment time exposures may be
made that are measured off on the chronometric scale, and
the variety of compound-reaction stimuli can be given.
1 In his "Mental and Physical Tests — Simpler Processes," 1914, pp. 263 ff.,
Whipple figures and describes a disc tachistoscope of his own construction and one
now commonly in use. In his instrument the point-exposure times are calculated
from the relative positions of weights upon the two counterbalancing arms that
route the disc. In the present arrangement the times are shown on the chronometric
scale.
XVIII. PRACTICE IN ASSOCIATING COLOR-NAMES
WITH COLORS1
BY WARNER BROWN
It has long been known that the process of recognizing
and naming a color takes more time than the process of recog-
nizing and naming an isolated printed word, such as the word,
for example, which designates the same color.2 The following
experiments represent an attempt to gain a clearer under-
standing of this phenomenon.
The first hypothesis which presented itself was that
words can be recognized and named more rapidly because
we have had more practice in doing this than in naming
colors.3 Accordingly a practice experiment was contrived
on the basis of Cattell's familiar color-naming test.4 Experi-
1 From the Psychological Laboratory of the University of California.
2 James, W., 'Principles of Psychology,' 1890, Vol. I., p. 559. Cattell, J. McK.
'Ueber die Zeit der Erkennung und Benennung von Schriftzeichen Bildern und Farben,
Philos. Stud., Vol. 2, 1885, pp. 635-650.
3 This explanation of the phenomenon is clearly stated by Cattell in the account
of its discovery which he gives under the title, 'The Time it Takes to See and Name
Objects' (Mind, Vol. II, 1886, p. 65). He says, "The time was found to be about
the same (over \ sec.) for colors as for pictures, and about twice as long as for words
or letters. Other experiments I have made show that we can recognize a single color
or picture in a slightly shorter time than a word or letter, but take longer to name it.
This is because in the case of words and letters the association has taken place so
often that the process has become automatic, whereas in the case of colors and pictures
we must by a voluntary effort choose the name."
The same interpretation is given by J. O. Quantz in his monograph, 'Problems
in the Psychology of Reading,' PSYCHOL. REV. MONOG., No. 5, 1897, p. 10. "The
association is of the same sort in words as in forms or colors, for the connection between
the written symbols and the spoken sound of any given word is just as arbitrary as is
that between a particular geometrical form and its name as uttered. But the asso-
ciation between forms or colors and their names, being less necessary than between
written and printed (spoken?) words has been less frequently formed and the former
has remained a voluntary process while the latter has become automatic through
repetition."
4 Cattell and Farrand, 'Physical and Mental Measurements of Students of
Columbia University,' PSYCHOL. REV., Vol. 3, 1896, p. 642. Wissler, C., 'The Correla-
tion of Mental and Phsyical Tests,' PSYCHOL. REV. MONOG., No. 16, 1901, p. 8.
Hollingworth, H. L., ' The Influence of Caffein on Efficiency,' Arch, of Psychol., No. 22,
1912, p. 16.
45
46 WARNER BROWN
ence had shown that the Columbia test was weak in the fol-
lowing points: Not all the color names were equally familiar;
they were not all equally hard to say (for example red, yellow;
blue, violet); there were strong brightness contrasts between
some of the colors; the chance arrangement of the colors
resulted in some bad sequences; the one-centimeter squares
were too small, making it difficult to 'keep the place' with the
eye. The test was accordingly modified in these respects:
The color squares were increased in size to one inch; the
sequence was so arranged that no color square was placed
next to another of the same color and a color was not per-
mitted to occur less than twice nor more than three times in
any row; only four different colors were used in any one set
and these were all either 'light' (white, pink, brown, gray)
or 'dark' (black, red, blue, green);1 the colors all had one-
syllable names; all of these names were highly familiar.2
It was expected on the hypothesis of Cattell and Quantz
that sufficient practice would make it possible to read off
the color names as rapidly from the colors themselves as
from a printed list. If the difference in speed depends upon
previous practice it should, by further practice, be possible
to reduce the time consumed in reading colors but not possible
to reduce to any considerable extent the time required to read
a list of words. In order to test the truth of this hypothesis
it was necessary to show not only that the speed of color
naming can be increased by practice but also that the speed
of reading words can not be increased so much by an equal
amount of practice. For the practice in reading words, lists
were typewritten with the one hundred color-names arranged
in the same order as the colors themselves. The words in
1 The colors used were the papers supplied by the Milton Bradley Company, of
Springfield, Mass., under the following designations: Black, White, Neutral Gray
No. 2, Engine Colored Paper No. 26 (brown) and No. iB (pink), Red, Green, and Blue.
1 The modified form of the Columbia test recommended by Woodworth and Wells,
'Association Tests/ PSYCHOL. REV. MONOG., No. 57, 1911, p. 49, meets most of the
difficulties mentioned above, but unfortunately it was not published until after the
present experiments were partly completed. It may be noted that in the Woodworth
and Wells test the colors appear on a white background whereas in the form here used
the squares were larger and juxtaposed without background.
COLOR-NAMES 47
each line were separated by a comma and one space; the lines
were separated by a triple space. For each set of colors there
were, of course, four distinct lists of words, corresponding to
the four arrangements of colors which were encountered on
beginning in the four different corners of the color-set. For
every practice trial in associating the colors with their names
there was a practice in reading the words from the corre-
sponding list.
A record-blank, including complete directions, was given
to each worker at each practice sitting; it read as follows:
DIRECTIONS FOR THE EXPERIMENT ON NAMING COLORS
There are two boards of colors. Each board contains 25 squares of each of 4
colors, and there is a different color in each corner of the board. There are 4 type-
written lists of colors for each of the boards, and each list begins with the name of
the color in one corner of the board, and gives the names of the colors in the order of
their appearance on the board.
The purpose of the experiment is to measure the maximum rate of speaking when
reading the lists of words or naming the colors, and to see how much this rate can be
increased by practice.
First day's work. Take the time with a stop-watch for reading aloud, as fast
as you possibly can, the words on the typewritten list beginning with Black. Enter
the time, in seconds and fifths of a second, opposite "List black" in the table below.
Then take the time for calling out the names of the colors, as fast as you possibly can,
from the board, beginning with Black in the upper left-hand corner and reading by rows
from left to right. Enter the time opposite ' Board black ' in the table. Then enter the
time for each of the remaining items in the table, being careful to take them in the
order indicated by the numbers.
1. List black 3. List white
2. Board black 4. Board white
5. List blue 7. List brown
6. Board blue 8. Board brown
9. List green n. List pink
10. Board green 12. Board pink
13. List red . 15. List gray
14. Board red 16. Board gray
Second and succeeding days. Use only one board of colors and the lists which
belong with it. Do not look at the other board or its lists, nor allow any one to read
them in your hearing. Record the times for the right (left)1 hand half of the table in
the order given, and do nothing with the other half of the table.
Twelfth day. Exactly the same as the first day.
Forty-five students took part in the experiment. All
1 If the subject was to practice the 'dark' colors the word right was expunged;
if he was to practice the 'light' colors the word left was expunged.
48
WARNER BROWN
practiced for twelve practice-periods. Most of them worked
twice a week, but a few practiced daily. Twenty-five of the
forty-five were women. Twenty, of whom ten were women,
practiced on the 'dark' colors. Twenty-five, of whom fifteen
were women, practiced on the 'light' colors. As no essential
difference appears between the light and dark colors the data
have been combined for the entire forty-five workers.1
The condensed data are presented in Table I. The table
TABLE I
GAIN BY PRACTICE IN NAMING COLORS AND READING WORDS
Average of 45 Subjects
The time is the average of the 4 trials made each day.
Colors :
Colors:
Colors:
Words:
Words:
Words:
Ratio:
Av. Time
Av. Gain
Av. Gain
Av. Time
Av. Gain
Av. Gain
Time for
Day.
Required
to Name
in Speed
Over
in Speed
Over
Required
to Read
in Speed
Orer
in Speed
Over
Colors
Divided
Them,
Sees.
ist Day,
Sees.
ist Day,
Per Cent.
Them,
Sees.
ist Day,
Sees.
ist Day,
Per Cent.
by Time
for Words
j
S5.8
35-2
•59
2
50.9
4-9
8.8
33-o
2.2
6.3
•54
3
46.4
9.4
16.8
31-6
3-6
10.2
•47
4
45-2
10.6
19.0
30.8
44
12.5
.46
5
43-7
12. 1
21.7
30.2
5-o
I4.2
•44
6
42.8
13.0
23.2
30.4
4.8
13-6
.41
7
42.4
134
24.0
29.9
5-3
I5-I
.42
8
41.4
144
25.8
29-5
57
16.2
.40
9
41.4
144
25.8
29.4
5-8
16.5
.41
10
41.1
147
26.4
29.0
6.2
17.6
.42
ii
40.7
I5-I
27.1
29.4
5.8
16.5
•38
12
41.4
144
25.8
29-3
5-9
16.8
.41
shows the average time required by the 45 subjects for naming
the 100 colors and for reading aloud the 100 words. The time
is the average for the four trials which were made each day.2
1 On the first day of work, when records were made for all of the subjects with both
light and dark sets (t. e., the first practice record with one set and the first check record
with the other set) the times were as follows:
Time required to name 100 dark colors 56.0 sec.; 100 'dark' words 36.0 sec.
Time required to name 100 light colors 55^8 sec.; 'light' words 35.2 sec.
This insignificant advantage of the light sets remains unchanged through the course of
practice. Most persons prefer to work with the light colors on esthetic grounds.
Some subjects complain of getting the tongue twisted around the words, blue and
black in the dark sets because of the identity of their initial sounds.
1 These four trials did not differ greatly from one another. As a rule the first
trial was better than the others except that on the first day, and to some extent on the
COLOR-NAMES
49
The practice gains are shown both in seconds and in per cent.
In both cases the amount of gain is computed on the basis of
the speed on the first day of work. The table further shows
the ratio between the time required for colors and the time
required for words.
In Table la the records are shown for the tests which
TABLE la
TESTS ON UNPRACTICED SETS, FOR WHICH RECORDS WERE MADE ON THE IST AND
I2TH DAYS OF PRACTICE
Column headings as above.
I
12
55-9
474
8.5
15.2
35.8
32.0
3,8
10.6
i.56
1.48
TABLE Ib
SEPARATE STATEMENT FOR MEN AND FOR WOMEN FOR THE IST AND I2TH DAYS OF
THE REGULAR PRACTICE WORK
Figures for Women in Italics
Headings as above.
I
58.9
35.6
1.66
I
53-3
35-i
1.52
12
42.0
16.9
28.7
29.9
5-7
16.0
1.40
12
39-9
13-4
25.1
2Q.O
6.1
1.74
1-38
were made on the first and last days with different sets of
colors and words.
In Table Ib the data of the first and last days are arranged
to display the fact that women excel men in speed in naming
colors, but that men improve more with practice.1
From the data of Table I. and from an inspection of the
curves of Fig. i it can be seen that the hypothesis on which
this experiment was based is probably not true. At the end
second day, there was improvement from trial to trial. The following figures were
obtained by averaging the records for the last ten days of practice:
ist trial 2d trial sd trial 4th trial
Time for 100 colors 42.3 42.3 43.0 42.8
Time for 100 words 29.0 30.4 29.9 30.5
Evidently the practice gains during this period occur in the intervals between sittings,
'overnight,' and not during the course of a sitting.
1 The superiority of women in naming colors has been observed by Woodworth
and Wells, PSYCHOL. REV. MONOG., No. 57, 1911, p. 51, and by Wissler, PSYCHOL.
REV. MONOG., No. 16, 1901, p. 17.
WARNER BROWN
of twelve periods of practice it is evident that only a very
slight further increase of speed in naming colors can be an-
ticipated, no matter how much more practice is taken; yet
the absolute rate in naming colors remains much slower than
the rate of reading the same words from the list and is even
slower than the word rate was before the beginning of practice.
Furthermore the life-long practice which we have had in
reading words has not brought that function to a maximum
speed; on the contrary it shows an amount of practice-im-
provement almost proportional to the improvement shown
in naming the colors. For every second gained in naming
colors at any stage of practice approximately half a second
has been gained in reading words. The ratio between speed
Sees.
60
55
50
45
40
35
30
25
Colors
Words
5 6 7
Days of Practice.
10 ii
12
in color naming and speed in word reading (the last column of
Table I.) shows no indication of approaching unity.1
From these data it seems safe to conclude that the dif-
1 The statements of this paragraph are true not only for the average results"given
the table, but for each individual subject who took part in the experiment] No
statement of the variability or probable error of the measurements has been made
tuse such a statement could have no direct bearing upon the interpretation of the
the present connection. The individual differences in absolute speed were
very large, but they do not in any way affect the results.
COLOR-NAMES 51
ference in speed between color-naming and word-reading
does not depend upon practice.
Further confirmation of this conclusion is found in the
fact that the effects of training in reading words are specific
for the particular words read and do not extend to other
words. It will be recalled that each person was trained upon
either the Might' or the 'dark' set, but that a test was made,
at the first and last sitting, of his speed with the other set
(the one he did not practice). The results of these tests are
indicated in Table la. The speed on the unpracticed sets at
the end of twelve days of practice is better than on the prac-
ticed sets on the second day of practice, but not so good as
on the third day. In other words, three days of direct practice
are better than two days of direct practice plus ten intervening
days of indirect practice. This, too, in a case where the
conditions regarding eye-movement and general adaptation
to work might lead us to anticipate a considerable amount of
transference of practice or ' formal' training. In the present
connection the significant fact is that the amount of transfered
practice is but little greater in the case of reading words than
in the case of naming colors. Apparently we must have
practice in reading specific words before we can attain great
proficiency in reading them. It can not, therefore, be safely
asserted that we read color names faster than we name colors
simply because of the large amount of practice which we have
had in reading words in general.
i
THE SECOND EXPERIMENT
It now seemed clear that the effects of previous practice
do not afford a sufficient explanation of the difference in
speed between color-naming and word-reading. Accordingly
the problem was attacked from another quarter. The intro-
spections of practically all of the students who had taken
part in the first experiment agreed upon one point: It is easier
to speak a printed word than to name a color because when
you want to name a color you have first to think of the name
(the word) and then speak it, whereas the printed word can
52
WARNER BROWN
be uttered without your having to think of anything. The
observations of our foreign-born students were particularly
clear on this point.1
On the basis of these introspections the hypothesis was
formed that the process of color-naming would be facilitated
by suggesting the word at the moment the color was pre-
sented. For actual experiment color-sets were prepared
which had entire words or parts of words printed on the face
of the colors themselves.
Nineteen students finally completed all the stages of this
experiment. They were first given fifteen periods, twice a
week, of practice in reading the lists and naming the colors
from sets upon which nothing was printed, just as in the
previous experiment. The color-set was named over three
times at each sitting and the list of words was read once.
TABLE II
Day
Time Required to Name
100 Colors, Sees.
Time Required to Read
100 Words, Sees.
Ratio: Time for Colors
Divided by Time for Words
I
53.8
35-5
•52
2
48.2
32.7
•47
3
46.2
31-7
•45
4
45-3
3I-I
.46
5
42.9
29.8
44
6
42.2
30.2
•39
7
4i-3
29.4
.41
8
40.7
28.8
.41
9
39-3
28.5
.38
10
39-5
29.0
.36
ii
40.4
28.8
.40
12
38.8
27.6
.41
13
38.2
27.7
.38
H
37-9
27.6
•37
15
37-i
27.4
•35
Only the 'light' set was used. The words of the list, instead
of being printed in a regular list with the ten words of a line
separated by commas, were now typewritten on separate
squares of paper, one inch square, which were mounted on a
board just as the color-squares were mounted, so that the
eye-movements involved were as nearly as possible the same
as for the colors. The data for these fifteen preliminary
1 Three Japanese and one Armenian took part in the experiment, but their records
are not included in the tabulations.
COLOR-NAMES 53
practice sittings are given in Table II. The figures agree
substantially with those presented in Table I.1
After the preliminary practice, which was only intended
to bring the students to such a point that their speed for simple
colors and lists of words would be nearly uniform from day to
day, experiments were begun with sets of colors arranged
just like the others except for words or letters typewritten
upon them. The following transcript of the directions gives
a sufficient outline of the course of this experiment.
DIRECTIONS
Sixteenth day. Read the list of words beginning with brown: then read the
simple color-set beginning with gray. Then read color-set 2 with b on brown; then
set 3 with w on white; then set 6, b on brown, w on white, p on pink, and g on gray.
Seventeenth day. Read the list of words beginning with gray: then the simple
color-set beginning with pink. Then read color-set 7, gr on gray; then set 4, p on
pink; then set 10, br on brown, wh on white, gr on gray, and p on pink.
Eighteenth day. Read the list of words beginning with brown: then the simple
color-set beginning with gray. Then read color-set II, own on brown; then set 12, ink
on pink; then set 15, own on brown, ink on pink, ite on white, and ay on gray.
Nineteenth day. Read the list of words beginning with gray: then the simple
color-set beginning with pink. Then read color-set 16, with full words on all colors.
Then read the simple color-set again beginning with brown. Then read color-set 16,
full words, again.
Twentieth day. Read the list of words beginning with pink: then the simple
color-set beginning with gray. Then read color-set 16, with full words on all colors,
two times. Then read the simple color-set beginning with brown.
Twenty-first day. Read the list of words beginning with brown: then the simple
color-set beginning with gray. Then read color-set 13, with ite on white; then set 14,
with ay on gray; then set 15, own on brown, ite on white, ink on pink, and ay on gray.
Twenty-second day. Read the list of words beginning with gray: then the simple
color-set beginning with pink. Then read color-set 8, with br on brown; then set 9,
with wh on white; then set 10, br on brown, gr on gray, wh on white, and p on pink.
Twenty-third day. Read the list of words beginning with brown: then the simple
color-set beginning with gray. Then read color-set 4, with p on pink; then set 5
with g on gray; then set 6, with b on brown, p on pink, w on white, and g on gray.
The data for the last eight days of this experiment are
presented in Table III. They are combined in the table so
that wherever two records of the same kind were obtained on
1 It may be noted that the rate of improvement is here almost the same as in the
earlier experiment in spite of the fact that the colors were practiced only three times
and the words only once instead of four times as in the earlier experiment. In view
of the fact already mentioned that the first trial of a sitting is usually the best there
is reason for believing that nearly the same results could be obtained in this work by
one trial per day as by four trials per day.
54
WARNER BROWN
TABLE III
TIME REQUIRED TO NAME 100 COLORS, TO READ 100 WORDS, AND TO NAME 100
COLORS WITH THE HELP OF PRINTED CUES
Day
Simple
Colors
Words
Colors on Which the Following Letters Were Printed as Cues
to Help in Naming the Colors
40.8 An initial consonant on one color.
16
38.1
28.1
39.4 An initial consonant on each color.
17
36.6
27-6
38.8 Initial pair of consonants on one color.
36.4 Initial pair of consonants on each color.
38.0 Vowel and final consonant on one color.
18
36.6
27.6
38.4 Vowel and final consonant on each color.
19
36.1
27.6
28.6 Entire word on each color.
20
357
27.2
28.3 Entire word on each color.
37.3 Vowel and final consonant on one color.
21
36.1
27.2
36.4 Vowel and final consonant on each color.
36.9 Initial pair of consonants on one color.
22
3S-o
27.9
'32.3 Initial pair of consonants on each color.
35.2 Initial consonant on one color.
23
35-3
27.9
34.6 Initial consonant on each color.
the same day only their average appears. When the entire
words are printed on the colors it is possible to read the
words without attending to the colors, but even in that case
the average speed is not so great as when the words are read
alone without the colored background, as may be seen in the
records of days nineteen and twenty. After having practiced
with the full words on the colored backgrounds some of the
students found it possible to read the color-names directly
upon seeing the initial letters without considering the back-
ground. This accounts for the fact that the records for the
twenty-second and twenty-third days, with initials, are con-
siderably better than the records for the sixteenth and seven-
teenth days under the same conditions.
From the results of this part of the experiment it may be
concluded that the association process in naming simple
objects like colors is radically different from the association
process in reading printed words. The presence of a visual
symbol of the sound does not greatly, if at all, facilitate the
process of association between color and color-name. Pho-
netic symbols which might suggest the name of the color do
not help us in naming it unless they are so clear that they
enable us to read the name itself directly without going
through the process of naming the color. The one association
COLOR-NAMES 55
process does not reinforce the other. The introspections of all
the subjects confirm the figures in declaring that the letters
printed on the colors do not serve as helpful cues or prompts,
but on the contrary actually interfere with the process of
association.1
CONCLUSION
The conclusions of these experiments seem to be entirely
negative. No facts have been adduced to explain why more
time is required to associate speech movements with a color
than with the corresponding printed word. But the evidence
does throw some light on the problem in so far as it eliminates
very definitely two lines of explanation which have been thought
possible. First, the phenomenon does not spring from a dif-
ference in the amount of practice which the two functions
have had in the past. Second the process of reading words is
not involved in the process of naming colors as a subsidiary
function. The two functions do not overlap, and in all
probability they depend upon distinct physiological processes.
1 A very similar problem has been attacked with the chronoscope by Bourdon.
"Sur le temps necessair pour nommer les nombres," Rev. Philos., Vol. 65, 1908, p. 426.
He finds that the time required to perceive and name a number of points of light
(not exceeding four) is only slightly greater than the time required to read arabic
numerals. Accordingly he infers that the process of perceiving a few points as a
number is as simple as perceiving the symbol of the number. Apparent conflicts
between this observation and the results in the case of naming colors are now under
investigation in this laboratory.
XIX. THE APPARENT RATE OF LIGHT SUCCES-
SION AS COMPARED WITH SOUND SUCCESSION1
BY BERTHA VON DER NIENBURG
It has often been observed that we perceive a duration
marked off by lights as shorter than an identical duration
marked off by sounds, a result readily explained by the
presence of after-images in the case of the light sensations.
Preliminary experiments with series of lights and of sounds
indicated, however, that not infrequently the light rate seemed
slower than the sound rate.2 This study was undertaken to
look into the subject more thoroughly, first from a descrip-
tive view point and later from a causal point of view.
The experiments were conducted during the period from
September, 1910, to May, 1911, in the psychological labora-
tory of the University of California. The subjects were
taken from the class in general psychology; their number
varied for the several parts of the work.
I. In the first group of experiments the light succession
and the sound succession were of equal rapidity. The ap-
paratus in the main consisted of a metronome and a minature
electric light, a telegraph sounder, and the necessary switches
and connections. The current flowing through the metro-
nome, which was placed in a distant room so that its ticking
1 From the Psychological Laboratory of the University of California.
8 Experiments were tried upon a class in psychology at the University of California,
a sound-series and a light series of equal rate (240 a minute) being given to all the
students together, who thereupon reported their independent judgments in writing.
In a first experiment the class was left in ignorance as to the relative rates of the
two series, the questions being in the form: "Are the two series of equal rate? If
not, which is faster?" The judgments were;
That the light-series was faster (L.F.) 24
That the two series were equal (E.) 55
That the light-series was slower (L.S.) 50
A fortnight later the same class was told that the two series were of equal rate,
and the students were asked to tell how the series seemed in this respect— wheth er
they seemed equal, or, if not, which seemed faster. The judgments were,
L.F. 41 E. 10. L.S. 70.
56
LIGHT AND SOUND SUCCESSION 57
might not be heard by the subject, could be sent either into
the telegraph sounder, thus marking off the intervals by sound,
or into the incandescent electric bulb, when the duration was
marked off by light. The entire apparatus was enclosed by
screens, so that no light could reach the eye save indirectly
through a small aperture (^ inch in diameter) which flashed
the light upon a screen.
The subject, who was seated before the screen upon which
the flashes appeared, was told that he would be given a series
of taps to be followed by a series of flashes, and that he was
to compare the rate with which the taps were coming and
the rate with which the flashes appeared. He was allowed a
trial in the beginning to insure a perfect understanding of the
experiment.
Eighteen subjects, ten women and eight men, were ex-
perimented upon. Each subject was allowed to form his
judgments in whatever manner seemed most natural to him.
Three rates were used, namely 61, 154, and 183 impres-
sions per minute, or each interval was approximatly i",
.43", and .32" in length. The series were arranged according
to the following plan. First, 30 taps at .43" following 30
flashes at the same rate were given, and the judgment was
recorded. This was repeated until 10 judgments had been
obtained. The order was then reversed, 30 flashes being
given first, with judgment. Next came 10 taps followed by
10 flashes, and a reversal; and 20 taps following 20 flashes,
and then the order reversed. This procedure was then
followed for the other two rates.
The percentages of respective judgments for the eighteen
people, combining, at first, all rates, lengths, and orders, are
as follows:
Fourteen subjects were used for one hour only, and while
60 judgments from each was the aim, yet because of irregular-
ities in the apparatus and in the subjects themselves the judg-
ments differed in number from 30 to 100. Four other subjects
gave four hours and passed from 180 to 230 judgments. I have
grouped these subjects into three groups: those passing from
30 to 50 judgments, those passing from 60 to 100 judgments
BERTHA VON DER N I EN BURG
'Light faster* 'Light equal to sound' 'Light slower'
37-5% 54-9% 7-6%
NUMBER AND PERCENTAGE OF RESPECTIVE JUDGMENTS PER INDIVIDUAL
Subjects
Total
Number of
Judgment!
Passed
'Light Faster'
Judgment!
'Light and
Sound Equal'
Judgments
'Light Slower'
Judgments
Number
Per
Cent.
Number
Per
Cent.
Number
Per
Cent.
Men
A
200
30
100
230
60
40
40
2IO
9IO
ISI
18
48
46
25
4
19
US
426
I5'5
oo.o
48.0
2O.O
41.7
10.0
47-5
54-8
44-7
46.8
49
9
50
184
28
36
17
79
452
24-5
30.0
50.0
80.0
46.7
90.0
42.|
37-6
50.2
497
3
2
7
J
32
10
II.7
IO
7^6
5-2
3-S
B
E
L ;
N
p
s
T
Sub-totals and average per
cents
Percentage of sum of each
class of judgments, of total
judgments
NUMBER AND PERCENTAGE OF RESPECTIVE JUDGMENTS PER INDIVIDUAL
Subjects
Total
Number of
Judgments
Passed
'Light Faster'
Judgments
'Light and
Sound Equal'
Judgments
'Light Slower'
Judgments
Number
Per
Cent.
Number
Per
Cent.
Number
Per
Cent.
Women
C . ...
£
180
60
£
£
80
50
690
i, 600
14
32
2
25
10
12
H
1
23
I76
602
3S-o
53-3
I.I
41.7
25.0
IS.O
35-0
63.3
46.0
32.3
25-5
38.5
37-6
26
27
178
27
23
67
IS
I
74
27
465
917
65.0
45-o
98.9
45-0
57-5
83.8
37-5
i-7
92.5
S4-o
58.1
67.4
S4-i
57-3
I
8
7
i
ii
21
49
81
1-7
13-3
17.5
i-3
27-S
35-o
9-6
7-1
7-4
S-i
D
F
G
H
/
K
M
O...
R... .
Sub-totals and average per
cents
Percentage of each class of
judgments, of total judg-
ments
Grand totals and average per
cents for all 18 observers . .
Percentage of sum of each
class of judgments, of total
judgments
and those passing from 180 to 230 judgments and giving equal
weight to each individual's results we have the following
table:
LIGHT AND SOUND SUCCESSION 59
Judgments
No. of Obserrers
L. F.
E.
L.S.
30 to 50
60 to 100
180 to 230
7
7
4
*&
36.46%
53-8%
52.08%
61.6%
9-3%
9-28%
1-94%
It will be noticed that there is very little difference in the
results of the first two groups. The difference in the results
in the third groups I judge to be due to the personnel of the
group. I think that from these figures I may conclude that
the number of judgments does not noticeably affect the final
results and that it is therefore not distorting the facts to
throw these individuals together.
With two exceptions each individual varied to a large
extent in his successive estimations. With no person was
the sound declared faster for a majority of the judgments,
while in the case of seven persons it was never considered
faster at all. In the case of twelve of the observers, the
greater number (the 'plurality') of their judgments were of
equality, and with the remaining six observers, the light
rate of succession was deemed faster than the sound rate.
II. To obtain a numerical evaluation of the differences
in the apparent ratings of the succession of these series, the
apparatus was now arranged so that the rate of the flashes
could be altered at will. While the sounder was still operated
through the metronome, the electric bulb was put on another
circuit. The physical light was now continuous; but by
revolving on a kymograph before it a wheel from whose
circumference eighteen acute angled notches were cut, the
effect of flashes was given to the subject who saw the light
through a small aperture in a black screen. A piece of
ground glass, placed directly at the back of the wheel, served
to make the light more distinct. By means of the kymo-
graph, the rate of succession could be varied at will. In
this part of the work the duration of each flash of light was
the same as that of the dark interval which followed it. The
sound rate remained constant, that, is 154 beats per minute.
The light rate was varied, by steps of eight beats per minute,
between the limits of the variable judgments, which proved
to be between no and 166 beats per minute. Twenty clicks
6o
BERTHA VON DER N I EN BURG
followed by twenty flashes were given throughout this group.
The light succession first compared was at a rate of no a
minute, then 126, and so to 142, 158, 118, 134, 150, and 166.
Ten judgments for each series were obtained. Five subjects
designated alphabetically in the following table, were used,
the first four persons having been subjects in the previous
work.
The approximate points at which the succession of lights
were judged to be of equal rapidity with the I54~beat-per-
minute sound-rate were:
Person
A..
C.
D.
E.
134
138
138
126
146
In each of these cases, therefore, the light rate was esti-
mated to be (relatively to the sounds) faster than it really was.
In the former part of the experiment, where equal rates of
succession were judged, the results for such of these subjects
as there acted as observers read for the 154-impressions-a-
second rate:
L. F.i
E.i
L.S.i
A..
oc%
c%
B
2 ?%
07 rOf
c
5%
QC%
P
17-5%
V.5 fO
62.5%
20%
Where the two series to be compared were kept equal, the
judgments do not show the same uniformity as in those
series where the rate of the light succession was varied.
In the entire group of experiments it was evident that
there were variations in the judgments of the same individual
from day to day, and from individual to individual.
III. A. Attention was now given to the various factors
that might cause some of the variations in the judgments
passed upon the equal rate of succession. The influence of
1 These abbreviations stand here and elsewhere for judgments 'Light Faster
than Sound,' 'Light Equal to Sound,' and 'Light Slower than Sound.'
LIGHT AND SOUND SIUCCESSON
61
the rate itself was first observed. Three rates were used,
183, 154, and 61 beats per minute; i. e., each double phase
was approximately 0.3 2", 0.43", and i" long. From 20 to
1 20 judgments were given for each rate of six subjects. The
percentage of the respective judgments for all are given in
the accompanying table:
NUMBER AND PER CENT. OF JUDGMENTS FOR VARYING RATES
Rates: 61
154
183
Sub-
Judgments of
ject
L. F.
E.
L. S.
L. F.
E.
L. S.
L. F.
E.
L.S.
No.
*
No.
%
No.
i
No.
i
No.
%
No.
i
No.
i
No.
*
No.
t
A..
57
71-3
23
28.8
-
—
38
95
2
5
_
—
56
70
24
30
_
—
E ..
7
35
II
£5
2
10
32
53-3
28
46.7
—
—
9
45
II
55
—
—
F...
2O
IOO
-
—
I
2-5
39
97-5
-
—
I
0.8
119
99.2
-
—
1...
—
—
19
55
I
5
9
22.5
3i
77-5
-
—
3
15
17
85
-
—
L...
—
—
40
IOO
—
—
3
5
57
95
—
—
43
3S-8
77
64.2
—
—
T ..
26
37-1
41
^8.6
3
4-3
7
17-5
25
62. <;
8
20
82
82
13
13
5
5
90
23-9
154
72.9
6
3-2
90
32-6
182
64.0
8
3-3
194
41-5
261
57-7
5
.8
These results would indicate that the light appears re-
latively faster as the rate of succession is increased. In-
dividual differences, however, exist.
B. The influence of different lengths of the series was now
tested. Three series, of 30 beats, and 20 beats, and 10 beats,
respectively, were used. The percentages are made up of
from 20 to 1 20 judgments by eight individuals for each series.
INDIVIDUAL TABULATION OF PERCENTAGE OF RESPECTIVE JUDGMENTS FOR SERIES
OF VARYING LENGTHS.
Length of Series
10
20
30
Sub-
ject
Judgments of
L. F.
E.
L.S.
L. F.
E.
L.S.
L. F.
E.
L. S.
No.
*
No.
*
No.
i
No.
i
No.
i
No.
$
No.
*
No.
*
No.
i
A ..
93
77-5
27
22.5
-
—
46
76.7
H
23-3
—
—
12
60
8
40
_
t ...
—
—
40
IOO
-
—
I
i
99
99
—
—
I
2-5
39
97-5
-
—
G . .
5
25
II
50
4
20
II
55
9
45
—
—
13
65
7
35
-
—
L. .
3
75
36
90
I
2-5
6
30
H
70
—
—
3
15
17
85
—
—
H ..
7
8.75
73
91.2
-
24
24
76
76
—
—
15
37-S
2S
62. S
-
—
A ..
6
30
12
60
2
IO
13
65
7
35
—
—
12
60
7
35
I
5
O ..
i
2.5
39
97-5
—
3
30
7
70
—
—
35
87.5
5
12. S
—
T ..
40
44.4
48
53-3
2
2.2
58
58
28
28
H
H
17
85
3
15
-
—
24-5
71.2
4-3
42.4
55-8
1.8
51-6
47.8
.6
62 BERTHA VON DER NIENBURG
There is here a tendency for the apparent rate of light
succession to increase as the series is lengthened. This is
not only manifested in the percentages of the eight subjects
together, but in those for each individual save one. All
subjects agreed that the thirty-beat series was unnecessarily
long. Those persons who formed their judgments immedi-
ately felt that the ten-beat series allowed sufficient time to
make the judgments, while those who formed no estimation
until the final beat preferred the twenty-beat series.
C. Since it was considered by many of the observers to
be much easier to form judgments when the sound succession
was given first, the effect of the order of the light and sound
successions was next considered. The same eighteen subjects
were experimented upon with the following results:
Per Cent. Distribution of Judgments
Sound-Light Light-Sound
L. F. E. L. S. L. F. E. L. S.
38 51 ii 42 52 6
These combined results indicate a slight increase in the
Might faster' judgments when the order is reversed from
sound-light to light-sound. But taking into account the
variation in individual cases, we may conclude that while
the order may increase the ease with which the judgments
can be made, its effect upon the quality of the judgment is
slight.
D. All the observers felt that it was easier to form an idea
of the sound intervals than of the light intervals, and gave as
their reason, that in the latter the flashes disappeared gradu-
ally, instead of sharply as the sound did. By raising and
lowering the light before the notched wheel described on
page 59 the ratio of light to darkness could be changed. The
effect of this change upon the rate at which the light suc-
cession was judged to be of the same rapidity as the sound
succession was studied in experiments upon four persons.
First of all, with the duration of the flash equal to that of
the dark phase, and with the sound-series preceding at a rate
of 154 beats per minute— there was found for each individual
the rate at which the light-series seemed equal to that of the
LIGHT AND SOUND SUCCESSION
sound series. While no point was considered the ' equality'
consistently by any person, yet one rate was judged to be so
more often than others. This rate I shall designate as the
* point of apparent equality.' The ratio of the light to the
total cycle of light-phase plus dark-phase was then increased
to 75 per cent., and ten judgments made; and then decreased
to 25 per cent., and ten judgments recorded; and then back
to 50 per cent. The order of these shifts from one ratio to
another was inconstant, so that sometimes we would go from
50 per cent, to 25 per cent, and then to 75 per cent., or again
from 25 per cent, through 50 per cent, to 75 per cent., and so
on. For the various persons the results were as follows :
Sub-
ject
Apparent
Equality Point
of Light Rate
When Sound
Rate Equals 1541
Ratio of Light to Light-plus-Dark
25 Per Cent.
50 Per Cent.
75 Per Cent.
L. F.
E.
L. S.
L. F.
E.
L. S.
L. F.
E.
L. S.
A ..
B ..
C..
£>..
128
148
134
ISO
76
24
76
18
24
24
24
6(i)
72(d)
72(d)
-(i)
8.0
16
2
16
78
78
76
78
3
22
6
20
60
22
58
20
38
22(d)2
ioo(d)
20(i)
4o(d)
A shows a slight tendency to increase his Might faster' and
'light slower' judgments at the expense of the equality
judgments in the 75 per cent, ratio, while in the 25 per cent,
ratio there is a marked increase in the 'light faster' judg-
ments. B shows a great increase of judgments 'light
slower' at the 75 per cent, ratio, and nearly as great a one for
the 25 per cent. C shows a like tendency for the 25 per cent,
ratio, but exhibits just the opposite tendency for the 75 per
cent, ratio. D shows an increase of 'light slower' judgments
for 75 per cent, and of 'light faster' judgments at 25 per cent.
In three cases, therefore, and contrary to expectation (since
the interval between flashes is now decreased) the 75 per cent,
ratio of light to darkness decreases the apparent rapidity of
the light, and the 25 per cent, ratio increases it.
E. After a month's time, the same observers together with
a new one, E, were used as subjects to see whether the ten-
1 Here and in the following table, *d' indicates decrease in the apparent rate, 'i'
increase and 'N' no change.
* Determined afresh for this portion of the experiment.
64
BERTHA VON DER N I EN BURG
dencies displayed above would exist when the gradation
method was used. The work was conducted in the manner
described above;1 that is, with the sound rate at 154, the light
rate was varied by steps of eight beats per minute from no
beats to 166 beats. This was done for each of the three ratios
of light to dark, ten estimations being recorded for each rate
compared. The approximate points where the light rate was
judged equal to the sound rate were:
RATE OF LIGHTS PER MINUTE WHICH SEEMED EQUAL TO 154 SOUNDS PER MINUTE
Subjects
Ratio of Light to L. + D.
25 Per Cent.
Ratio of Light to L. -f D.
50 Per Cent.
Ratio of Light to L. + D.
75 Per Cent.
A..
134 (N) (i)2
134
132 (I?) (d?)
B
130 (I) (d)
138
118 (I) (d)
c
136 (I?) (d)
138
126 (I) (i)
D
130 (D) (i)
s
126
137 (D) (d)
E
142 (D
146
136 (D
^'s point where the light rate seems to equal the 154
sound rate is somewhat higher than it was in the earlier group
of experiments; he no longer exhibits the same tendency as
before, that is, to regard the light rate as faster for the 25 per
cent, ratio than for the other ratios. 5's equality point is the
same as before, but now the light succession appears to come
much faster for both the 75 per cent, and 25 per cent, ratios,
which is just the opposite of the earlier results. C's equality
point now is somewhat lighter. His judgments are consistent
for the 75 per cent, ratio with those of a month earlier, but not
for the 25 per cent, ratio. Z)'s point of equality has fallen
fourteen beats per minute, and instead of an increased ap-
parent light rate for 25 per cent, as at first, it is decreased.
To the new subject, the apparent light rate is increased by the
greater ratio of light, and to a less degree is increased by the
smaller ratio.
In most cases, then, the 75 per cent, ratio has apparently
quickened the light rate, and the 25 per cent, ratio has also
quickened it. In the previous group of experiments, the
apparent rate was increased or decreased for the 25 per cent.
1 See page 60.
JThe lower-case letters indicate, 'increase' and 'decrease' respectively, in the
earlier groups (see p. (63)).
LIGHT AND SOUND SUCCESSION 65
and the 75 per cent, ratio about indifferently. The conclusion
seems warranted that the judgment of the rate is not based
greatly upon the interval between impressions. For were the
apparently greater rapidity of the light-succession due to its
blur and after-image filling in the interval between im-
pressions, and so making them appear to come closer together,
then the 75 per cent, ratio of light to darkness should greatly
increase the rate of the light succession, and the 25 per cent,
ratio should decrease its apparent rate in somewhat the. same
degree. This, however, did not happen in the above experi-
ments.
F. Throughout the previous work, a record had been
kept of the natural and spontaneous manner in which the
subjects formed their judgments. Four different ways pre-
dominated. Some said that they got an idea of the suc-
cession of the beats in the first series given, and noticed the
likeness or difference in the rate of the second series to it.
This mode will be designated by the term " general impres-
sion." Others seem to have judged by various synchronous
muscular rhythms, of which I shall count as the second
group those in whom this rhythm was less voluntary and
conventional, occurring, e. g., in the forehead, or in the back
of the head. In a third group are placed those who counted;
and in a fourth group, those who tapped.
Dividing twelve of the people used in the first part of the
work according to their mode of making a judgment, three
fall in Group I., three in Group II., four in group III., and two
in Group IV. The percentage of respective judgments for
the varying groups is as follows, when light rate and sound
rate were physically equal (154 per sec.).
Group
L. F.
E.
L. S.
I. .
73 6
2C 6
8
II
20 o
zyu
70 8
92
Ill
21 O
R T
IV
zi.y
38.75
51-25
IO.O
These figures indicate that those who simply get an
'impression' give nearly three fourths of their judgments as
'light faster,' while those judging by muscular rhythm and
66 BERTHA VON DER N I EN BURG
those who counted give as such less than a quarter of their
judgments. In the latter two groups the number of ' equality '
judgments is increased to more than one half the entire
number, while the Might slower7 judgments are greater than
in the first and the fourth group.
G. Ten subjects who had never been experimented upon
previously were taken to make a further study of the direct
effect upon the judgments themselves of deliberately follow-
ing these different ways of forming their judgments. They
were first given the sound-light series with the eight light-
rates and the 154-beat-per-minute sound-rate. One judgment
was passed upon each series. Then ten judgments were
made in the series in which the sound and light successions
were both 154 per minute; the first procedure of one judgment
on each series with a varying light rate was again gone
through. Nothing was said as to ways or means of judging,
so that each subject employed the method most natural to
him. The above process was then repeated twice with other
definite methods of forming the judgments now imposed.
If the original method used had been by 'general impression,'
the person was asked to count for one series, and to beat for
the next. Where the 'general impression' method had to be
imposed, to be certain that there would be no muscular
rhythm, the subjects were asked to say and repeat, 'There
is a black cat,' aloud as fast as possible. Great care was taken
that this was in no way made to fit in with the rhythm of either
series. This procedure after it had been tried one or two
times by the subject in no way hindered the person from giving
attention to the series and did stop any involuntary vocal
synchronisms, so that the person would be judging by his
impression alone.
When left to do as they would, six of the subjects beat,
two counted, and two judged by their impression of succession.
Throughout this part of the work it was noticed that while
each individual may attempt to judge in the same manner,
many variations still remain. For instance, in beating, a
person may move his whole arm, his hand entire, or use his
finger only slightly. He may synchronize the sound and
LIGHT AND SOUND SUCCESSION 67
keep beating that rate throughout the light series, forming
his judgments of the latter by the way its flashes fall in with
his beating, or he may change his tapping as the flashes change,
or he may cease to tap entirely for the light. He may tap a
decided rhythm, he may synchronize well for the sound and
not for the light, or he may not be able to do so for either.
In cases where the tapping changed with the oncoming of
the flashes, it was noticed that in some cases this change in
the tapping would be reflected in the answer, while in other
cases the difference in tapping was no indication of what the
judgment would be. Similar differences were noticed in the
counting. If a person tapped with marked irregularity, his
counting was of the same sort. With all this variety in these
two aids to the formation of time judgments it could hardly be
expected that harmonious results for different individuals
who are beating and then counting could be obtained. The
average percentages of judgments by the three modes em-
ployed are, however, as follows:
LIGHT SERIES AND SOUND SERIES AT EQUAL RATE (154 PER MINUTE)
General Impression Counting Tapping
L. F. E. L. S. L. F. E. L. S. L. F. E. L. S.
52.5% 32% iS-5% 59% 29% 12% 55.5% 28% 16.5%
The later tables that give the judgments in more detail
show the great difference in the different individuals. The
number of 'light slower' judgments seems marked when we
remember that in the first group of persons examined 7.6 was
the average per cent, of such judgments given. The apparent
equality points for these observers were, as found by the
gradation method:
General Impression Counting Tapping
I4I.8 142 146.6
As elsewhere, the judgments by this method do not coin-
cide with those made on the equal rates. In the six cases
where the natural method was to tap, the average percentages
indicate a falling off of 'light faster' judgments for the im-
posed methods with an approximate increase of 50 per cent,
in the 'light slower' judgments when judging by the General
Impression method. Results:
68
BERTHA FON DER NIENBURG
EFFECT OF DEPARTING FROM NATURAL METHOD
Subjects
Natural Method
Imposed Methods
Tapping
General Impression Counting
Judgments of
L. F.
E.
L. S.
L. F.
E.
L. S.
L. F.
E.
L.S.
I
2
3
4
100%
60
20
25
90
80
l\
60
IO
20
5
15
15
80%
50
1°
60
80
10
40
60
40
40
20
10
10
40
10
100%
60
20
50
60
70
40
80
50
40
2O
IO
Average Per Cent.
62.5
31.67
5.83
53-33
35
11.67
60
38.33
1.67
Counting
General Impression
Tapping
8
80%
30
10
30
10
40
30
80
30%
20
40
50%
50
20
30
30
20
Average Per Cent.
55-0
20.0
25.0
55-o
25-0
20.0
50.0
25.0
25.0%
General Impression
Tapping
Counting
9
10
50%
65
40
10
35
60%
20
40
80
80%
40
20
60
57-5
20.0
22.5
40.0
20.0
40.0
60.0
IO.O
30
Those who naturally counted show individually a consider-
able change in result under the other methods, though when
these results are combined there is but a slight tendency to
quicken the light during the general impression and to quicken
the sound while tapping.
When counting and tapping are the imposed methods the
'light slower' judgments are markedly increased with one of
the subjects.
In only four cases does the natural method tend to give a
greater number of equality judgments.
Two individual records in this group deserve special
attention. Instead of judging of the succession of beats as
did the other subjects, one of these judged of the succession
of the intervals, so that in tapping and counting, instead of
matching his beat to that of the metronome, he counted and
tapped after its beat. His results are as follows :
Judgments for 20 S-L, with both sound and light at the
equal rate of 154 per minute,
General Impression
65%L.F. 35%L.S.
Counting
40%L.F. 60% L.S.
Tapping
20% L.F. 80% L.S.
LIGHT AND SOUND SUCCESSION 69
Equality point when sound was at 154, and the light rate
varied (gradation method) :
General Impression Counting Tapping
154 138 154
A contradiction in the results obtained by the two methods
thus is evident. While this subject's original method was
that of the general impression, he said after he had tried the
others, that tapping was by far the surest.
The other person had a noticeably high ' equality point';
in fact he was the only person of the twenty-nine subjects
experimented upon who ever gave as an apparently identical
rate, a higher rate of light succession than of sound succession.
His results read:
General Impression , Counting Tapping
Equality point by gradation method (16
judgments each) 160 150 154
F. E. S. F. E. S. F. E. S.
Distribution of judgments per cent, when light-
rate = sound-rate (154). (10 judgments
each) — 60 40 20 80 — 20 65 15
From these various modifications of the experiment we
may perhaps be justified in drawing the following conclusions.
1. The commonly accepted statement that of equal times
marked off by light and sounds the light-limited durations
seem shorter than the sound-limited is by no means univer-
sally true, when applied to the apparent rate of succession of
series of light and of sound impressions.
The experience that the light succession is less rapid than
the sound succession comes not infrequently, and with some
observers comes indeed more frequently than does the op-
posite experience, that the light succession is more rapid.
2. With persons who are practiced the impression that of
two equal rates the light rate is the slower does not appear to
be influenced directly by the amount of such practice. But
when there is no practice at all, as in the class experiments, a
greater number of persons take the light series to be slower
than take it to be faster.
3. The apparent difference of rate between light succession
70 BERTHA VON DER N I EN BURG
and sound succession does not seem to be directly connected
with the persistence of the light effect upon the retina, which
makes the blank lapse of time between the impressions less
in the case of light than in that of sound. For in the first
place, this could not explain those not infrequent cases where
the light rate seems slower; and, secondly, the effect of arti-
ficially varying this blank lapse without changing the rate
itself affects in no simple and direct way the apparent rate
of the impressions. The change of the blank interval to an
interval less than has become familiar seems more frequently
to have an effect similar to a change in the opposite direction
(i. e., an increase of the blank interval); although different
persons respond differently to this experiment, and differently
at different times.
4. The higher the rate of the two kinds of succession here
compared, the more pronounced becomes, with most obser-
vers, the illusion that the light series is the more rapid.
5. The longer the series of impressions to be compared, the
more pronounced is the illusion that the light series is the
more rapid.
6. The order in which the two series is given affects the
result: the impression of greater rapidity in the light series
comes, with most observers, more frequently when the light
series precedes the sound series.
7. The method of forming the judgment of the relative
speed of these successions differs greatly with different ob-
servers. Those who naturally incline to assist their judg-
ments by counting or by noticing some hidden organic rhythm,
in general have less frequently the sense of greater speed in
the light series than have those who depend upon their
'general impression' (whatever that may mean) or upon an
overt tapping by hand or finger. Yet the illusion of greater
speed in the light does not seem to depend upon the presence
or absence of mental aid from any noticeable organic rhythm,
whether voluntary or involuntary.
XX. A MEMORY TEST WITH SCHOOL CHILDREN1
BY ARTHUR H. CHAMBERLAIN
THE PROBLEM STATED
With a view to determining the power of recall in school
children, a series of tests were made. It was sought to ascer-
tain the effect upon the power to recall when:
1. A number of objects are displayed (a) singly; (b) three
together.
2. The objects chosen interest, we might suppose, par-
ticularly (a) the boy's mind; (b) the girl's mind.
3. The subjects tested are of different school grades.
4. The subjects tested are of different sex.
OBJECTS CHOSEN
Fifteen objects were selected as follows: pocket knife, roll
of string, marble, watch, key, flat-iron, threaded needle,
thimble, scissors, doll, pencil, notebook, two-cent stamp,
five-cent nickel, and match.
These objects fall into three groups: First, those that
interest particularly the boys and are handled by them in
their daily routine. These are the first five objects listed.
Second, those that might be expected to interest the girls.
These are the second five objects named. Third, those that
are of equal interest to both sexes. These consitute the final
five objects. All of the objects are found constantly in the
child's environment. Choice was made of objects that were
not too greatly different in size.
THE SUBJECTS
The subjects were chosen from the third, fifth and eighth
grades, sixty from each grade, one hundred eighty pupils,
all told. They were equally divided, thirty boys and thirty
girls from each grade. They represented several different
1 From the Psychological Laboratory of the University of California.
71
72 ARTHUR H. CHAMBERLAIN
schools, in various sections of the city where the tests were
made.
APPARATUS USED
The apparatus for the tests was simple: A circular disc
of wood, one-half inch thick and eighteen inches in diameter
served as a stand upon which, near its outer edge, the objects
were arranged at equal distances apart. A second disc of like
thickness and twenty-four inches in diameter, had an opening
whose size could be varied as desired. The larger disc was
placed four inches above the smaller, which rested flat upon a
desk or table. The upper disc was held in its horizontal
position and made to revolve over the lower disc as a wheel
revolves on an axle. This was accomplished by means of a
rectangular block having either end cut to a step-cylinder and
housed or shouldered into the inner faces of the two discs
at their centers.
ARRANGEMENT OF OBJECTS
Those five objects supposed to be more familiar to the girls
were placed in sequence — the threaded needle, flat-iron,
thimble, doll and scissors. In the same way those objects
pertaining chiefly to the boys were arranged in sequence —
marble, knife, watch, key and roll of string.
MODES OF DISPLAY
During the tests there were two modes of display. First,
the objects were shown in such a way that, at all times, three
were visible, while as each new object entered this group, one
of the older members of it dropped out. Second, they were
displayed singly. Any one subject was tested with only one
mode of display.
METHOD OF PROCEDURE
Each subject was allowed one minute for observation of
the fifteen objects, whatever was the mode of display. In
the first test, the subject was placed directly in front of the
cut-out sector and, as the upper disc revolved, the objects
came into view, three at a time. The subject moved with
MEMORY TEST 73
the disc, thus -keeping directly in front of the opening. In
the second test, the same order of arrangement was main-
tained, but the opening in the disc revealed only one object
at a time. By allowing three seconds for each exposure, with
one second interval, a total of one minute was given as in
the other test.
In every instance note was made of the object at which
the observation was begun. The immediate recall was
tested at the close of the experiment by having the pupil
name all the objects he could remember. A list of these
objects was recorded by the operator in the order in which the
pupil named them. The subject was also given a sheet of
paper having a circle described upon it and corresponding
to the disc. Upon this circle he was asked to locate the objects
in the order in which they were placed on the disc. Any
additions to the list of objects originally recalled were placed
to the observer's credit. Record was then made of the total
number of objects recalled by each pupil, out of a possible
fifteen; of the order of recall, that is, the sequence in which
the objects were named; and of the order preserved in placing
the objects on the circle. An object is said to be in a 'correct*
position when it is located upon the circle in a position exactly
corresponding to its original position on the disc. Two objects
were said to be in a 'relatively' correct position when they
simply changed places in location, — were transposed. Or, if
two objects in adjacent positions on the disc were given a
corresponding location upon the circle, but not arranged
properly as regards other objects in the group, the placing
was said to be relatively correct. As only objects not in-
cluded in the list of those correctly placed are included in the
' relatively correct ' column, the average for the former is usu-
ally greatly in excess of the latter. Whenever the average for
correct placing is relatively high, the average for relative
placing is considerably lessened. There are twenty judgments
when the objects are displayed singly and twenty when they
are displayed three together.
74
ARTHUR H. CHAMBERLAIN
THE RESULTS
Table I shows the results of the test in the three grades
with the two methods of exposure.
These results are then analyzed in several tables that
bring out: (i) The effect of the different methods of display
without regard to grade, sex or arrangement of objects; (2)
The effect of the particular grade of pupil upon total recall,
correct and relative placing and the like; (3) The effect of
sex upon total recall, the recall of boys' and girls' groups,1 etc.
TABLE I
Sex
Av. No.
Recalled
Av. No.
Correctly
Placed
Av. No.
Relatively
Placed
Av. No.
Recalled,
Boys'
Group
Av. No.
Recalled,
Girls'
Group
M.V.
of Total
Recalled
Grades:
One at a time ....
Boys
74
5-7
.7
2.9
2.7
•17
Girls
8.1
54
•9
2.6
3
•5
Both
7-75
5-55
.8
2-75
2.8S
.61
Three at a time. .. .
Boys
9.1
6.4
•9
2.9
3-7
.92
Girls
8
5-5
•7
34
3
4
Both
8.55
5-95
.8
3-iS
3-35
.66
Grade 5 :
One at a time
Boys
10.5
3
2-5
3-6
3-3
•7
Girls
9.2
2-3
34
2.8
3-3
.04
Both
9.85
2.65
2-95
3-2
3-3
•37
Three at a time. . . .
Boys
9-5
3-7
2.9
2.9
3-2
4
Girls
9-5
2.4
3-7
3-3
3
.1
Both
9-5
3-05
3-3
3-i
3-i
•25
Grade 8:
One at a time
Boys
Q 2
Si
I Q
2 7
•5 A
.04.
Girls
96
' i
7-i
i.y
4
**/
3-8
J'T
3-2
•WT
.24
Both
94
6.4
I-I5
3-25
3-3
.14
Three at a time. . . .
Boys
10
6.4
2-3
3-i
34
4
Girls
IO.I
6.6
2.1
3-i
3-7
•5
Both
10.05
6-5
2.2
3-i
3-55
.46
Table II. shows the collective results with the two methods
of display.
TABLE II
Method of Display
Av. No.
Recalled
Av. No.
Correctly
Placed
Av. No.
Relatively
Placed
Av. No.
Recalled,
Boys' Group
Av. No.
Recalled,
Girls' Group
M.V.
of Total
Recall
Singly
9
9-37
4.87
5.166
1.633
2.1
3-07
3.12
3-IS
3-33
**i
1.46
Three at a time ....
1 The letters G. and B. will be used throughout as referring to the objects spoken
of as the girls' and boys' group respectively.
MEMORY TEST
75
The display of the objects three at a time gives an ad-
vantage in the average number of objects recalled, as well as
in the average number correctly placed. When shown three
together the relative placing also is better than when the
objects are displayed singly. The tendency to recall the G.
or B. group receives very slight advantage in any one method
of display over another. The M.V. of total recall varies only
slightly in the two methods.
Table III. shows comparative results in grades three, five
and eight.
TABLE III
Grade
Av. No.
Recalled
Av. No.
Correctly
Placed
Av. No.
Relatively
Placed
Av. No.
Recalled,
Boys' Group
AT. No.
Recalled,
Girls' Group
M.V.
of Total
Recall
3
8.15
9.67
9.72
575
2.8S
6-45
.8
3-12
1.67
2-95
3-15
3-17
3-i
3-2
34
1.63
I-3I
1.30
The pupils of the fifth grade show a marked superiority
over those of the third grade in the average number recalled.
The eighth grade is not greatly in advance of the fifth in this
particular. In correct placing, the third grade is almost
abreast of the eighth, while the fifth grade drops back to
one half the showing made by the third. In relative placing,
the fifth grade students far excel those of the third and are
much superior to the eighth. The fifth and eighth grades
show about equal power in recall of the B group, while the
third grade makes nearly as good a showing. In every grade
the average for recall of the G group is slightly better than that
of the B group.
Table IV. shows the effect of sex in all grades.
TABLE IV
Sex
Av. No.
Recalled
Av. No.
Correctly
Placed
Av. No.
Relatively
Placed
Av. No.
Recalled,
Boys' Group
Av. No.
Recalled,
Girls' Group
M.V. of
Total Recall
Boys .
Girls .
9.28
9.08
5-15
4.88
1.86
1.86
3-02
3.16
3-28
3-2
1-43
1.30
Table V shows the relation of boys to girls in the different
grades.
76
ARTHUR H. CHAMBERLAIN
TABLE V
Grade
Sex
Av. No.
Recalled
Av. No.
Correctly
Placed
Av. No.
Relatively
Placed
Av. No.
Recalled,
Boys' Groupjj
Av. No.
Recalled,
Girls' Group
M.V. of
Total Recall
3
5
8
Boys
Girls
Boys
Girls
Boys
Girls
8.25
8.05
10
9-35
9.6
9-85
6.O5
545
3-35
2-35
6.05
6.85
.8
.8
2.7
3-55
2.1
1.25
2.9
3
3-25
3-05
2.9
345
3-2
3
3-25
3-iS
34
345
1-55
145
i-SS
1.07
1.22
I.38
In the third and fifth grades, the boys have a slight ad-
vantage over the girls in the total number recalled and in the
number correctly placed. The objects composing the G
group are recalled somewhat better by each sex. The ratio
between the recall by girls and boys of the B group and the
G group is only slightly different from the ratio between girls
and boys in the total recall.
CONCLUSIONS FROM THE STUDY
The results of the various experiments would seem to
justify the following conclusions:
1. Recall is stronger when the objects are seen three at a
time than when shown singly.
2. The average for total recall shows a considerable
increase from the third to the fifth, with an almost negligible
increase from the fifth to the eighth grades. This difference
is emphasized when we consider that the age-difference be-
tween the fifth and eighth grade was approximately twice as
great as that between the third and fifth. In other words,
ability to memorize or to recall does not increase regularly with
advance in age or experience.
3. The total average of recall for all grades and with all
methods of exposure of objects shows the girls not to be super-
ior to the boys. This is not in accord with the usual outcome
of experiments in memory. No clear difference is discernible
between the boys and the girls in the attraction exerted by
the so-called boys' group and girls' group of objects; for both
sexes the girls' group was slightly more attractive.
XXL PRACTICE IN ASSOCIATING NUMBER-
NAMES WITH NUMBER-SYMBOLS1
BY WARNER BROWN
In a recent study2 the writer employed the difference
between the time required to perceive and name a series of
colors and the time required to read the same color-names
when they are printed out in type as a typical instance of the
general rule that it takes longer to perceive and name a simple
object than to perceive and name a word. At that time
reference was made to an experiment by Bourdon in which the
time for calling out the number of points of light in a small
group was said to be no greater than the time for naming the
corresponding arabic numeral.3 On the basis of this experi-
ment Bourdon argues that the process of perceiving a small
number of points as a number is no more complicated than
the process of perceiving the symbol of the number. If this
is correct it means that it is possible in this case to associate
the name of an object with the object itself as quickly as
with the symbol of the name and this would make it seem
probable that the association processes in color-naming involve
difficulties peculiar to themselves and are not typical of the
general situation in which simple objects are perceived and
named.
What follows is an attempt to discover whether number-
naming is really a process which is free from the time-con-
suming difficulties of color-naming. A practice experiment
was devised in which number-naming was subjected to the
same analysis that was applied in the previous case to color-
naming. The material for the experiment was all prepared
with the typewriter. It consisted of four sheets, each con-
1 Studies from the Psychological Laboratory of the University of California.
2 This REVIEW, p. 45.
3 Rev. philos., Vol. 65, 1908, p. 426.
77
78 WARNER BROWN
taining ten lines of ten items. The first sheet contained the
type-written words one, two, three, and four. There were 25
words of each sort, arranged in irregular order with not more
than 3 nor less than 2 of a kind to a line. Four different sets
of sequences of words were used to prevent memorization of
any particular sequence. The lines were separated by triple
space and the words were separated from each other by the
space of three letters. The second sheet contained the arabic
numerals corresponding to the words of the first sheet so
arranged that each symbol occupied a position, which was
relatively the same as the center of the corresponding word
in the sheet of words. In this way the eye-movement factor
was kept as nearly constant as possible. The third sheet
(known as 'dots') was made up of dots arranged to represent
the numbers as follows: one was represented by a period',
two by a colon', three by a colon with a period after it; four
by two colons. The fourth sheet represented the numbers by
an appropriate number of oblique strokes ('scores') made
with the key used in printing fractions on the typewriter.
The time was measured which was required to read aloud
the one hundred items on each of these sheets. The sheets
were read one after the other in the order given above, and
then all four were read again, so that the time stated for
each sheet on each day of practice is the average of two
records, the first and fifth, second and sixth, etc.1
The accompanying table gives the results of eleven days'
practice with this material on the part of twenty-four stu-
dents. None of these students knew the real purpose of the
experiment. They were all encouraged in the supposition,
which came naturally to all of them, that they would be able
with practice to read the "dots" or "scores" as fast as the
words or symbols.
1 The experiment with colors referred to above, made it clear that there is no
considerable difference in time between successive trials on the same day. The
average time for naming 100 colors was found to be 42.3, 42.3, 43.0, and 42.8 seconds
for 4 successive trials; the time for 4 successive trials of reading 100 words was 29.0,
30.4, 29.9, and 30.5 seconds. There can be no serious objection, therefore, to com-
paring the rate of reading one sheet with the rate of reading another sheet when their
sequence is the same each time. It is understood, of course, that the order of the
items on the sheet is different on successive sheets.
NAMES AND NUMBER-SYMBOLS
79
TABLE FOR ELEVEN DAYS OF PRACTICE ON THE PART OF 24 PERSONS, SHOWING
THE AVERAGE TIME REQUIRED TO NAME IOO ITEMS PRESENTED AS WORDS, AS ARABIC
NUMERALS, AS GROUPS OF DOTS, OR AS GROUPS OF SCORES; AND SHOWING FURTHER
THE RATIO OBTAINED BY DIVIDING THE TIME FOR THE ARABIC NUMERALS INTO THE
TIME FOR EACH OF THE OTHER PERFORMANCES.
Ratio of Arabic to :
Day
Word
Arabic
Dot
Score
Word
Dot
Score
I
30.1
27.8
39-6
36.0
1. 08
.42
1.29
2
28.9
26.5
35-9
34-2
1.09
•35
1.29
3
27-6
25.8
344
33.2
1.07
•33
.29
4
27.1
25.0
34-7
32.5
1. 08
•39
•30
26.6
2S.O
34-o
324
.06
•36
.29
6
26.S
24.7
33-0
31-7
.07
•34
.28
7
26.3
24-3
32-3
30.5
.08
•33
•25
8
25-7
24.2
31.8
30-9
.06
•32
•27
9
25-5
23.8
3i-9
31.0
.07
•34
•30
10
25.1
23.6
31-2
304
.07
•32
.29
ii
25.0
234
30.9
29.9
.07
•32
.28
The result shows that the time required to perceive and
name the number of a small group of marks is longer than the
time required to perceive and name the corresponding word
or the arabic symbol of the number. In this respect the
results agree perfectly with the experiment in naming colors,
and support the general dictum that the naming of objects is
slower than the naming of words.1
The same peculiar inhibitions appear in the reading of
the "dots" or " scores" which are encountered in color-
naming. Some persons are more troubled by the "dots"
and others find more difficulty in the case of the "scores"
but no one notes any considerable disturbance of this kind in
the case of the words or arabic symbols.
This experiment agrees with color-naming in the essential
point that the ratio between the time required to name an
object and the time required to name its symbol resists the
1 As a matter of fact Bourdon's statement that the dots are named as fast as the
symbols does not seem to be fully supported even by his own figures. His interpreta-
tion of his data raised a doubt which the results of the present experiment tend to
clear away. There is no striking conflict between the original data of the two experi-
ments. The method of serial reactions which has been used in the present case un-
doubtedly tends to magnify the loss of time in the association process and might
reasonably be expected to show a greater difference in time between the two processes
than would be shown by Bourdon's method of single reactions. But it is not probable
that the serial reactions would show a difference unless the single reactions also gave
some difference and as a matter of fact Bourdon's reactions do give some difference.
So WARNER BROWN
action of practice. This seems to argue in both cases that
previous practice is not the basis of the relative rapidity of
the latter process. So, too, the fact that the process of naming
the symbol is itself capable of a very material improvement
through practice precludes our speaking of it as automatic in
comparison with another process (the slower one of naming
the object) which improves no faster. It does not appear
probable that differences in previous practice have much to do
with the relative speed of the two processes.
The present experiment confirms the inference drawn from
one part of the color experiment that phonetic symbols such
as letters do not seem to be responsible for the advantage in
speed of one association process over the other. It might be
thought that the sight of the different letters would guide the
complex movements of the vocal organs in uttering the word,
and so facilitate the reaction to a word, but it appears that
the words with their phonetic symbols can not be read quite
as fast as the arbitrary arabic symbols which contain no
phonetic elements. This is true for nearly every individual
person and in spite of the fact that the spacing of the words
and figures on the page gave nearly normal conditions of eye-
movement for the words and rather unusual conditions for the
arabic symbols.
In conclusion it may be said that the causes of the delay
of the association processes in naming a simple object remain
as obscure as ever. But on the negative side it seems clear
that the greater speed with which words are named does not
depend upon an advantage in practice and does not depend
upon the suggestiveness of the letters in the words.
XXII. INCIDENTAL MEMORY IN A GROUP OF
PERSONS1
BY WARNER BROWN
The study which is reported below leads to the conclusion
that the factors which make it easy or difficult for an individ-
ual to recall certain of the incidental observations of his past
experience also tend to affect in the same way the collective
memories of a large group of persons.
The material for the investigation was obtained by having
the members of a large college class write down, in a limited
time, the names of all of the advertisements which they could
remember having seen recently in the street-cars. They also,
after making out the list, wrote down answers to certain
questions about the advertisements, but that has nothing to
do with the present report. The experiment was performed
twice. The first time 175 persons wrote lists. The lists
contained in all 896 items and these items were found to
include mentions of 215 different advertisements. Thus the
average person recalled 5.1 advertisements, and the average
advertisement was mentioned 4.2 times. Table I is arranged
TABLE I
The Number of
Items in the List
The Number of Per-
sons Giving a List of
this Length
The Total Number of
Items in the Lists of this
Length
The Relative Frequency
of Occurrence of
Such Items
0
21
I
8
8
23.1
2
17
34
24.8
3
. 15
45
16.2
4
22
88
22.6
16
80
20.8
6
20
120
18.0
7
13
91
21.9
8
12
96
19.7
9
9
81
IS-7
10
7
70
14.4
ii
6
66
14.4
12
3
36
11.4
13
3
39
9-5
H
3
42
12.5
From the Psychological Laboratory of the University of California.
8l
82
WARNER BROWN
according to the length of the lists and shows the number of
persons who gave a list of each length from o to 14 items.
Table II. shows, roughly, the number of mentions accorded to
TABLE II
The Number
of Items Men-
tioned
The Number of
Mentions Received
by Each of These
Items
The Average
Position of These
Items in the Lists
The Number of
Times One of These
Items is Found
at the Head of a
a List, Per Cent.
The Number of
Times One of These
Items is Found in
the 8th or a Lower
Position, Per Cent.
;s3
152
(1
54
5-3
5-7
n-5
22.6
19.2
34
2
5-0
n.8
I9.I
8
3
4.6
ii. i
I4.8
H
4 or 5
3-9
14.8
14.7
8
6 or 7
5-2
7.8
19.6
8
7 or 8
4-5
11.9
17.0
7
8 or 9
5-o
10.2
18.6
6
9, 10, ii
44
17-5
10.5
4
12-15
3.6
16.6
7-4
3
15-18
3.8
28.0
IO.O
3
18-20
4.0
28.1
12.3
2
23,30
.8
24-5
iS-i
I
55
3-5
2O.O
5-5
I
55
2.9
30-9
5-5
I
58
3-o
24.1
5-2
different advertisements. Thus 105, or nearly half of them,
are mentioned by but one person, while three of them are
mentioned more than 50 times each. These three 'best'
advertisements (Arrow collars, Spearmint gum, and a local
confectioner), receive almost one fifth of all the mentions.
In what has just been said we have before us the essential
points upon which the investigation is to be based. Some of
the advertisements are much better remembered than others;
and at the same time some persons remember a larger number
of advertisements than other persons do. Is it true then,
that the advertisements which are remembered by the greatest
number of persons are the ones which permit of the easiest
recall on the part of those persons who can remember several?
The answer is obtained by finding whether the advertisements
which are forgotten by most persons are written down later
than those which can be recalled by more people. The result
shows (Table II.), that in the average, one of the 105 which
are mentioned only once is not mentioned until five others
have been mentioned. The most popular advertisement has
INCIDENTAL MEMORY 83
an average position of third. In other words the average
person will write down an item which many other persons can
remember, sooner than he will write down one which only a
few can remember. Less than 9 per cent, of the 105 straggling
items are found at the head of a list, while about 24 per cent,
of all the mentions of the most popular one are found at the
top of a list.
Table II. has been arranged to show the result when the
whole number of mentions is broken up into 16 approximately
equal groups of from 50 to 60 cases each, on the basis of
the frequency with which the items were mentioned. The
table shows the average position of the item in the list;
the items most frequently mentioned stand ahead of the rarer
items in the lists. It also shows for each advertisement or
group of advertisements the proportion of its mentions which
stand at the head of a list; those most frequently mentioned
are more apt to be mentioned at the very start. The corre-
lation of rank
6SZ)2
in order of mention, in the above table, with average position
in the list is .85, and with proportion of leading mentions it is
.76.
While conducting this investigation the writer labored
under the impression that those persons who mentioned only
a few items would be apt to have peculiar reasons for re-
membering the few advertisements which they could recall,
and that they would mention many wild and eccentric ex-
amples. The analysis of the data shows that such can not
be the case. The sporadic item very seldom occurs early in
any list; it can only rarely occur in a short list. The last
column of Table I. shows the relative frequency with which the
items in lists of different lengths are mentioned. This is
found by listing all of the items in all of the lists of a certain
length and entering opposite that item the total number of
times that it is mentioned in all of the lists, i. e., the frequency
of the item. The total amount of all of these frequencies is
then divided by the total number of items concerned. The
84 WARNER BROWN
results of this computation make it evident that the items
which occur in a very short list are items which are more often
remembered than the items which occur in longer lists. As
the length of the list increases it includes more and more
sporadic or infrequent items. The original data show that of
the 42 items contained in the one-item or two-item lists three-
quarters were mentioned by more than ten different people,
while of the 42 items contained in the fourteen-item lists
three quarters were advertisements receiving less than ten
mentions and one quarter were rarities, mentioned by less
than four persons. Turning to Table II., the last column, it
is evident that the less frequently mentioned items are more
apt than the others to find mention toward the end of a long
list, in the eighth or a still lower position. The correlation
between rarity of occurrence and low position in the list is .78.
The facts warrant the conclusion that the items in the
short lists are not determined by individual or special con-
ditions, but are simply the items which are most easily re-
membered. There seems to be good reason to believe that
the same factors, whatever they are, which cause an adver-
tisement, or other similar incidental impression, to be recalled
early in the memory of one individual cause it to be recalled
early in the memory of another, regardless of the number of
items which may, or may not, follow after it. The difference
between individuals in this respect seems to be a difference in
the number of the items recalled, and not in the kind or iden-
tity of the items. Items, which, for any reason, are difficult
to recall appear late in long lists and do not appear at all in
short lists.
This conclusion is significant for experimental work in
memory as it puts a new value upon the relative position
of the items in the recalled series. Apparently the first
items to be recalled are generally those which make a uni-
versal appeal; the special personal appeals are reported later,
or not at all. Moreover there seems reason to believe that
"poor" incidental memory involves, at least with this ma-
terial, no other abnormality than poverty or " weakness."
So far as the actual advertisements which were used in
INCIDENTAL MEMORY 5
the investigatipn are concerned, it is important to note that
some make a much more lasting impression than others.
The difference can not be expressed by saying that one ad-
vertisement appeals to more persons than another; it must
be stated as a difference in the strength of the appeal.
The good advertisement makes an appeal so strong that it
can not be forgotten; the poor advertisement is forgotten by
all except those persons who can remember very weak im-
pressions.
The results of this investigation are fully confirmed by a
repetition of the experiment some five months later with
another college class, among the members of which there were
only a few who had taken part in the first experiment. As a
result, perhaps, of the interest aroused by the first experiment,
the average number of items per student rose from 5.1 to 9.9.
The number failing to report any advertisements fell from 21
to 3 although the class was only a third smaller. There were
no lists of one or two items presented. In spite of the larger
number of items per student, the variety of the items increased
to such an extent that the average number of mentions per
advertisement only increased from 4.2 to 4.6. Under these
changed conditions all of the conclusions of the first experi-
ment were confirmed. The correlation between frequency
of mention and high position in the list was found to be .76.
Of the straggling single items only 7 per cent, were found at
the head of a list, while 24 per cent, of the mentions of the
most popular advertisement headed lists. The correlation
between popularity and primacy was found to be .61. On
the other hand the correlation between infrequency and a
position somewhere lower than ninth on the list was found to
be .76.
In conclusion it may be said that the items which appeal
to the largest number of persons make the strongest appeal
to most of those persons, and that those items which appeal
to only a few make a weak appeal even to them.
VOL. XXII. No. 2 March, 1915
THE PSYCHOLOGICAL REVIEW
A PROPOSED CLASSIFICATION OF MENTAL
FUNCTIONS
BY GEORGE A. COE
Union Theological Seminary
Whenever anything is declared to be a function of mind
we should be able to discover both the general sense in which
the term * function' is used, and also the setting of the par-
ticular function in question within a functional whole. This
is as much as to say that classification of mental functions
should have a place in functional psychology that will cor-
respond to the position now occupied in structural psychology
by lists of mental elements and modes of combination. Up
to the present time such a systematic background has been
lacking. As a consequence the undefined fringe of meaning
in discussions of functions leaves still too much room for
misunderstanding one another, or even oneself. Further,
the lack of classification implies that we are not yet ready
to begin describing functions in terms of functional laws.
Such is the unsatisfactory situation out of which the present
article attempts to take a single step. The results are neces-
sarily preliminary and tentative; the most that I can hope for
is that other investigators will be sufficiently interested to
make good my deficiencies.
The approaches thus far made toward a classification of
mental functions fall into the following classes:
(a) Affirmations of the purposive character of mind, with-
out any list of specific functions.1
1 E. g., J. E. Creighton, 'The Standpoint and Method of Psychology,' Phil. Rev.,
March, 1914; H. Munsterberg, 'Psychology, General and Applied,' 1914, and R. M.
Ogden, 'Introduction to General Psychology,' 1914.
87
88 GEORGE A. COE
(b) The oft-made assertion that the fundamental functions
of all life, mind included, are nutrition and reproduction.
At a later point I shall ask what, as a matter of fact, mind
does with these two vital processes. At once, however, I
would point out that some of the so-called 'irradiations' from
primitive hunger and love — for example, science — have char-
acters of their own which it requires some violence to call
either nutritive or reproductive.
(c) To each item in a structural classification of mind
Angell has added the question, What is its function? There
results what might be called an engineer's drawing of mind as
an adjusting mechanism. It goes far toward supplying the
functional classification that I am seeking, and as a conse-
quence I shall borrow rather freely from it. That it needs
supplementing, however, should be clear from these two
considerations: First, Angell's list of functions is not based
upon similarities and differences among the functions them-
selves; he merely finds and describes a function for each
element of structure. Second, his genetic method keeps his
eyes fixed upon the earliest mental reaction, the terminus a
quo, whereas our problem — the direction of mental movement
— requires us to consider also the most developed reaction as
a terminus ad quern. I find no fault with Angell for not
answering questions that he does not raise, but functional
psychology must surely incorporate into itself a fuller de-
scription of the interests of developed mind. After we have
named early utilities, and even after we have made such
generalizations as that mind extends the control and organ-
ization of movements, something in the nature of function
still remains over. To illustrate: If you should ask what are
the functions of a dividing engine, I might answer by showing
how each wheel and lever contributes to the accurate control
of movement, and I might generalize by saying that this
instrument as a whole has the function of so adjusting our
motions as to enable us to make extremely minute divisions
of a surface. This would be a functional description, no
doubt, yet beyond it lies the destination of the whole, namely,
certain sciences in the interest of which the dividing engine
MENTAL FUNCTIONS 89
exists at all. Just so, the proposition that mind increases
the extent and the fineness of our adjustments needs to be
supplemented by inquiry into the terminal meaning of the
whole.
(d) A fourth approach to a functional classification pro-
ceeds as follows: Mental functions are correlative with inter-
ests; interests have their roots in instinctive satisfactions;
therefore an inventory of instincts would be ipso facto a list of
the functions of mind. Let us, then, look to our original
nature, that is to our unlearned tendencies to react in specific
ways, to give attention to specific sorts of object, to take satis-
faction in predetermined kinds of mental occupation. The
program is attractive, and we shall see that it yields results
that have an important bearing upon our problem, though
not quite the results that are commonly expected. For, first,
the * original' nature of man means the part of his nature that
is disclosed antecedently to all culture, that is, before the
mind has performed some of its most characteristic acts.1
Second, the broad mental areas traditionally called instincts
are disappearing from the psychologic map, and in their
stead there is appearing a vast, indefinite number of narrow
adjustment acts. For example, Thorndike says that "reach-
ing is not a single instinct, but includes at least three somewhat
different responses to three very different situations."2 Thus,
the farther back we go in our mental history the greater the
difficulty of functional classification, unless we constantly
look forward as well as backward. On the other hand, the
very minuteness and rigor of Thorndike's analysis reveal
certain general, forward-looking tendencies. Thus, there is
a tendency to be or to become conscious;3 there is an original
'love of sensory life for its own sake';4 there is spontaneous
preference for experiences in which there is mental control;5
finally, there is a native capacity for learning.6 In short,
1 E. L. Thorndike, 'The Original Nature of Man/ 1913, ^198 f.; also 'Education/
1912, Ch. v.
2 'Original Nature/ 50.
3 Ibid., 170 f.
«i4i.
6 141 f.
•171.
90 GEORGE A. COE
there are 'original tendencies of the original tendencies . . .
original tendencies not to this or that particular sensitivity,
bond or power of response, but of sensitivities, connections
and responses, in general.'1 Here, I take it, is where interests,
in the proper sense of the term, come in. If we are to define
our mental functions by our interests, we must consider not
merely tendencies to this or that sensitivity, but also and
particularly our tendencies to organize or do something with
our sensitivities. Some results of Thorndike's analysis of
such tendencies I shall take over into my own classification.
(e) Some of the conditions for a general classification of
mental functions are fulfilled in recent discussions of value.2
Here function is treated as function; it is not confused with
elements of structure, nor is a given function identified with
its earliest, crudest form. Sense of direction from something
to something is here. Urban's list of values, in particular,
conveys a sense of the general direction of the movement of
mind. What is still needed is something like a combination of
Angell, Thorndike, and Urban. The reason why lists of
values need supplementing is twofold : First, they do not com-
prehend mind as a whole, for example, its biological aspects.
Second, several types of value, as will presently appear, are
not simple functions, but functional complexes.
These converging lines in recent psychology may be sum-
marily described as follows: (i) All mental process what-
soever is purposive, and it should be analyzed from this as
well as from the structural standpoint — that is, mental functions
must be determined. (2) The human mind is functionally as
well as structurally continuous with the animal mind, so that
a classification of functions must include the biological point
of view. (3) The termini of mind, by which functions are
defined, include conscious interests, or self-defining ends.
(4) Several specific functions of both the biological type and
the conscious-interest type, have been defined here and there
in scattered places.
What remains to be done is to systematize these results;
1 170.
'The chief classifications of value are summarized by J. S. Moore 'The System
of Values/ Jour. Phil., VII (1910), 282-291.
MENTAL FUNCTIONS 91
to discover, and if possible, fill remaining gaps; and to show
the relation of the resulting functional concepts to older,
more current psychological categories. The whole must, of
course, be description, not evaluation. The work of func-
tional psychology is not to tell us what we ought to prefer,
but to determine, as a matter of observable fact, what mind
does actually go toward and 'for.' Two main divisions,
each with several subdivisions, are implied in what has already
been said.
A. Biological Functions. — To occupy the biological stand-
point— which is simply a point of view used temporarily for
certain purposes, and not necessarily more true or funda-
mental than other points of view — is to think of living beings
without reference to any approvals or preferences, any
'better and worse.' The biological functions of mind consist
in quantitatively determinable increases in range of response
to environment. Our subdivisions of biological functions,
accordingly, are as follows:
1. Increase in the spatial range of objects responded to.
2. Increase in the temporal range of objects responded to.
3. Increase in the range of magnitudes to which response is
made.
4. Increase in the range of qualities responded to.
5. Increase in the range of environmental coordinations to
which coordinated responses are made.
This list will remain the same whether we approach the
facts from the behaviorist standpoint or from that of tra-
ditional psychology. I call these functions mental for two
reasons: Because they characterize mind in its most conscious
as well as its less conscious stages, and because these 'direc-
tions of movement' though they are established before we
reflect upon them, become, after reflection, conscious pur-
poses.
The relation of this analysis to the popular categories,
nutrition and reproduction, requires a word of explanation.
To begin with nutrition, what has mind, as a matter of fact,
to do with it? (a) Mind connotes changes in the feeding
reaction that fall under one or more of the above-listed
92 GEORGE A. COE
functions. But the law here is a general one; it applies
likewise to protection from weather, from accidents, and from
enemies, and it applies also to social organization, science,
and art. As far as range of response is concerned, then, we
need no special nutrition category, (b) Mind connotes
success in a competitive struggle over a limited supply of
food. Increase of mind makes a difference here, but in what?
Can the difference be expressed in terms of nutrition ? No; for
nutritive functions would go on at least as well if no com-
petition occurred, or if the mentally inferior animal had
happened to get the food instead of the mentally superior
one. The difference made by mind is that some new object
or quality is responded to, and that the more differentiated
response tends to be perpetuated by inheritance. Here the
function appears to be not nutrition but the production of a
more specialized individual, (c) It is at least as correct to
say that mind moves away from as toward nutrition. For,
correlative with the growth of mind goes restriction of feeding
to specialized kinds of food, and consequent increase in the
mechanical cost of getting it. The ocean brings food to an
oyster; a cat must hunt for its living. Everywhere the dis-
criminative appetite is the expensive one. (d) If we scru-
tinize cases in which feeding appears to be the end of conscious
effort, we find, almost if not quite invariably, that the very
act of consciously seeking food gives to nutrition the place
of means to some experience beyond itself. The labor move-
ment illustrates this principle on a large scale. Even if the
central stimulus of this movement could be identified as
hunger (which is doubtful), the conscious end of the strug-
gle is home life, leisure, culture, the education of children,
free participation in the determination of one's destiny, (e)
But it may be said that the social integration of men has as
one of its most obviously important consequences the sta-
bilizing of the food supply and a more even distribution of it.
Civilization will soon reach a point at which famines can no
longer occur. What, it may be asked, is the meaning of the
present movement for agricultural instruction, and indeed
for vocational training in its whole extent, if not just this,
MENTAL FUNCTIONS 93
that men want enough to eat? Here, indeed, is excellent
material for answering the question what mind is about when
it seeks food. The crucial question for us is whether the
direction of the mind's movement here can be defined as
from hunger to repletion. Of course food is an object of
conscious desire. So is getting to Albany on time an object
of desire on the part of one who is travelling from New York
to Buffalo by way of the New York Central. The road to
our social ends certainly takes the food-supply route. But,
as in the case of the labor movement, social food seeking that
begins instinctively awakens, sooner or later, a consciousness
of the social values broadly called cultural, and these it is
that define the specifically mental destination or function.
Turning now to the question whether reproduction should
be accounted a mental function, we find the course of evolu-
tion not at all ambiguous. Reproduction is most prolific in
the lowest ranges of life. Mental development is clearly
correlated with decrease in the birth rate. How many fac-
tors are involved in this decrease I will not attempt to say,
but certainly mind is one of them. Herbert Spencer realized
this fact,1 though he did not bring out the full significance of
it. John Fiske's two essays on human infancy2 carry us
much farther. Mind individualizes the various living beings
that are involved, first the offspring and then the parents.
The obvious mental function is not reproduction of existing
types, but the production of certain new, more specialized
types. Mind does not stimulate reproduction any more than
it stimulates hunger; it does not increase fertility any more
than it increases assimilation. But, just as mind specializes
foods and increases the cost of feeding, so it individualizes
living beings and increases the cost of each individual. The
whole may be viewed as on the one hand an increase of in-
hibitions, and on the other hand a focalizing of dispersed at-
tention. In short, the biological functions of mind can be
altogether expressed as increase in the range of objects and
1 'Principles of Biology,' Part VI., especially Chs. XII. and XIII.
2 Reprinted under the title, 'The Meaning of Infancy,' in the Riverside Educa-
tional Monograph series, Boston, 1909.
94 GEORGE A. COE
qualities responded to, and in range of coordination of re-
sponses.
B. Preferential Functions. — Our discussion of nutrition
and reproduction has already brought us face to face with
conscious preferences, that is, mind defining its own direction.
We may take for granted, I suppose, that satisfactions are,
in general, a sign of unimpeded mental action, and that we
can tell one another about our satisfactions. One may, in-
deed, be mistaken as to what one likes, that is, as to what it
is in a complex that makes it likable, but such mistakes can
be discovered and corrected, chiefly by further communi-
cation. The functions of our second main division, then, are
always qualitative (implying a ' better and worse'), and they
are scientifically known through communication by means of
language. Thus it is that many preferences have already
been successfully studied, such as color preferences, the likes
and dislikes of children with respect to pictures and with re-
spect to future occupations, merit in handwriting, merit in
English composition, merit as a psychologist, the comic,
persuasiveness, even moral excellence.1 Such experimental
studies have the effect not merely of discovering preferences,
but also of adding precision to preferences already recorded
in the world's literature. Would that a Hollingworth might
have been present throughout human evolution to record the
growth of human preferences. As the case stands, we must
combine experiment upon present preferences with the less
precise study of life as reflected in literature, art, and insti-
tutions.
Where shall we look for a basis for the systematic sub-
division of preferential functions? Suppose we compare
early types of reaction with late ones, say, Thorndike's pic-
ture of original nature with value-analyses, which represent
developed interests. Let us begin with the fact that there is
satisfaction in merely being conscious. To be conscious,
then, we may count as the first preferential function. Note,
next, that satisfaction attaches to mere movement of atten-
1H. L. Hollingworth gives a list of 'order of merit' researches in 'Experimental
Studies in Judgment,' Archives of Psychology, New York, 1913, 118 f.
MENTAL FUNCTIONS 95
tion from one object to another, as in 'love of sensory life for
its own sake.' May we not say that a second preferential
function of mind is to multiply its objects? A third appears
in the preference for experiences that include control of ob-
jects. A fourth is closely related thereto, namely, the ar-
rangement of objects in systems — it is a function of mind to
unify its objects. This is seen all up and down the scale from
the spontaneous perception of spatial figures in the starry sky
to the ordering of an argument.
These four preferential functions appear to be fundamen-
tal, that is, not further analyzable. If we turn, in the next
place, to the usual value categories to see whether we may
not find further unanalyzable functions, we come upon the
interesting, not to say strange, fact that ethical, noetic, re-
ligious and even economic values presuppose a function that
they do not name. Each of these types of value depends upon
the existence of a society of inter-communicating individuals,
yet it seems not to have occurred to anyone to include a
social category — simply and specifically social — in discussions
of either functions or values. Should not the fifth prefer-
ential function in our list, then, be the function of being social,
of having something in common with another mind, in short,
of communicating? The justification, not to say necessity,
for recognizing a simply social function of mind, exists not
alone in the social presupposition of several recognized values,
but also in a long series of genetic studies which, from one
angle after another, have revealed the fundamentally social
nature of consciousness.1
There remains for consideration our esthetic experience.
Doubtless it involves functions already named, particularly
1 It is true that these are commonly studies of content rather than of function,
and that 'P and 'thou' appear therein as 'idea of I* and 'idea of thou.' For the pur-
poses of merely structural analysis this is doubtless sufficient. That is, structural
analysis as such has no place at all for the experience of communication. On the
other hand, communication will loom large in any adequate general analysis of mental
functions. In an article 'On Having Friends: A Study of Social Values ' (Jour. Phil.,
XII. (1915)* iS5-i6i), I essay a functional treatment of one easily accessible social
experience. Satisfaction in having a friend I show to be satisfaction in a second
experiencing. Social experience like this is distorted whenever attempts are made to
construe it without the -ing.
96 GEORGE A. COE
the functions of unification and communication. But it
seems to contain also an attitude somewhat different from
those already named, the attitude of contemplation— the
taking of satisfaction in objects merely as there, without re-
gard to anything further that may happen to or with them.
Hence I add contemplation to the list.
The preferential functions, then, are these:
1. To be conscious.
2. To multiply objects of consciousness.
3. To control objects, oneself included.
4. To unify objects, oneself included.
5. To communicate, that is, have in common.
6. To contemplate.
Some omissions from this list require explanation. Play
is omitted because it involves a complex of I and 2, generally
3 also, sometimes all six, and because it is fully exhausted
therein. Truth is omitted because, as far as it is not an ab-
straction from actual intellectual functioning, I hold it to be
analyzable without remainder into functions already named,
especially 4 and 5.1 No ethical function appears because the
three objectives that it includes — control, unification, sociali-
zation— already have appropriate recognition in the list.2
As to economic value, it seems to be exhausted in the notion
of control within a social medium. Finally, religion is with-
out a place in the list because it offers no particular value of
its own. Religion is not coordinate with other interests, but
is rather a movement of reinforcement, unification, and re-
valuation of values as a whole, particularly in social terms.3
It will be asked, no doubt, whether the functions of mind
can be named without any direct reference to instinctive de-
sires. In addition to what has already been said concerning
nutrition and reproduction — that they are, so to say, con-
stants that find a supply at every level of mentality — it may
now be added, as a general truth, that mental activity upon
the objects of instinctive desire does not satisfy the desire in
1 Cf. A. W. Moore, 'Truth Value,' Jour. Philos., V (1908), 429-436.
1 Cf. J. H. Tufts, 'Ethical Value,' Jour. PhUos., V (1908), 517-522.
«G. A. Coe, 'Religious Value,' Jour. Philos., V (1908), 253-256.
MENTAL FUNCTIONS 97
its initial form, but modifies the desire itself. For example,
what has at first only a derived interest as means to something
else may acquire an interest of its own, become an end. This
is surely the way that science has come into being, and very
likely art also. The evolution of parental and of conjugal
relations offers abundant examples of the truth that the dis-
tinctive work of mind with our desires is to differentiate and
recreate them. Our list of mental functions, accordingly,
need not specify particular desires, but only the primary
ways in which mind works among them.
A question may arise, also, whether higher desires or ideal
values ought not to appear in the list. Is not the most dis-
tinctive achievement of mind in the realm of desires, it may
be said, the mastery of certain ones in the interest of others?
I agree that 'the desires of the self-conscious' must be recog-
nized as having a character of their own,1 and that a list of
mental functions must do justice to them. "The valuation
of persons as persons constitutes a relatively independent
type, one which presupposes a differentiation of object and
attitude."2 The list as it stands, however, will be found to
do justice to this differentiation. Here are self-control, self-
unification, self-socialization, with the implication that all
this applies to any and every self, both actualized selves and
ideal selves.
Finally, inasmuch as no pleasurable sense-quality of ob-
jects is mentioned in the list, but only 'objects, oneself in-
cluded,' doubt may arise whether the functions here named
are not merely formal and contentless. Functions would
indeed be merely formal if they were so defined as to imply
indifference to the specific qualities of things. 'Pure in-
tellect' is certainly a mere abstraction, never a function. In
a list of 'preferential functions,' however, satisfactions are
everywhere presupposed, not ignored. Granted both pain-
ful and pleasurable objects as data, our question is what
mind does with such data. Psychology is of course free from
the old hedonistic fallacy that the only thing we can do with
1 A. O, Lovejoy, 'The Desires of the Self-Conscious,' Jour. Philos., IV (1907),
29-39.
2 W. M. Urban, 'Valuation, its Nature and Laws,' London, 1909, 282; see also 269.
98 GEORGE A. COE
pleasures is to seek for them, and that the only thing we can
do with pains is to avoid them. What we try to do in the
presence of such data is to control and organize them, sifting
out an item here, deliberately enlarging an item there, all in
the interest of being, so to say, at home with oneself and with
one's fellows. In short, the preferential functions here named
represent minds as mutually attaining freedom in the world
as it is. Such minds are as concrete as anything can con-
ceivably be.
COLOR THEORY AND REALISM1
BY KNIGHT DUNLAP
The Johns Hopkins University
The suggestion made by McGilvary concerning a possible
color theory in harmony with a doctrine of sensational
realism2 seems to me very important. No color theory, so
far as I am aware, has attempted to do more than satisfy
the psychological requirements (some have been contented
with much less) ; but it is possible that we may find a theory
which will satisfy both the psychological requirements and
the requirements of realism. If such a theory be found, it
will be a matter of interest and importance.
I should like to point out, in the first place, that none of
the prominent theories can be given a realistic turn. The
Franklin theory, and the Hering theory with its variants,
fall short on the psychological side in so many places (as, for
instance, in the inability to account for the perception of green
without red) that they may be dropped out of consideration.
The Hering theory, it is true, would perhaps be satisfactory
to the realist, if it were psychologically defensible; but the
Franklin theory being a physiological interpretation of the
current form of the Young-Helmholtz theory, is not acceptable
to the realist, for reasons which I will mention below.
The Young-Helmholtz theory, as at present held, is
satisfactory to the psychologist, largely because it is, as I have
elsewhere explained,3 merely a psychological schematization,
elastic enough to take in all of the accepted data of color
vision, but not concerned with definite hypotheses in the
physiological unknown.
The Young-Helmholtz theory assumes red, green, and
1 Written in 1912, but withheld from publication on account of the crowded
condition of the REVIEW.
2 1912, Philosophical Review, XXI. (2), 171. If I am misrepresenting Dr. Me-
Gilvary's realism (which is more than possible), I offer my apologies.
5 'System of Psychology,' p. 73, footnote.
99
ioo • KNIGHT DUNLAP
bluish-violet (indigo) as the three primary colors; not because
they are the only ones on which the theory can rest; but
because it happened to start out with them for simplicity's
sake, and has as yet seen no sufficient reason for changing.
The statement of Helmholtz on this point is significant:
"Der Wahl der drei Grundfarben hat zunachst etwas Will-
kurliches. Es konnten beliebig jede drei Farben gewahlt
werden, aus denen Weiss zusammengesetzt werden kann.
Young ist wohl durch die Riicksicht geleitet worden, dass die
Endfarben des Spectrum eine ausgezeichnete Stellung zu
beanspruchen scheinen. Werden wir diese nicht wahlen,
so miisste eine der Grundfarben ein purpurner Farbenton sein,
und die ihr entsprechende Curve in Fig. 119 zwei Maxima
haben, eines im Roth, eines im Violett. Es ware dies eine
complicirtere, aber nicht unmogliche Voraussetzung. So viel
Ich siehe, giebt es bisher keine anderes Mittel, eine der Grund-
farben zu bestimmen, als die Untersuchung der Farbenblin-
den."1
So far, no exception can be taken to Helmholtz. But in
the consideration of color blindness, he made a mistake which
was quite excusable at the time, and assumed that in the
ordinary cases of dichromopsia the patient sees green and
violet, since the natural explanation of dichromopsia seemed
to be the absence of one of the three processes. Hence, he
found no reason for abandoning Young's primaries. We now
know (or at least it is generally believed) that the two colors
seen in typical dichromopsia are yellow and blue or some
color near blue. This gives us the alternative of abandoning
Young's set of primary colors, or of stating the parachro-
mopsias in terms other than those of absence of one of the
three processes. The latter alternative has been chosen by
adherents to the Young-Helrnholtz theory,2 and for psycho-
logical purposes is quite satisfactory so far. It might be
thought, however, from the statement quoted above, that
1 'Physiologische Optik,' Zweite Abschnitt, § 20 (pp. 292-293 in the first edition).
* Opponents of the three-color theory make no mention of this, but refute the
theory in its older form, which is much more satisfactory for their purposes. So far
as I can find, the only modern statements of the theory in English text-books are in
Greenwood's 'Physiology of the Senses' and my 'System of Psychology.'
COLOR THEORY AND REALISM 101
Helmholtz would rather have considered the certification of
yellow-blue dichromopsia as an indication that the three prim-
aries should be changed.
The acceptance of green and red as primaries prevents a
realistic interpretation of the phenomena of color vision, since
a really red object is (probably) seen as yellow by the di-
chromopsic individual, whereas red is not compounded from
yellow, but yellow from red, according to the theory. There
may be some way of avoiding this difficulty, but it seems to
me that the acceptance of the Young-Helmholtz theory in
the present form, or of the Franklin physiological theory based
on it, entails the assumption that colors are purely private
content. So far as we can assume psychologically, some
person or animal might see as blue, or as any other color, or
even as the note b|?, the object I see as red.
It seems to be possible to modify the Helmholtz theory in
such a way as to make it compatible with sensory realism;
and possibly the modification would be just as satisfactory
psychologically. We might eventually have to assume four
colors instead of three, but for the present we may deal with
three, with the addition of white (gray). This addition is
necessary because of the fact that all sorts of parachromopsias
seem to agree in the perception of white in quite a normal
fashion (or at least it is now generally believed that this is
the case).
If the three colors are yellow, the spectral color usually
called blue-green, and the purple complementary to green,
the theory seems to work out very well. As a matter of fact,
if the total range of colors be studied without prejudice, this
triad is seen to be at least as satisfactory (so far as mere
qualitative comparison goes), as any other. Let us call the
three colors yellow, peacock, and mauve (the hues commonly
indicated by the last two of these terms are near enough to
the colors in question to serve the purpose), and the white or
gray neutral. We may then use the letters Y, P, M and N
without confusion. Red is a mixture of Y and M; green a
mixture of Y and P; blue and violet are mixtures of P and
M. Practically all colors contain some N. Hence the fact
102 KNIGHT DUNLAP
that in complete achromopsia only N is seen, in certain para-
chromopsias only the Fof the red (Y + M) combination, and
in others (possibly) only the P of the blue and violet (P + M)
combinations, is perceived, is quite intelligible. It is quite
conceivable, 'in other words, that the M curve of the 'slow'
end of the spectrum might alone be lacking, or the M curve
of the * rapid' end, or even both together. Whether the
common cases of dichromopsia with shortened spectrum are
due to the absence of the M curves alone, or to the absence of
the slow M curve and the P curve, we can not decide from the
evidence so far. In the cases with unshortened spectrum,
the P curve alone may be missing. If there are transitional
cases between these two forms of dichromopsias (which some
investigators deny), they can also be accommodated in this
omnibus theory; for it may be pointed out that there are end-
less possibilities of difference in sensitiveness of the processes
present.
The occurrence of pure white in normal vision is to be
explained by the facts (which are established regardless of
theory) that the stimulation of any color process renders it
progressively less sensitive. If the eye is exposed to the
influence of yellow light, it sees progressively (expressed
loosely) less and less yellow, and more and more white. The
apparent intensity of any component, in other words, is
relative not only to the 'natural capacity' of the eye, but also
to its condition at any moment; and this condition depends
largely on preceding stimulations. We have to assume further
that any mixture of light rays to which the eye has become
well adapted, will therefore not stimulate the color processes,
although it will stimulate the N process. It is a significant
fact that the eye becomes as well adapted to yellow gaslight
or to the bluish light of the mercury vapor arc, as to daylight,
if it is subjected to one of these stimuli alone.
There is however a further necessary statement, which is
not a statement of fact unless our hypothetical theory be
true; and that is, that the three processes can not in any case
be simultaneously stimulated. Any two of them may be
set in action, but an additional stimulus, adequate, when
COLOR THEORY AND REALISM 103
occurring alone, to excite the third process, simply has an
inhibitory effect on the first two (although acting as an
increased stimulus to the TV process), unless this third stimulus
becomes relatively more intense than the first two, in which
case it alone is effective, the other two serving to lessen its
specific effect. This statement rather complicates the theory,
but, for that matter, the complication indicated occurs in
some form in every theory, even in the current Young-Helm-
holtz theory. While this realistic color theory seems to me
highly interesting, and is probably just as workable as the
accepted form of the Young-Helmholtz theory, I should not
be inclined to substitute it for the latter unless the former
theory should be shown to be the psychologically superior
one. A sufficient factor of community in the world of objects
may be given in the relational elements alone, and I think we
might accept this point on a basis of realism rather than of
idealism. It is however, well to bear in mind the realistic
theory of color, as it apparently offers welcome possibilities
of psychological research.
POINT SCALE RATINGS OF DELINQUENT BOYS
AND GIRLS1
BY THOMAS H. HAINES
Bureau Juvenile Research, Columbus, Ohio
The Yerkes-Bridges2 point scale for measuring intelli-
gence appeals to the worker in the province of the standardi-
zation of the development of human intelligence most strongly
as a very useful rationalization of the Binet-Simon method.
1. It puts the whole process on such a basis that is is con-
stantly self-perfective, — the norms approach the reality in
direct proportion to the increase in number of records of
normal children summarized.
2. Another advantage is that it becomes a simple matter
to plot curves for the growth of intelligence in different races,
in different sexes, and in different classes of society in the
same locality. All resulting curves and tabulations, being
made by the same examinational methods, are directly
comparable.
3. It is not to be denied, also, that the point scale method
of rating intelligence overcomes some arbitrariness in method,
in that it allows the individual to make his credits anywhere
along the line of some twenty tests, whereas the Binet method
makes the passing of a given year depend upon a fixed stan-
dard, passing four out of five given tests.
4. Partial credit given for partial results also commend
this method, as for example in 'words given in three min-
utes,' 'three words used in one sentence,' 'arrangement of
weights,' and 'counting backward from 20 to I.'
It was a very fortunate feature of the plan of the origi-
nators of the point scale, that they pursued the natural
1 Read before the American Psychological Association, at a meeting at Phila-
delphia, December 29, 1914.
* See 'The Point Scale. A New Method for Measuring Mental Capacity,' by
Robert M. Yerkes and J. W. Bridges, The Boston Medical and Surgical Journal,
Vol. CLXXL, Number 23, December 3, 1914, pp. 857-866.
104
POINT SCALE RATINGS 105
method of development, and relied almost entirely upon the
Binet material and methods. Nineteen of the twenty point
scale tests are Binet tests. Analogies are extra.
This makes it a very simple matter to conduct the two
examinations on the same person at the same time, or
rather to secure both a Binet and point scale rating of
a given person from one examination. One has simply to
follow the order of the point scale sheet. When this is
complete, it is only a matter of half a dozen short tests
to complete the Binet rating. This practice has been pur-
sued by the author in examining two hundred delinquents in
industrial schools in Ohio. The idea in mind, in the begin-
ning, was to try out the point scale in comparison with the
Binet-Simon method, to see what it was worth, as it was
frankly recognized that the data summarized in the point
scale, so far, from normal children, are rather limited, as com-
pared with those from the use of the Binet scale.
It was soon apparent that results from the two methods
showed a very close parallel, the point scale results tending a
little higher, as would be expected with abnormal adolescent
minors, because of the wider opportunities offered for securing-
credits, for any given year of intelligence age. But the
paralleling is so close in the low grade morons, and the agree-
ment is so close between the two methods in excluding in-
telligence defect in the higher grade delinquents, that dis-
parity between the two ratings seemed at once to afford
reasonable ground for doubt as to the value of the Binet
findings.
We formed, therefore, a class of doubtful cases, — cases in
which Binet rating made the child less than twelve years old
mentally, and point scale credits were beyond twelve years.
The close correlation of results by the two scales is the first
result of interest in this work. The second is the value of
the point scale as a check upon the Binet scale in determining
the line between intelligence defect and normal intelligence.
At the best, it is a delicate matter to decide between a high
grade moron and one who has no intelligence defect, and each
examiner must develop his own standards. But where a
106 THOMAS H. RAINES
boy makes 11.4 years Binet and 80 points by point scale, or
14 years, it is certainly safe to consider his Binet record as in
some measure accidental, and that he is more likely to make
good, than the boy whose Binet is 1 1.4 years, and whose point
scale is 71 points, or n.6 years. The latter seems much more
likely to prove himself a high grade moron, to have an irre-
mediable defect in equipment on the side of intelligence. Of
course, even here there is doubt. The point scale is a method
contributing to the finer shading of our doubts, and the more
precise statement of intelligence defects.
In scoring by the Binet method, we adopted the common
procedure of requiring four credits to pass a given year, mak-
ing the highest so passed a basal year, and allowing one year
additional for each five credits. Half credits were often
given, as e. g., in two definitions superior to use in the nine-
year tests, copying one design in the ten-year tests, or giving
solution to one problem in twelve-year tests. Mentality was
reckoned to tenths of years. In estimating mental age by
the point scale, the spirit of the method was violated to this
extent, that we assumed the lines of the curve between each
two consecutive years to be straight, and we set down the
equivalent to a point rating, in years, reckoned to the nearest
tenth of a year.
The classification adopted in the first two hundred cases
examined, one hundred at the Girls' Industrial School, and
one hundred at the Boys' Industrial School, resulted in the
following grouping. The average mentalities are given for
each group.
GIRLS.
1. 34 low and medium morons, ranging from 8 to 10.5 years
mentally.
Binet average, 9.4 years. Point scale average = 54.5
points, or 9.1 years.
2. 24 high-grade morons, ranging from 10.5 years up.
Binet average, 10.9 years. Point Scale average = 70
points, or 11.5 years.
3. 13 doubtful cases.
Binet average, 11.5 years. Point scale average = 80.1
points, or 14 years.
POINT SCALE RATINGS
107
29 no intelligence defect.
Binet, 22 = 12 yrs. -f
5 = 12 "
2 = 12 " -
Point scale average = 84
points.
BOYS.
1. 2 high-grade imbeciles, each making 32 points = 6.6 years,
point scale.
2. 40 medium and low-grade morons.
Binet average, 9.4 years. Point scale average = 56.2
points, or 9.5 yrs.
3. 25 high-grade morons.
Binet average, 10.9 years. Point scale average = 81.2
points, or 14.5 years.
4. 20 doubtful cases.
Binet average, 11.4 years. Point scale average = 81.2
points, or 14.5 years.
5. 13 no intelligence defect.
Binet 6=12 yrs. + Point scale average of the seven
I = 12 yrs. 84.6 =15 years + 2j points.
Six of the thirteen were less than two years retarded.
The mean or average variations from these averages so
far as calculable, both for years, and for points, are embodied
in the following table:
Av. Binet
M.V.
Av.
P.S.
M.V.
Equivalent
Years
M.V.
GIRLS
Low and medium morons .
High morons
94 yrs.
IO.Q yrs.
.39 yrs.
.4.3 yrs.
54-5
7O
5-5 =
-z.c =
9.1 yrs.
ii yrs.
.58 yrs.
.32 vrs.
Doubtful intell. def
No intell. def
1 1.$ yrs.
22 are 12 -f- yrs.
.24 yrs.
80. 1
«1
3-5 =
7.C =
14 yrs.
it; yrs.+
BOYS
Low and medium morons .
High-grade morons
Doubtful
5 are 12 yrs.
2 are 12 — yrs.
94 yrs.
10.0 yrs.
11.4. yrs.
.50 yrs.
.42 yrs.
.20 yrs.
|6.2
67
81 ?
5-3 =
3-5 =
2.Q =
2 pts.
9.5 yrs.
1 14 yrs.
14.. c yrs.
.64 yrs.
.24 yrs.
No defect
•* -t- /iu
o are 1 2 yrs -p
84.6
27 —
1C vrs -4-
I is 12 yrs.
2 pts.
These figures demonstrate the point scale, at the present
state of development, to be quite as accurate a means of
measuring the intelligence of high grade defectives as is the
loS THOMAS H. RAINES
Binet scale. In the groups directly comparable, by mean
variations, we find the low and medium morons, both boys
and girls, yielding considerably larger mean variations for
point scale equivalents in years, and the high grade morons,
both sexes, yielding considerably smaller mean variations
for point scale equivalents in years. One could, of course,
throw these comparisons either way by manipulating the
data. The effort in grouping was to give equal weight to
each set of figures. That this was done with fair success is
indicated by the close approximation to each other, of the
average mental ages, obtained by the two methods, in each
of these two groups for both sexes.
The mean variations from the point scale averages, for
the four groups, for each sex, indicate that we have made the
three highest groups of about the same ranges of mentality,
whereas the lowest has nearly twice the range. The low and
medium grade morons could readily be set apart into two
groups each, having a range equivalent to that of the high
grades morons. The M. V. of each group would lower ac-
cordingly. It is a great satisfaction to feel that we have
even a tentative means of assessing the intelligence develop-
ment, between ten and fifteen years, in the normal child.
Small and insignificant as are the point differences in these
years, the point scale begins to let in the light upon these
differentiations, which the Binet scale has left in the dark.
When more data are collected, qualitative or analytic studies
of the differences in the point scale findings with children
rating within these years, both normal and abnormal, may
be expected to yield significant results, — significant for
diagnosis of the mentality of the exceptional child.
In view of the disparity of the results obtained by the
two methods, in our group of ' doubtful intelligence defect,'
two things at once occur to one. (i) It will be interesting to
note the future history in the particular cases, to see whether
such a doubtful child proves himself to be 11.5 years mentally
as per Binet or 14 years as per point scale. There must
remain a blot upon his intelligence, until he disproves it, for
he is likewise retarded by both ratings as compared with
POINT SCALE RATINGS 109
the group above. It is rare for members of this group to
achieve the 82 points of 15 years, while the higher group
averages 84 points.
(2) The other matter in regard to this disparity holds in
regard to the 'no defect' group also. In both groups and
for both sexes the points attained by point scale rating cor-
respond to higher years on the point scale curves, than are
attained by the Binet rating. One cannot avoid the sus-
picion that the numbers of normal cases so far rated in these
higher years by the point scale, may be so few and so ex-
ceptional that we have averages which are too low for given
years, and that more work with normals will bring these
figures higher. The close correspondence between the two
ratings for the lower groups suggests this.1 A thousand
normal children, ten to fifteen years old, rated and grouped
by point scale, would, in any event, bring needed light in this
region. The same results studied analytically would be of
great service in furthering insight into mental organization
and development in late childhood and adolescence.
1 Professor Yerkes states that point scale and Binet ratings, resulting from his
own examinations of more than fifty normal children, show that the Binet ratings
for children below the mental age of eight are too high, and for children above the
mental age of eight they are too low. Professor Thorndike comes to a similar view
by taking Dr. Goddard's results of the examinations of two thousand school children
in Vineland, New Jersey, and working out the averages and median values. For
Thorndike's results see the Psychological Clinic, December 15, 1914.
A PRELIMINARY STUDY OF THE DEFICIENCIES OF
THE METHOD OF FLICKER FOR THE
PHOTOMETRY OF LIGHTS OF DIF-
FERENT COLOR1
BY C. E. FERREE AND GERTRUDE RAND
Bryn Mawr College
SYNOPSIS
A satisfactory method of photometry should combine the following features.
(i) It should enable one to detect small differences in luminosity and to reproduce
results for a given observer with a small mean variation and for a number of observers
with a comparatively small mean variation. That is, the method should possess an
adequate degree of sensitivity. (2) It should be known either to possess of itself logical
sureness of principle or its results must agree in the average with those of some method
which can be shown to have this sureness of principle. The method of flicker probably
satisfies the first of these requirements better than the equality of brightness method.
It does not, however, possess of itself the needed sureness of principle, nor have its
results been shown to agree in the average with any method which is accorded sureness
of principle. Points are enumerated in the paper appended which raise doubt with
regard to the correctness of the photometric balance obtained by the method of flicker.
Only one of these is discussed, namely, the influence of the time element in the exposure
of the eye to the lights to be compared. With regard to this point, it is shown from
experimental data (i) that the sensations aroused by lights differing in color value rise
to their maximum brightness at different rates; and (2) that the single exposures used
in the method of flicker are much shorter than is required for these sensations to rise
to their full value. The eye, therefore, is very much underexposed to its stimulus by
the method of flicker. That is, the rate of succession used in the method of flicker is
too fast for the single impressions to arouse their maximum effect in sensation and too
slow for the successive impressions to add or summate as much as they would need
to do to rise to their full value or perhaps even to a higher value than would be given
by the individual exposures. Only one other possibility for a correct balance remains, —
equality is attained at some value lower than the full value. This can not be assumed,
however, without violating well-known laws relating to the factors which influence
persistence of vision.
The principal point of discussion, then, is to what degree it should be held that the
difference in lag between the sensations aroused by the single exposures used in the
method of flicker is obliterated in a succession of exposures. Broadly considered, three
positions are possible with regard to the point for the rates of succession that are
employed in the method of flicker, (i) The difference is not obliterated at all. In
this case the photometric balance should deviate from the true balance in direct pro-
1 Paper read by C. E. Ferree at the Philadelphia section of the Illuminating
Engineering Society, January 16, 1914.
1 10
FLICKER PHOTOMETRY m
portion to the difference in lag for the single exposures. (2) The difference is in part
obliterated, but it is still present to a degree which renders the method untenable for
precise work. And (3) the difference is entirely obliterated or so nearly so as to be of
no practical consequence to the validity of the method. The second is approximately
the position taken in this paper. The following evidence is offered in support of this
position, (a) At high intensities of light the writers get by the method of flicker a
deviation from the true photometric balance, as determined by the equality of bright-
ness method, in a direction which corresponds to the difference in lag between these
colors at high intensities as determined both in their own laboratory1 and by Broca
and Sulzer. (£) At low intensities they get a difference in lag for the colors which is in
the same direction as the deviation obtained by Ives and Luckiesh at low intensifies
(the reverse Purkinje effect), (c) A change in the relative lengths of exposure to the
two lights in the method of flicker produces a deviation from the equality of brightness
balance which again corresponds in direction to the changes that are produced in the
sensations aroused by the single exposures when similar changes are made in the
relative lengths of exposure. And (d) determinations made at several intensities of
light by the method of flicker show a deviation from the equality of brightness balance
which is many times the smallest difference in brightness that can be detected by the
method. Moreover, in their own results the writers find that these deviations in every
case correspond to the difference in lag given by lights of the same order of magnitude
of intensity, so far as can be judged from the determinations of lag that have been made
up to this time. When, however, determinations have been made on a larger number
of observers, individual differences may be found in the amount and distribution of lag
just as they have been found in the amount and direction of the deviation of the flicker
from the equality of brightness balance. Later in the interests of a fairer comparison
the writers hope to make in every case compared the determination of lag and the photo-
metric determinations on the same observer.
The writers have preferred to call the work of which this
paper is a brief report a preliminary study for the following
reasons, (i) Only one of the points directly pertaining to
the method of flicker that should be investigated has been
taken account of in the work. And (2) to complete the chain
of evidence needed for this point, a more especially directed
and perhaps more careful determination should be made
than has yet been done of the time required for visual
sensations colored and colorless to rise to their maximum of
intensity. Such a study with especial reference to the needs
of photometry is now in progress in our laboratory, but is as
yet unfinished.2
1 See this paper, footnote I, pp. 125-130.
2 In the work now in progress in our laboratory, attention will be paid to the
following points. In case of colors, care will be taken to use lights of a small range of
wave-length. The intensities of the lights used will be specified photometrically and
radiometrically. The white light will in addition be specified either spectro-photo-
metrically or spectro-radiometrically. For the sake of the comparisons needed in
H2 c. E. FERREE AND GERTRUDE RAND
A satisfactory method of photometry should combine the
following features, (i) It should enable us to detect small
differences in luminosity and to reproduce our results for a
given observer with a small mean variation and for a number
of observers with a comparatively small mean variation.
That is, the method should possess an adequate degree of
sensitivity. (2) It should be known either to possess of itself
logical sureness of principle, or its results must in the average
agree with those of some method which can be shown to
possess this sureness of principle. Methods having these
features have been developed for the photometry of colorless
light. The problem of the photometry of colored light,
however, has presented great difficulty.
METHODS OF PHOTOMETERING COLORED LIGHT.
The methods for photometering colored light may be
grouped under two headings: the direct methods and the in-
direct methods. In the former class we have the method of
direct comparison or, as it is sometimes called, the equality of
brightness method. Of the latter class the method of flicker
has received the greatest amount of attention and has been the
most favored. It will be the purpose of this paper (i) briefly
to compare the relative advantages and disadvantages of
the method of flicker and the equality of brightness method
with regard to sensitivity; (2) to show that the method of
flicker, so far as it has been developed up to the present time,
does not seem to possess of itself the sureness of principle
needed to meet the requirements of a satisfactory method;
and (3) to show that as yet its results have not been found
satisfactorily to agree in the average with those of any method
which can be shown to have this sureness of principle. In a
the photometric work, all determinations for lights differing in composition will be
made at the different intensities employed with stimuli equalized photometrically.
Comparative results will be obtained for the same observers for the best of the methods
already in use, and three new methods will be introduced. In part, results will be
obtained for observers who have also been employed in the work on the method of
flicker. The work will be done for different intensities of light, and both under
dark and light room conditions. In a survey of the work done up to the present
time, one can not help but note that too little care has been taken to observe even
some of the most essential of the above conditions.
FLICKER PHOTOMETRY "3
later paper a new method of photometry will be described which
possesses approximately as high a degree of sensitivity for color
work as the method of flicker and gives results which agree
much more closely in the average with those obtained by the
equality of brightness method. The second of the above points
will be shown as follows. It will be pointed out that at the rate
of speed at which the impressions are given in the method of
flicker, the eye is very much underexposed to its stimulus. It
can reasonably be assumed that this underexposure causes a
reduction of the intensity of the sensation, and should lead,
therefore, to a false estimation of the brightness of the colors.
In fact, at the rate of rotation of the exposure apparatus
required for lights of the order of intensity employed in
practical work, this reduction produces for the observers we
have used an effect similar to the Purkinje phenomenon.1
At least a deviation from the equality of brightness values
is found in our results for such intensities which accords well
with the Purkinje phenomenon. That is, reds and yellows
are underestimated in brightness, and blues and greens are
overestimated. And (b) it will be shown that flicker itself,
the phenomenon on which the equalization at the photo-
metric screen is based, is subject to variations depending upon
a number of factors the effect of which has not in all cases been
adequately studied and in some cases not even recognized.
An investigation of one of these alone, the effect of varying
1 We do not mean to draw too close an analogy here between the effect on the
brightness of sensation produced by keeping the intensity of light constant and reduc-
ing the time of exposure of the eye to the light, and the effect produced by keeping the
time of exposure of the eye constant and reducing the intensity of the light employed
(the Purkinje phenomenon). In attempting to interpret the effect produced by the
short exposures used in the method of flicker, our data should be taken primarily
from the results showing the relative rise of sensation to its maximum for white light
and lights of the different colors. (See discussion of the development time of sensa-
tion, pp. 118-130). It is quite possible and in fact quite probable from Broca and
Sulzer's results, for example, that for a part of the upward course blue and green
rise faster than red, and conversely for a part of the course red rises faster than blue
and green. (Yellow was not used by Broca and Sulzer.) The results of Broca and
Sulzer are cited on this point, not by any means because their method of making
the determination is the freest from criticism of any that have yet been used, but
because they alone have attempted to plot the comparative curves for the different
colors and white light at different points from the threshold to the maximum.
114 C. E. FERREE AND GERTRUDE RAND
the ratio of the time of exposure of the eye to the lights to be
compared, is enough to lead one seriously to question whether
the method of flicker can be safely used in the work of hetero-
chromatic photometry, at least not without calibration, and
perhaps not without an amount of calibration which is in
itself prohibitive of the use of the method in practical work.
The third point will be covered in the following way. (i) It
will be pointed out that the only method that has thus far
been used as a standard with which to compare the method
of flicker has been the equality of brightness method. The
selection of this method as a standard has been recommended
among others by Whitman, Wilde, and Schenck, and a com-
parison of the results of the two methods, more or less com-
plete, has been made by a number of experimenters. And (2)
it will be shown both from our own work and from a very great
preponderance of the work done by others who have made the
comparison, that the results by the method of flicker do not
agree in the average with those obtained by the equality of
brightness method; and, therefore, that justification for the
adoption of the method of flicker can not yet, at least, be
fairly claimed through its agreement in result with the
equality of brightness method.
The Equality of Brightness Method. — With regard ,to sensi-
tivity in the photometry of lights of different color, the
equality of brightness method has the following disadvan-
tages, (i) Small differences in luminosity can not be de-
tected because the actual difference present is masked by
the difference in color quality. (2) Results for a given
observer can not be reproduced within a small limit of
variation, because the ability to do this in turn presupposes
the ability to detect small differences which, as has just been
stated, can not be done. (3) Results can not be reproduced
from observer to observer within a small limit of variation
because (a) the sensitivity to color varies more among ob-
servers than does, for example, the sensitivity to brightness,
hence there is a variable amount of the disturbing factor of
color present for different observers; and (b) because the
standard or pattern for the judgment of equality differs
FLICKER PHOTOMETRY 115
more from individual to individual when the factor of color is
present than when it is not. That is, in any photometric
judgment the observer must decide for himself what he will
call equality and make all his judgments conform to this
pattern or standard. When color is present to interfere with
the judgment of equality, the selection of this standard varies
more for different observers than it does when no color is
present. With regard to all the points on which sensitivity
depends, therefore, the equality of brightness method may
be said to possess a low degree of sensitivity.
The Method of Flicker. — The method of flicker possesses
greater sensitivity than the equality of brightness method.
That is, smaller differences in the luminosity of the photo-
metric surfaces can be detected, and the judgment of equality
is surer and more reproducible.1 This is because the disturb-
ing factor of color difference in the impressions to be compared
is eliminated from the judgment. That is, instead of being
given simultaneously, the stimuli are given in succession and
at such a rate that all color differences between them disap-
pear, and the brightness impressions are permitted to develop
in sensation unobscured by differences in color quality.
The use of the phenomenon of flicker to detect a difference
in brightness between two illuminated surfaces can best be
understood possibly by considering the phenomena that take
place when successive impressions of colored and colorless
light are made upon the retina at different rates of speed.
When the retina is exposed successively to colorless lights
differing in brightness, the following phenomena take place.
When the rate of succession is low, the impressions remain
1 This higher degree of reproducibility can be claimed perhaps only for the judg-
ments given by a single observer. It does not seem to obtain to any considerable
extent, so far as results are available for comparison, when results are compared from
observer to observer. For example, in a group of eighteen observers Ives gets differences
as great as 159 per cent, for .481 /i, 114 per cent, for .498 /*, 26 per cent, for .518/1.
18 per cent, for .537*1; 13 per cent, for .556/4; 10 per cent, for .576/0; 28 per cent, for
•595 M, 65 per cent, for .615 /i; 86 per cent, for .635 /i; and 122 per cent, for .655 p.
The percentage of average variation from the mean for these observers is 17 per cent,
for .481 /i; 134 per cent, for .498 /i; 6 per cent, for .518 /t; 3 per cent, for .537/4; 2.75
per cent, for .556/1; 2.2 per cent, for .576/1; 5.4 per cent, for .595 //; 9.5 per cent, for
.615/1; 13-2 per cent, for .635/1; and 19.3 per cent, for .655/1. (Philos. Mag., 1912,
24, Ser. 6, pp. 853-863.)
n6 C. E. FERREE AND GERTRUDE RAND
more or less separate and distinct. At rates higher than thisr
we have in order Fechner's colors,1 flicker, and the fusion of
the two impressions into a uniform gray. When the eye is
exposed successively to colored and colorless light, the follow-
ing phenomena take place. At low rates, we have again the
more or less separate successions of the two impressions. At
rates slightly higher than this, we have first a phenomenon
that may be called by analogy color flicker, and then an inter-
mingling of color and brightness flicker. At still higher rates
we have color fusion, brightness flicker, and complete color
and brightness fusion. Thus, both in case of colored and
colorless light, brightness flicker seems to be a phenomenon
due solely to the succession at certain rates of speed of im-
pressions differing in luminosity or brightness. Moreover,,
the phenomenon is very sensitive to changes in the luminosity
of the successive impressions. That is, a very slight change
in one of the impressions will produce flicker when there is no
flicker, or will cause a noticeable change in its amount when
there is flicker. Flicker thus becomes a very sensitive means
of detecting brightness difference. This sensitivity, however,,
is not so great in case of colored as it is in case of colorless
light. It would in fact in all probability be very low were it
not for the fortunate fact that color fusion takes place at a
very much lower rate of succession than brightness fusion.
Concerning the ease and sureness of making the judgment,,
then, the case with regard to the method of flicker may be
summed up as follows: By giving the impressions to be
compared to the retina successively at a certain rate of speed,
the disturbing element of color difference, which so interferes
with the detection of brightness difference when the im-
pressions are given simultaneously, is eliminated, and the
phenomenon of brightness flicker stands out clearly in a field
uniform as to color quality. That is, by using a method of
successive impressions we have succeeded in eliminating the
1 Fechner's colors are best observed when the successions are made by rotating
discs made up of white and black sectors, or by discs specially constructed for the
purpose. This phenomenon occurs at a rate of speed near the upper limit required
to give separate impressions, and consists of impressions of color mingled with the
more or less separate impressions given by the white and black sectors.
FLICKER PHOTOMETRY n7
feature that renders the comparison of the brightness of the
simultaneous impressions so difficult to make, namely, the
difference in color quality between the impressions to be
compared. The judgment, then, is easy, and the principle
on which the equalization is based seems to be clear. The
method has come to have many supporters, but other things
besides the sureness of judgment must be taken into account.
This brings us to a consideration of our second point, namely,
the method of flicker when applied to the photometry of
lights of different color does not seem to possess the sureness
of principle needed to meet the requirements of a satisfactory
method. We have two reasons for making this assertion.
In the first place, as we have already stated, at the rate of
speed at which the impressions are given in the method of
flicker, the eye is very much underexposed to its stimulus.
And in the second place, flicker, the phenomenon on which
the equalization is based at the photometric surface, is subject
to many variations depending upon a number of factors the
bearing of which on the application of the phenomenon to
photometry, has not in all cases been adequately studied, and
in some cases not even recognized. A few of these may be
suggested in passing, (i) The intensity of illumination and
the influence it exerts on the speed of alternation that has to be
used in order to give the method maximum sensitivity.
(2) The different rates of speed required for the fusion of the
different colors, and the varying lower limit this difference
puts upon the rates of speed that can be used. (3) The effect
of the saturation of the colors used on the fusion rate. (4)
The effect of field size. (5) The effect of the ratio of the time
of exposure of the eye to the lights to be compared; etc. A
better knowledge than we now have of the effect of these
factors is, the writers believe, of fundamental importance in
the employment of the phenomenon of flicker in the photom-
etry of lights of different colors. At a later date they hope
to report the results of a systematic study of these factors.
In the present paper, the effect of only one of them will be
considered, namely, the ratio of the time of exposure of the
eye to the lights to be compared. An investigation of this
uS c. E. FERREE AND GERTRUDE RAND
point alone is enough to lead one seriously to question whether
the method of flicker can safely be used in the work of hetero-
chromatic photometry, at least not without calibration, and
perhaps not without an amount of calibration which is in
itself prohibitive of the use of the method in practical work.
THE ACTION OF LIGHT ON THE EYE UNDER THE CONDITIONS
IMPOSED BY THE METHOD OF FLICKER.
Both of the above points will probably be more easily
understood if a brief consideration is given to the way the
eye responds to colored and colorless lights when the im-
pressions are given to it in the manner they are given in
the method of flicker. The eye is not an ideal sense organ,
that is, it does not respond at once with its full intensity of
sensation at the beginning of stimulation, nor does the sensa-
tion cease with the cessation of stimulation. It takes, for
example, an interval of time for the sensation proper to a
given stimulus to rise to its maximum; and also an interval
to die away after the stimulation has ceased, depending for
its length upon several factors.1
The interval of time required for a sensation to rise to its
maximum will be called in this paper the development time of
sensation. Plateau in 18342 first expressed the belief that
1 There are two phases to this after-effect, positive and negative. The positive
alone concerns us here. In this phase, which is often called the persistence of vision,
the original sensation tends to persist in its original color and brightness. More
accurately described, however, it rapidly loses in color and rapidly darkens. In the
negative phase there is a brightness reversal, that is, what is light in the original
sensation becomes dark, and the color changes to the complementary color. The
negative phase is much longer than the positive. The length of the positive depends
upon many factors: the intensity of the stimulus, the time of exposure of the eye
to the stimulus, the state of adaptation of the eye, the general illumination of
the field of vision, the brightness of the local preexposure and post-exposure, etc.
Unless the eye is put under very especial conditions of stimulation, the duration of the
positive phase is very short indeed, in fact, momentary. For a further discussion of
this point with reference to the method of flicker, see appendix.
2 As early as the time of Bacon it was noted that there is a period of inertia in
vision. ("At in visu (cujus actio est pernicissima) liquet etiam requiri ad eum
actuandum momenta certa temporis: idque probatur ex iis, quae propter motus
velocitatem non cernunter; ut ex latione pilae ex sclopeto. Velocior enim est praeter-
volatio pilae, quam impressio speciei ejus quae deferri poterat ad visum"."— 'Novum
Organum,' lib. II., Aph. XLVI.). Later Beudant ('Essai d'un Cours Elementaire
FLICKER PHOTOMETRY 119
color sensation does not come at once to its maximum. He,
and later Fick1 in 1863, showed that when a sector of white
paper passes very rapidly only once before the eye it looks
to be a dark gray. With the experiments of Exner in 1868,
the work of determining the development time of visual
sensation was definitely begun. Different methods of making
the determination have been used by different investigators,
and different results have been obtained. There is, however,
among the different results a certain amount of agreement.
At least the order of magnitude of the development time can
be fixed within certain limits. The chief points of interest
in these investigations have been (i) to compare the develop-
ment time of the different sensations of color with each other
and with that of colorless sensation; and (2) to determine
whether the intensity of the stimulus has any effect upon
the development time. All who have made the comparison
have found that each of the color sensations has a development
time different from the colorless sensations; all with the excep-
tion of Diirr and Berliner, that each of the colors has a different
development time; and all with the exception of Diirr, that an
increase of intensity shortens to some extent the development
time of all sensations.
A table (Table I.) has been prepared showing the develop-
ment time obtained by each of these men for the different
et General des Sciences Physique: Partie Physique,' p. 489, 3 me edition) also stated
that an object which moves with extreme rapidity before the eye is not perceived
because impressions are not made on the eye instantly. Plateau (' Nouveaux Memoirs
de 1'Academie Royale des Sciences et Belles Lettres de Bruxelles,' 1834, 8, p. 53)
made the observation that when a bit of white paper passes very rapidly before the
eye, it appears not white but gray. He was the first to express the belief that color
sensation also does not come at once to its maximum of intensity. Swan (Trans.
Roy. Soc. Edinb., 1849, 16, pp. 581-603) observed that the "light of the sky seen
immediately over a ball in its descent through the air, seemed less bright than at those
parts of the retina where the action of light had not been interrupted by the passage
of the dark body"; and conducted some experiments to determine the intensity of
light sensation with short exposures. Exposing the eye to lights of different intensities
for intervals ranging from i/ioo to 1/16 of a second, colorless 'lights of different
intensity produce like portions of their total effect on the eye in equal times.' While
he does not directly determine the interval required for the light sensation to come to
its maximum, he estimates it from the results of his experiments with short exposures
to be about i/io of a second.
:Fick, A. Archiv fur Anatomic und Physiologie, 1863, p. 739.
120 C. E. FERREE AND GERTRUDE RAND
TABLE I
SHOWING A COMPARISON OF THE DEVELOPMENT TIME OF VISUAL SENSATION WITH
THE AVERAGE TIME OF EXPOSURE OF THE EYE TO ITS STIMULUS USED IN OUR
EXPERIMENTS WITH THE METHOD OF FLICKER.
Sensation
Development Time
Average Time of
Exposure of Color by
Method of Flicker
Exner1
1868
White
5 intensities:
.0178 sec., when
Kunkel2
1874
Different colors
.118-2.87 sec-
.05 7-. 133
the value of the
colored sector was
180°.
Charpentier*
1887
1896
White,
White,
5 intensities
.OI4-.049
5 intensities
Durr5
1902
White,
.090-. 148
2 intensities:
N^artius*
IQO2
Different colors
White
.266
,;54i .
6 intensities:
.0213 sec., when
Broca and Sulzer7
McDougall.8 ..
1903
IQO4.
Different colors
White
Different colors
White
.OI3-.093
.020-.090
8 intensities:
.03 1 -.125
.07 -.125
12 intensities:
the value of the
colored sector ran-
ged from 45°-3i5°-
Biichner9
1006
Different colors
White
.049-. 2
.100-. 108
3 intensities:
Berliner10
I QO7
Different colors
.033-.230
.no
*Exner, S., 'Ueber die zu einer Gesichtswahrnehmung nothige Zeit.,' Sitzungs-
berichte der Kaiserlichen Akademie der Wissenschaften, Math.-Phys. Classe, 1868,
58, pp. 601-632.
2 Kunkel, A., 'Ueber die Abhangigkeit der Farbenempfindung von der Zeit/
P finger's Arckiv, 1874, p, p. 197.
3 Charpentier, 'Sur la periode d'addition des impressions illuminismes,' Corn-pies
Rendus Societe de Biologie, 1887, 4, pp. 192-194.
4 Lough, 'The Relations of Intensity to Duration of Stimulation in Our Sensations
of Light,' PSYCHOLOGICAL REVIEW, 1896, j, pp. 484-492.
6 Durr, E., 'Ueber das Ansteigen der Netzhauterregung,' Philosophische Studien,
1901-1903, 18, pp. 215-273.
6 Martius, G., 'Ueber die Dauer der Lichtempfindungen,' Beitrdge zur Psychologie
und Philosophie, Leipzig, 1902, I, Heft 3.
7 Broca, A., and Sulzer, D., Comptes Rendus der Seances de VAcademie des Sciences,
1902, 134, pp. 831-834; 1903, 137, pp. 944-946; 977-979; and 1046-1049.
•McDougall, W., 'The Variation of the Intensity of Visual Sensation with the
Duration of the Stimulus,' British Journal of Psychology, 1904-1905, I, pp. 151-189.
9 Biichner, M., 'Ueber das Ansteigen der Helligkeitserregung,' Psychologische
Studien, 1906-1907, 2, pp. 1-29.
FLICKER PHOTOMETRY 121
colored and colorless sensations; and, for comparative pur-
poses, the average exposure time that was used in our experi-
ments for all the colors in the determination of their bright-
ness by the method of flicker. In choosing this time of ex-
posure for the method of flicker, in order to secure for the
method the greatest possible sensitivity, we used the slowest
rate of succession of colored and colorless sectors that could
be employed.
An inspection of this table will show that while the results
for the development time of sensation differ quite a little
among themselves, they agree in one very important partic-
ular, namely, they are all much greater than are the intervals
that are used in the longest exposures that are permissible
by the method of flicker. That is, by the method of flicker,
the eye is very much underexposed to its stimulus. The
effect of this under exposure is obviously to cause a reduction
in the intensity of sensation. That is, the rate of succession
of impressions used in the method of flicker is too fast for the
single impressions to arouse their maximum effect in sensa-
tion and too slow for the successive impressions to add or
summate as much as they would need to do to cause the in-
tensity of the sensation aroused by each light to rise to its
full value, or perhaps even to rise to a higher value than
would be given by the individual exposures. In fact as will
be shown in an appendix to this paper the sensation can
not be expected to rise to its full value through summa-
tion if the Talbot-Plateau law be true, however rapid is
the rate of succession of the individual impressions (see ap-
pendix). Even when a rate is reached at which complete
fusion takes place, both for the color and brightness com-
ponents in sensation, there is according to the Talbot-
10 Berliner, 'Der Ansteig der reinen Farbenerregung im Sehorgan,' Psychologische
Studien, 1907, 5, pp. 91-155.
W. Swan in an article entitled 'On the Gradual Production of Luminous Impres-
sions on the Eye; Part II., being a description of an instrument for producing isolated
luminous impressions on the eye of extremely short duration, and for measuring their
intensity,' Trans. Roy. Soc. Edinb., 1861, 2, pp. 33-40, has described a very ingenious
but complicated apparatus for getting short periods of stimulation of the retina, but
apparently neither he nor any one else has ever used the apparatus described.
122 C. E. FERREE AND GERTRUDE RAND
Plateau law, a reduction in the intensity of each sensation
which is the same as would be gotten were the intensity of
each light to be reduced in proportion to the time of ex-
posure of that light to the total time of exposure of both
lights, and no further increase in the rate of succession pro-
duces any change in the effect.1 The possibility then of the
sensations which, as is shown by the work on development
times, are unequal for the single exposures used in the
method of flicker, reaching equality by rising to their full
value seems to be ruled out. In terms of the Talbot-Plateau
law they could not reach their full intensity through an effect
of summation, however fast the rate of succession be made,
1 Ewald ("Versuche zur Analyse der Licht- und Farbenreaktionen eines Wirbel-
losen " (Daphnia pulex), Ztschr.f. Psychol. u. Physiol. d. Sinnes., 1914, 48, pp. 285-325;
and "The Applicability of the Photochemical Energy Law to Light Reactions in
Animals," Science, 1913, 38, pp. 236-238) has made an interesting contribution with
regard to the effect of the intermittent action of light on the eye which it may not be
out of place to mention here. The facetted eye of the daphnia was used in his experi-
ments. When exposed to light this eye responds by turning towards the light, and
when lights of different intensities are used it turns towards the stronger light. After
having determined the sensitivity of this response to difference in intensity of light by
exposing the eye to a number of lights of different intensities acting continuously on
the eye, he undertook to make a comparison of the effect of light acting continuously
and intermittently. The intermittence was gotten by rotating a sectored disc in front
of one of the lights. The lights were so chosen that the same amount of energy acted
upon the eye in a given unit of time from both the continuous and intermittent sources.
That is if a ratio of total open to closed sector of the value l/io was used, the light in
front of which these sectors were rotated was made ten times as intense as the light
acting continuously. The sectored disc was then rotated at different speeds. When
a speed of 30 revolutions per second was attained the eye remained stationary. That
is at this speed of rotation the two lights produced equal effects on the eye, — which is,
of course, no more than a demonstration of the Talbot-Plateau law for the primitive
eye. But when the speed was made slower than this, the eye invariably turned towards
the light which was acting continuously. That is when the rate at which the im-
pressions were given to the eye was made slower the result was to weaken the effect
on the eye even though the same amount of light was received by the eye in a unit of
time in both cases. Ewald's results show then that, so far as the primitive eye is
concerned, when light impressions are given to the eye at certain high rates of succession
(analogous to the fusion rates for the human eye) there is a reduction in the amount of
response aroused which is the same as would be produced were the intensity of the
light reduced by an amount proportional to the ratio of the time of exposure to the
light to the total time of the observation; and when they are given to the eye at rates
slower than these the effect on the eye is the same as if the light acting on it had been
still further reduced in intensity.
FLICKER PHOTOMETRY 123
let alone attain it at the rates which are employed in the
method of flicker.1
1 There seems to be only one other possibility that the method of flicker should
give the true photometric balance between lights of different color values, namely, that
the sensations aroused should reach equality at some value lower than the full value.
That this is extremely improbable is shown by the following consideration. The
weaker sensation or the sensation which has the slower rate of development for a single
exposure would have to rise in value because of summation effect resulting from the
succession of exposures until it became equal to the stronger sensation. To produce
any effect of summation each individual impression would have to last over in sensation
until the next impression of its kind is received which, since the impressions alternate,
would be the next impression but one. And to produce the particular effect required
here, not only would each excitation have to last over until the next one is aroused, but
the weaker one would have to last over more strongly than the stronger one, else the
effect of the summation would not be to produce the gain of the weaker on the stronger
which is required to bring the two to the true photometric balance. That is, the
advocate of this point of view would say that even though for the single exposure one
color is weighted more than the other, the effect of this is obliterated in a succession of
impressions and the two rise to equal value, because the weaker sensation would carry
over more strongly hence would gain more relatively in the process of summating than
would the stronger sensation. This is not at all in accord with the experimental evi-
dence available at this time on the relation of the positive after-effect or persistence of
sensation to the original sensation. Goldschmidt ("Quantitive Untersuchungen iiber
positive Nachbilder," Psych. Studien, 1910, 6, pp. 159-252) and others show, for ex-
ample, that the stronger the original sensation, the more strongly does it tend to
carry over after the light is cut off. Goldschmidt also concludes from his experiments,
which is a very important point for this discussion, that the tendency of the sensation
to carry over is, so far as its brightness is concerned, independent of the color. That
is, suppose that a photometric balance was obtained for green and red lights of com-
paratively high intensities by the method of flicker. Then according to Broca and
Sulzer's curves, also the results obtained in this laboratory, green would attain to a
higher brightness value for the single exposure than would be attained by red. Hence if
green is not to be overestimated by the method of flicker, red must carry over more
strongly as the impressions succeed each other than does green, and thus make up by
a summation effect the deficiency shown in the single exposure. But according to
Goldschmidt's results this greater tendency to carry over could not be assumed for
red, either because of its color value or because of its weaker intensity, and there is
no other aspect of the sensation which could have any bearing on the question in hand.
Moreover, this hypothesis is rendered still more untenable by the experimental fact
that the situation at low intensities is reversed. That is, at low intensities red, as
shown by the curves for difference in lag (see Fig. 2, p. 127), attains to a higher
value than green for the single exposure. Then if red is not to be overestimated
by the method of flicker and in direct proportion to the values given to the two sensa-
tions in the single exposure, green must be carried over more strongly in the succession
of impressions than is red. The explanation of both of these points would require
not only that the color value of the stimulus exerts an influence on the carrying over
of the brightness aspect of the sensation, but that this influence reverses in passing
from high to low intensities. For a discussion of how highly improbable it is for the
rates of succession used in the method of flicker that one impression could last over
until the next impression but one is received in any amount that could be of con-
siderable consequence to the method, see appendix.
124 C. E. FERREE AND GERTRUDE RAND
It seems fair to conclude, then, that instead of getting by
the method of flicker the sensations that should be aroused by
the lights with which we are working, we get sensations of lower
intensity. But it may be asked what if there is a reduction of
the intensity of the impressions received? Equalization is all
we are working for and the intensity of both impressions is
reduced. Is it not possible, therefore, to find a ratio of time of
exposure to each light such that the amount of reduction in
the intensity of both impressions will be equal ? This would be
comparatively simple if the rate of development for all the
colored were the same as for all the colorless sensations. The
intervals of exposure could be made equal as is ordinarily done
when sectored discs are used and as apparently must be done
when the exposures are given by means of a rotating prism.
But the development time for color sensation is not the same
as for colorless sensation, and, moreover, the consensus of evi-
dence is that the rate of development is not the same for any
two of the color sensations. Thus from the standpoint of the
unequal reduction in intensities produced by the method of
flicker, the task of selecting a proper ratio of exposure time of
colored to colorless light, in case of the different colors, is one
that requires a great deal of accurate knowledge if the method
is to have the sureness of principle needed, — more, the writers
think, than we now possess.1
1 One scarcely needs point out in this regard that there is apparently no point
in the intensity scale for which a given reduction in intensity for colored light gives
the same change in luminosity in sensation that it does for white light. Beginning
with the spectrum of fully saturated colors and comparing the effect of reduction by
equal amounts of colored and white lights equal in photometric value, the blues and
greens are found not to decrease in luminosity so fast as the white light, and the reds
and yellows are found to decrease faster. Or as the phenomenon is ordinarily ex-
pressed, there is a relative lightening of the blues and greens and a relative darkening
of the reds and yellows. Nor is the phenomenon of unequal change confined to the
lower intensities. It is more striking for these intensities, but it occurs also for the
higher intensities. This conclusion is drawn from the statement made by several
writers that beginning with the spectrum of fully saturated colors and increasing the
intensity of light, all the colors are found to tend towards white, and in so doing to
change their luminosities at different rates. (For example, see Helmholtz, H., 'Ueber
Hrn. D. Brewster's neue Analyse des Sonnenlichts,' Pogg. Ann., 1852, 86, p. 520; also
Handbuch der physiologischen Optik, zw. Aufl., 1896, pp. 465-466; Chodln, A., 'Ueber
die Abhangigkeit der Farrbenempfindungen von der Lichtstarke,' Sammlung physio-
logischer Abhandlungen von Preyer, 1877, I, p. 33 if.; Briicke, E., 'Ueber einige
FLICKER PHOTOMETRY 125
We have discussed here, moreover, the effect of underex-
posure at only one rate of rotation of the exposure apparatus.
The situation becomes still more complicated when this rate
is changed. If it were changed, as it must be to preserve the
sensitivity of the method in passing from high to low illu-
mination, the whole scale of magnitude of the underexposure
would change, and a shift in the relative evaluation of the
luminosities of the different colors might very well be ex-
pected from the shape of the sensation curves as they rise to
their maximum. In fact this shift is found in the work of
previous investigators1 who have made the comparison at
Empfindungen im Gebiete der Sehnerven,' Sitzungsber. der Wiener Akademie, Math.-
Natur. Klasse, 1878, 77, Abth. 3, p. 63.) As we have already stated, however (p. 113),
we do not mean to draw too close an analogy here between the effect on the bright-
ness of sensation produced by keeping the intensity of light constant and reducing
the time of exposure of the eye to the light, and the effect of keeping the time of
exposure of the eye to the light constant and reducing the intensity. The degree to
which the analogy holds can scarcely be considered as fixed until more work is done
showing the way in which the luminosity curves for the different colors rise to their
maximum as the time of exposure of the eye to the different colored lights is increased.
1 See, for example, the phenomenon called by Ives the "reverse Purkinje" effect
(Philos. Mag., 1912, 24, Ser. 6, pp. 170-173); later demonstrated and discussed by
Luckiesch (Electrical World, March 22, 1913, p. 620). These writers have found that
the red end of the spectrum shows a relatively higher luminosity value as compared
with the green end by the method of flicker at low than at high illuminations. From
the shape of Broca and Sulzer's curves for the rise of visual sensation, to its maximum,
for example, this result might very well be due to the difference in the relative lumin-
osity value of the colors caused by the difference in the length of exposure given to the
eye in the method of flicker at the faster rates of speed required for the higher illumina-
tions and at the slower rates for lower illuminations. That is, the longer exposures
given by the slower speeds of rotation allow the colors to attain a higher intensity.
For example, the speeds used by Ives for what he calls 10 Illumination Units range,
for the different colors for five observers, from 7 to 10 cycles per second, and for 250
Illumination Units from 10 to 22 cycles per second.
Broca and Sulzer's curves (Fig. i) are appended here for one order of intensity
of stimulus (Comptfs Rendus, 1903, 137, p. 978. For other intensities see p. 945).
The curves given were selected because they alone show a comparison between the
results for colored and white light. It will be seen from these curves that for exposures
less than .07 sec. (approximate value), blue and green rise to a higher value than red;
for exposures ranging from .07 to .11 sec., blue rises to a higher value than red, and
red higher than green; and for exposures ranging from .11 sec. to about .25 sec., red
rises to a higher value than blue or green.
There is also a very strong probability that the relative lag in sensation for the
different wave-lengths is not the same for lights of low intensity as for lights of higher
intensities. In fact the results that have been obtained so far in this laboratory in
determining the development time of the sensations aroused by red, yellow, green, and
126 C. E. FERREE AND GERTRUDE RAND
intensities low enough to necessitate a decided reduction in
speed of rotation. And that the shift is different from the
normal effect on the brightness of the colors produced by a
decrease of illumination, is shown by the fact that according
to the results of these investigators it is in a different direction
from that given by the equality of brightness method.
Moreover, in the later paper it will be shown that a change in
this evaluation amounting to several times the smallest dif-
ference in luminosity that can be detected by the method, is
also produced by working the reverse variation, that is, by
keeping the rate of rotation constant and changing the ratio
of value of colored to colorless sector. It is difficult, there-
fore, to avoid the conclusion that the type of exposure used in
the method of flicker is an important factor in the cause of its
disagreement with the results obtained by other methods.
blue lights of spectrum purity show that at low intensities red and yellow rise more
rapidly in photometric value than green and blue. This result is quite marked in those
parts of the curves representing an exposure time of the same order of magnitude as
is used in the method of flicker. In order to show this point we have appended here
too.
50.
FIG. i
three curves representing the relative rates of development of red, yellow, green, and
blue at intensities which we will designate for the present as low and intermediate; and
of red, yellow, and green for a higher intensity. These determinations were made by
FLICKER PHOTOMETRY
127
Reexamining the case, then, with regard to the underex-
posure of the eye by the method of flicker, we find that the
short exposure times necessary to the method cause a re-
M. A. Bills of this laboratory. Later, results will be given for red, green, blue, and
yellow at a number of intensities, and specifications will be made of the intensities
employed in both photometric and radiometric units. The colored lights used in
determining the curves given below were obtained from a spectrum of good definition
and were in each case equal in photometric value, as they should be if results are to be
used in interpreting the action of light on the eye under the conditions imposed by the
\
0-05-
0.1
FIG. 2
method of flicker when the photometric balance is attained. In constructing these
curves time of exposure is plotted along the abscissa and brightness of color along the
ordinate. The curves in Fig. 2 are for the low intensity; in Fig. 3 for the intermediate
intensity; and in Fig. 4 for the higher intensity.
It is very probable that there is considerable individual difference in the amount
and distribution of lag. A rigid test of the correspondence of difference in lag to the
direction of deviation of results gotten by the method of flicker from those obtained by
the equality of brightness method would require that the photometric determinations
and the determination of lag should be made for a given quality and intensity of light
by the same observer.
If it should be found that there is an individual difference in the amount and dis-
tribution of lag, the result would supplement very nicely the explanation why a much
higher degree of reproducibility is gotten by the method of flicker than by the method
128 C. E. FERREE AND GERTRUDE RAND
duction in the action of the standard and comparison lights
on the eye. If this reduction were equal in amount, quite
enough difficulty would be encountered. But it is not
of equality of brightness only when the results of a single observer are considered (see
footnote, p. 115). That is since the factor of color difference which so disturbs the
judgment in the equality of brightness method is eliminated in the method of flicker,
we should get correspondingly a higher degree of reproducibility for a single observer
and for different observers, were there not some factor present in the method of flicker
and not in the equality of brightness method, which varies from individual to individual.
So supplemented the explanation would be as follows. In the method of flicker the
6.oy 0-< o-if o. a 0.25-
FIG. 3
judgment for a single observer shows a higher degree of reproducibility than in the
equality of brightness method because of the elimination of the disturbing factor o
color difference; but a false balance is established by the method, the deviation from
the true balance depending in direction and amount for different observers upon the
difference in the amount and distribution of lag in the rise of the sensations towards the
maximum. The difference in the amount and direction of this deviation from the true
balance from observer to observer is the cause of the relatively low degree of reproduci-
bility of results when the work of different observers is compared.
f / I
.• •' /
fit
/!:
f i '
C ' / (
\ //
lit
944
I , <
. ! f
•II
A / (
tl?
Ill
i
f !'
X v
FIG. 4
a 0,0
130 C. E. FERREE AND GERTRUDE RAND
equal in amount, and we have no adequate information as to
the amount of the inequality. Until we have more informa-
tion with regard to the amount of the reduction and the
effect it produces, any successful attempt to regulate the
relative duration of the exposure to colored and colorless light,
even for a single color at a single intensity, can scarcely be
more than the result of chance. Surely to expect to accom-
plish this for all intensities of all colors by a single ratio of
exposure, more especially by means of a I :i ratio, as has been
the practice in the past, is, it would seem to the writers, to
ignore the sensation principles which underlie the method.
Obviously this ratio requires calibration, and to give the
method the sureness of principle required, it would seem that
the calibration might have to be made for each color at the
particular intensity at which the work is to be done. This
calibration might possibly be accomplished by means of an
accurate knowledge, if that knowledge could be secured in
sufficient detail, of the temporal course of visual sensation as
it rises to its maximum for the given intensity of light used; or
it might be done by comparing the results obtained by the
method of flicker with the results obtained by some other
method adopted as a standard. So far the tendency seems
to have been to look for this standard within the subject of
photometry itself. As has already been stated, several
writers have signified a desire to make the equality of bright-
ness method a standard, and comparisons have been made of
the results obtained by the method of flicker and the equality
of brightness method.
A consideration of the results of these comparisons together
with the data collected by one of the writers in ten years use of
the flicker and equality of brightness methods in making the
brightness matches needed for the work in color sensitivity, has
influenced us to open the question anew in the interests of
the work on color sensitivity. In the work with these two
methods all done at comparatively high illuminations and
with a large number of observers, agreement has been rare.
For this reason the chief incentive to make the present study
has not been to establish disagreement, but to investigate
FLICKER PHOTOMETRY I31
further the causes of disagreement. The results of this study
seem to indicate that the type of exposure of the eye to its
stimulus by the method of flicker is an important cause of
disagreement.
So much for the theoretical considerations relating to the
method of flicker. To test the accuracy of some of the more
important points that have come up, a plan of experimentation
has been formulated and in part carried out. So far as the
results of that experimentation will be reported upon in this
paper, the following things will be shown, (i) By comparing
the time required for the sensations aroused by colored and
colorless light to reach their maximum of intensity, we have
already shown that as a general case the eye is very much
underexposed to its stimulus by the method of flicker, and
we have concluded that the effect of this underexposure
on the brightness component of sensation will be unequal
in amount for colored and colorless light, and should lead,
therefore, to a false estimation of the brightness of the
colors. That this conclusion is justified so far as our work is
concerned, will be demonstrated in part by comparing the
results of the method of flicker with those obtained by the
equality of brightness method, in which case the eye is fully
exposed to its stimulus, and showing that for the method of
flicker there is for our observers for the intensities of light
used and for the rate of rotation of the photometer head
required for these intensities, a characteristic underestimation
of the brightness of red and yellow and overestimation of the
brightness of blue and green. That this characteristic devia-
tion is due to the type of exposure used in the method of
flicker and not to some other factor will be further shown in
the consideration of our second point. (2) We have said
that the ratio of the time of exposure should be considered
as a factor influencing the results obtained by the method of
flicker. In order to confirm this judgment of the case, we
have varied this ratio, keeping the other conditions constant,
and have found that a corresponding variation is produced
in the results. That is, by changing the value of the colored
and colorless sectors in the rotating disc we have used to
132 C. E. FERREE AND GERTRUDE RAND
regulate the time of exposure in the method of flicker, corre-
sponding variations are obtained in the characteristic under-
estimations of red and yellow and the overestimations of blue
and green. These variations, it will be shown, moreover,,
are very much, greater than the changes in luminosity that
are required to be detected by the method of flicker, and arey
therefore, worthy of being taken into account in an evaluation
of the usefulness of the method, whatever method be adopted
as a standard for comparison. And (3) we have contended
that if the equality of brightness method be adopted as the
standard for work in color photometry, the method of flicker
does not satisfy the requirements, for it does not give results
which agree in the average with those obtained by the equal-
ity of brightness method. This will be shown both from
results of our own work and from a preponderance of the
work done by others who have made the comparison. In
our own work the comparison has been made for a series
of intensities which may be considered as at least fairly repre-
sentative of the higher intensities, they being considered more
favorable to agreement by Dr. Ives.1 Especial care has been
taken in this series to duplicate at one point the intensity
which Dr. Ives finds the most favorable to agreement.
The remainder of the paper will be taken up with the
demonstration of these three points.
I. THE UNDERESTIMATION OF THE LUMINOSITIES OF REI>
AND YELLOW AND THE OVERESTIMATION OF THE
LUMINOSITIES OF BLUE AND GREEN
Special tables have not been prepared for this point
because the results can readily be seen in the tables for points
II. and III. In these tables taken collectively the comparison
will be shown for a representative series of variations, both
of the ratio of the time of exposure to colored and colorless
light and of the intensity of the lights employed. In every
case underestimation is found to be characteristic for red and
yellow, and overestimation for blue and green.
1 Ives, H. E., 'Studies in the Photometry of Lights of Different Colors, Philos^
Mag,, 1912, 24, Ser. 6, pp. 149-188.
FLICKER PHOTOMETRY *33
II. THE VARIATION OF THE RATIO OF THE TIME OF EXPOSURE
TO THE COLORED AND COLORLESS LIGHT CAUSES A
CORRESPONDING VARIATION IN THE CHARACTERISTIC
UNDERESTIMATION OF RED AND YELLOW AND
OVERESTIMATION OF BLUE AND GREEN
As has already been stated, the work under this heading
has been undertaken in part to show the preceding point,
and in part to show that the amount of this underestimation
and overestimation is a variable function of the ratio of the
time of exposure to the colored and the colorless light. The
effect of the variation was determined both when the com-
parison was made between colored and colorless pigment
surfaces, and between colored and colorless lights. For the
pigment surfaces the standard red, green, blue, yellow, white,
and black of the Hering series of papers were used. For the
colored lights, two sources have been used: the Wratton and
Wainwright color filters, and the light of the spectrum.
Since the work with the spectrum as source has not yet been
finished, results will be given at this point from the work with
the filters. Of these filters, only the Alpha and Eta were
used. The former transmits a band of red from the end of
the spectrum to .65 ju, the latter, a band in the blue-green
from .52 n to .465 /z. These two alone were used for the
following reasons: (i) They are fairly representative of the
colors that show a relative change in luminosity with change
of intensity. And (2) the yellow, green, and blue filters each
transmits components that undergo opposite luminosity
changes with a change of intensity of the source. That is,
the best yellow of the series transmits also a green com-
ponent; the best green, a yellow component; and the best
blue transmits some of the violet.
The photometric apparatus employed was for the sake of
comparison made to conform very closely in its essential
features to that described by Dr. Ives.1 The general plan
of our apparatus is indicated in Fig. 5. It consists of a photo-
meter bar carrying the standard white light (A), a second bar
carrying the colored light (£), a sectored disc (C), and a
1 Ives, H. E., op. cit.j p. 161.
134 C. E. FERREE AND GERTRUDE RAND
screen (D) provided with a small aperture (0) through which
the light comes to the eye. The standard white light was
enclosed in a black light-proof box (E), which was provided
FIG. 5
in front with a circular opening 4 cm. in diameter for the trans-
mission of the light. In passing to the sectored disc, the
light was screened both from the observer's eye and from the
colored source by black screens properly placed. The light
which was passed through the colored filters was placed in a
similar light-proof box (F) provided with an opening 4 cm.
square for the transmission of the light. Above and below
this opening were grooves into which the color filters were
slid. The sectored discs were made of aluminum. The edges
of these discs were carefully bevelled and the surface was kept
freshly covered with magnesium oxide deposited from the
burning metal. The aperture in the screen through which
the light passed to the observer's eye was 3 mm. square.
The visual angle subtended by this aperture at the observer's
eye at 20 cm. distance was very small. A small angle was
needed to guard against the unequal sensitivity of the central
and paracentral portions of the retina to flicker, and against
the difference in their brightness sensitivity to colored ancj
colorless light. A 13-candle-power Mazda lamp was used as
source for the colorless light, and 13-cp., 52-cp., and i3O-cp.
lamps for the colored light. These lamps were operated on a
no D.C. circuit in series with an ammeter and finely gradu-
ated rheostat to guard against fluctuations in the current and
FLICKER PHOTOMETRY ^35
loss of efficiency in the lamps. Also fresh lamps were sub-
stituted at the beginning of each series of observations. As
a check on the results obtained from these lamps, several
series of observations were made using a standardized tungsten
lamp, street series, i6.6-cp. operated at 11.43 volts by a
storage battery for the colorless light, and a similar lamp of 67-
cp. operated by a storage battery at 10.35 volts for the
source of the colored light.
The method of making the flicker judgment was as follows:
A preliminary determination was made of the approximate
setting of the light which was being moved, to give equaliza-
tion. The speed of rotation of the sectored disc was then
reduced until flicker was obtained. The position of the light
was again adjusted until no flicker was obtained, and so on.
This variation in the speed of rotation of the disc and the
position of the light was continued until the position was
ascertained that gave no flicker for the lowest speed of rota-
tion. The final determination of this point was made by
moving the light in both directions until noticeable flicker was
obtained, and taking the average of these two readings.
The movement required to give flicker on either side of this
average position ranged usually from 2 to 9 mm. depending to
some extent upon the observer and the intensity of illumina-
tion used. Employing the above apparatus and method,
results were obtained for the highest intensity of colored light
used for a total open sector of 315°, 270°, 225°, 180°, and 45°;
and for the other intensities, for a total open sector of 315°,
180°, and 45°. In making the comparisons by the equality
of brightness method, the disc was rotated until one of its
edges bisected horizontally the photometric field. The
results are shown in Tables II.-V. They will be sum-
marized briefly as follows: (i) For all values of open sector
and for all intensities of light, there was an underestimation
of the luminosity of the red light and an overestimation of the
luminosity of the blue-green. (2) As the size of the open
sector was decreased, there was a corresponding increase in
the amount of the underestimation of the luminosity of the
red for all the intensities employed, and of the overestimation
136
C. E. FERREE AND GERTRUDE RAND
5
'.9
111
to 1
111
S «^
dies
1!*
-
m
lit
3ii
• 2 IS
Is
CO
Hjfi
co ^ ^ 4*"O vo co vo vr> ^^ vovo K- ON ^06 vo t>.oo t
^
fe ii
U
vo vo vo vo vo vo vo
vo
F^ $
1* III
|
§ 5 1§J
VO VO
•<U ««
co n- co -* co co
ill
FI
Us
GIT
s
t t ^ ^?
T + r + r +
>i*O s
fitls
§
VO
1 i^-iJ J MC^
O O "-i co ON t*»
QWM|g|
1 + 1 + 1 +
III
CJ
10 vo vo vo vo 10
VO
•5 QJ>3
co vo t^. co O 00 c4 -4"O -4- O 06 M •**• "4- vovd VO t^vo"
CO CO COOO OO t^ O O O ON ONOO OOOOOOOOOO tv.t-».l^J
LO CO
r^ t^.
^
J3 iu
O c/3
ooooooooooooooooooooc
toO vovoO vovoO voioO vovoO voO vovoO vo
83
IB-S «^ll
CJ
M1 Hi
0°
vo ^ co
vo HH to ^o to r<
c^ ON ON O ^* OO
3
( g g g
. O 2i <U LH <U t-
^ 3 « ^ ^feo'S ^^
M M M pQ M PQ
gl
§|| il| s||
Ii
V04->g VO^jg OO-ug
o ^
co'XJ CU co c» -3 &. co co^ Q* co
l-l UO M
111
d
o
co
FLICKER PHOTOMETRY
137
^|
•4
to
o w o
HI b» N vo t^ O^
COON NHI OOrftO^f-
4
fl
VO 00 00 10 t^VO
^•***aa*^-
J*
II
6
gS
to
to 10 toco
oo oo ONVO 00 ONOO oo oo ONOO oo
U «
^ a
O cfl
ill
to to
•!
3&*
N to
to •* to •«*•
6
%z
.
£
> 2
5
|
*J
c* ^
00 t^ to
u
|I
•r}- 10
1 +
oo t~- vo oo
1 + 1 +
•o
•^1
^|8
•Slo.
§
3
o
||l
2*°o%
ll*J
to
to
tO HI
to "«*• CO
$
*
5«w
1 +
1 + 1 +
1 ^
"o^o
s ^
B
3 rt
4; .2? c £j
1—1 rf
o
* w
>
g «.s~
to
s g
oo
MtO NVOtOto lOtOlO
H %
1
p§3fe
^ON'O f^t^R
M &
« S
s^oo^as^oo^^R
H 0
t;
1*
HI
< M
HI HI HI CS
Q
^
Ho
w
d -
b
i
!Q
^o
^ §
<
fo
*o /5
10 O to to O to
too «oioO to«oO totoO to
•4
«»
t— ( CO "^ HI OO ~^t"
HIOO ThhHOO TJ-HICQ ^HIOO Tj-
|1
C<1 l-l tO HI
ill
&§*
Distance of
White Light
Giving Equality
of Illumination
10
vo to
oo oo
i
w
oo o^ oo oo
C!
G &
1
T! *%>
•8 1* 3 J^
Cb CQ Cb CD
c
1-
« l41'=
gg.u Ad.||.g
55
•"a s c
hH '^ 0» CO rt H-4 CD
J?£ S c "" g S c
b
d2 § S o *g ° " | w
c
1s
^.TJ & 2 OH g
^*.22^Cy rtONctf^o
N ^ CX co t^^J OO +-> CX co
to vo
t
1
12
I
c
II
Jl 1
8* VO rt &
v:
vo to -2
to vc5 w -2
HI HI
138 C. E. FERRER AND GERTRUDE RAND
of the blue-green. (3) The amount of change in the photo-
metric value of the color produced by varying the ratio of
exposure to colored and colorless light was many times the
smallest amount of change that can be detected by the method
of flicker, and, therefore, must be considered of consequence
in relation to the application of the method to practical work.
Column I of these tables represents the source of white
light; Column 2, the source of colored light; Column 3, the
color used; Column 4, the distance of the white light from the
disc when judgment of equality is given by equality of bright-
ness method; Column 5, the value of the colored sector for
the method of flicker; Column 6, the distance from the disc
at which the white light has to be placed to give the judgment
of no flicker; Column 7, the difference in the distance the
white light was placed for the equality of brightness method
and the flicker method with 180° of colored sector; Column 8,
the change in the distance of the white light produced by
varying the value of the colored sectors in the method of
flicker; Column 9, the distance the white light has to be moved
in the equality of brightness method to change the judgment
from equality to just noticeably lighter or darker; Column 10,
the distance the white light has to be moved in the flicker
method to change the judgment from no flicker to just
noticeable flicker; and Column II, the number of revolutions
per second of the sectored disc for the method of flicker.
Tables IV. and V. represent the results of Tables II. and
III. expressed in percentage of luminosity at the photometric
screen.
Pigment papers are still used in a great many laboratories
for the investigation of color sensitivity, and because of their
convenience and ease of manipulation, they probably will be
used for many years to come for preliminary work and for a
certain class of investigations in which only comparative
results are wanted. In estimating the brightness or luminos-
ity of these pigment colors, the method of flicker is now
much more extensively used perhaps than any of the other
methods of making brightness comparisons. For this reason
we have considered it worth while to extend our work to the
FLICKER PHOTOMETRY
139
TABLE IV
OBSERVER A
Showing the Results in Table II. Expressed in Percentage of Luminosity
Disagreement
Between
Equality of
Change
Amount of Change that
Can Be Detected by
Source of
White Light
Source of
Colored Light
Color
Brightness
Method and
Flicker Method
with 180°
Open Sector
Produced by
Varying
Sectors
Equality of
Brightness
Method
Flicker
Method
13 cp.
13 cp. 151 cm.
Red
-is %
- 4-8%
5 %
4%
distant from
Blue-
photometric
green
+12.3
+ 6.3
3-7
.6
screen
52 cp. 151 cm.
Red
—20
- 7-5
7
4
distant from
Blue-
photometric
green
+30
+ 13
7-3
.6
screen
130 cp. 89 cm.
Red
— 21
- 9
7
•9
distant from
Blue-
photometric
green
+ 197
+ 7-3
6.8
•7
screen
16.6 cp.
67 cp. standard
Red
-23.6
- 8-3
8.7
i
standard
lamp, 89 cm.
Blue-
lamp
distant from
green
+33
+ 8.3
8.5
.8
photometric
screen
TABLE V
OBSERVER B
13 cp.
52 cp. 151 cm.
Red
-30%
-16%
7-5%
1-7%
distant from
Blue-
photometric
green
+ 17-
+ 18
7-7
1.2
screen
16.6 cp.
67 cp. standard
Red
-28.
— 12.4
6.5
1.6
standard
lamp, 89 cm.
Blue-
lamp
distant from
green
+31-7
+24.
9
2
photometric
screen
investigation of the effect of varying the value of the colored
and the colorless sectors on the brightness of the pigment colors
as determined by the method of flicker. Of the devices
available for applying the method to these colors, the Schenck
apparatus was selected as best suited to our purpose. As
colors to be investigated, the red, green, blue, and yellow of
the Hering series of standard papers were chosen. Sectors
of the value of 180°, 270°, and 300° were used. Values lower
140 C. E. FERREE AND GERTRUDE RAND
than 1 80° were not used because they could not be accurately
obtained with the type of photometer employed. Two in-
tensities of illumination were used, one of 390 foot-candles
(vertical component) received directly under a skylight and
diffusion sash of ground glass; the other, 5 foot-candles, the
illumination of a room lighted by windows. Space will be
given here only for the results for the higher illumination.
This illumination was carefully chosen far above the range of
intensities at which the Purkinje phenomemon occurs when
the eye is fully exposed to its stimulus, in order to subject
our demonstration to a rigid test. We were seeking, for
example, to ascertain whether an intensity might not be
found so high that the underexposure of the eye to its stimulus
by the method of flicker would not cause an underestimation
of the brightness of red and yellow and an overestimation of
the brightness of blue and green.1 That these underestima-
tions and overestimations occur at this high illumination and
by amounts many times the smallest brightness difference that
can be detected by the method, will be shown in Table VI.
Column I of this table shows the color used; Column 2, the
black-white value of the color estimated by the equality of
brightness method; Column 3 gives the value of the colored
sector; Column 4, the white-black value of the color estimated
by the flicker method; Column 5 gives the difference in the
result by the equality of brightness and flicker method with
1 80° colored sector; Column 6 gives the change produced in
the result by the method of flicker by varying the size of the
colored sector; Column 7 gives the amount of change that
can be detected by the equality of brightness method; and
Column 8, by the flicker method.
III. THE METHOD OF FLICKER DOES NOT GIVE RESULTS
WHICH AGREE IN THE AVERAGE WITH THOSE OBTAINED
BY THE EQUALITY OF BRIGHTNESS METHOD
Nothing will be added in this section except to make our
comparisons at the intensity of illumination found to be most
1 We have been careful to choose high intensities because Dr. Ives has contended
that at high intensities the disagreement between the methods of flicker and equality
of brightness tends to disappear.
FUCKER PHOTOMETRY
141
TABLE VI
OBSERVER A
Showing that the Underestimation of Red and Yellow and the Over estimation of Blue and
Green is a Characteristic of the Method of Flicker for Light of the Intensity Used
in this Work, and that the Amount of this Underestimation and Over-
estimation is a Variable Function of the Ratio of the Time of
Exposure of the Eye to the Colored and the Colorless
Light.
Color
Equality of
Brightness
Method
Flicker Method
Difference by
Equality of
Brightness
Method and by
Flicker Method
with 1 80°
Colored Sector
Change Pro-
duced by
Varying
Sectors
Change that
can be Detected
by
White-black
Value
Value of
Colored
Sector
White-black
Value
Equal-
ity of
Bright-
ness
Flicker
Method
Method
Red
White 64°
300°
White 58.9°
-174°
-I2.3°
8°
1.8°
_-*
Black 296°
Black 301.1°
270°
White 56.2°
1.2°
•
Black 303.8°
180°
White 46.6°
1.8°
Black 313.4°
Yellow
White 332°
300°
White 328.3°
-18.6°
-14.9°
9-5°
2°
Black 28°
Black 31.7°
270°
White 321.7°
1.8°
Black 38.3°
1 80°
White 313.4°
Black 46.6°
1.8°
Green
White 88.5°
300°
White 99°
+26.4°
+ 15-9°
9°
45°
Black 251.5°
Black 261°
270°
White 105.5°
•9°
Black 254.5°
180°
White 114.9°
45°
Black 245.1°
Blue
White 12.5°
300°
White 14.5°
+ 10.7°
+ 8-7°
5-3°
2.2°
Black 347.5°
Black 345.5°
270°
White 19.1°
14°
Black 340.9°
1 80°
White 23.2°
•9°
Black 336.8°
favorable for agreement by Dr. Ives.1 The plan of the ap-
paratus used in this work is indicated in Fig. 6. A spectro-
scope was used to give the colored light; a 32-cp. carbon
lamp (,F) was used as the source of the colorless light. This
lamp gave a light of the same quality as that used by Dr. Ives,
namely, the quality of the carbon standard of 4.85 watts per
mean spherical candle. When placed at 32.6 cm. from the
sectored disc (£)), 270 meter candles of light were reflected
1 Ives, H. E., op. cit., p. 173.
142 C. E. FERRER AND GERTRUDE RAND
from the disc. The eye piece was removed from the spectro-
scope and a lens system was used in its place consisting of two
lenses (A) and (5), one to render the light emerging from the
objective slit (C) parallel, and the other to focus it on the eye
30 cm. distant. Between the eye and the focusing lens (E)
was interposed the sectored disc (D). Thus the light reflected
from the sectored disc suffered no absorption in passing to the
eye. A stimulus-opening (E) 16 mm. in diameter was placed
in front of the disc 20 cm. from the eye. This subtended the
FIG. 6
same visual angle as the field size that Dr. Ives found to be
the most favorable. A pupillary aperture I mm. square
placed in front of the eye reduced the light reflected from the
white disc to the intensity called by Dr. Ives 270 illumination
units.1 The colors used were a very narrow band of the
spectrum in the region of .68 /*, .57 /*, .52 /*, and .47 /*, giving
the four pure colors red, yellow, green, and blue. The method
of making the comparison was as follows: The sectored
disc was turned so that its edge bisected horizontally the
photometric field, and the luminosity of the colored field was
altered by changing the width of the collimator-slit until it
equalled by the equality of brightness method the 270 illumi-
nation units. Using this slit width, then, the disc was rotated
and the position of the white light was adjusted until no
1 By using a pupillary aperture I mm. square, Dr. Ives has reduced the light
entering the eye by an amount which, so far as we can see, can not be determined.
He has established an arbitrary unit which he calls an illumination unit. We can
not, therefore, compare the intensities of light used by us in the preceding experi-
ments (pp. 134 ff.) with the 270 illumination units used by Dr. Ives. If one were to
judge, however, by the apparent brightness of the disc in the two cases, he would have
to say that the amount of light entering the eye was considerably greater for our
FLICKER PHOTOMETRY
'43
flicker was obtained. The flicker determinations were made
with 315°, 1 80°, and 45° total open sector as before. The
results are shown in Tables VII and VIII.
TABLE VII
OBSERVER A
Showing that the Underestimation of Red and Yellow and the Over estimation of Blue and
Green is a Characteristic of the Method of Flicker for Lights of the Intensity Used
in this Work, and that the Amount of this Underestimation and Over-
estimation is a Variable Function of the Ratio of the Time of
Exposure of the Eye to the Colored and the Colorless
Light. Intensity same as was used by
Ives
Wave-
length
Equality of
Brightness
Method
Flicker Method
Difference by
Equality of
Brightness
Method and
by Flicker
Method
with 1 80°
Colored
Sector
Change
Produced
by Varying
Sectors
Amount of Change
that Can Be
Detected by
No. of
Revolu-
tions per
Second,
Flicker
Method
Distance of
White Light
Giving
Equality of
Illumination
Value of
Colored
Sector
Distance of
White Light
Giving no
Flicker
Equality
of Bright-
ness
Method
Flicker
Method
.68 /i
32.6 cm.
3IS°
41.5 cm.
— 1 1.6 cm.
—3.6 cm.
2.4 cm.
•5 cm.
9.2
1 80°
44.2
•45
12
45°
45-1
•5
II
•57 /•
32.6
315°
37-4
- 6.9
-3-4
2.8
•4
9.8
1 80°
39-5
•4
H
45°
40.8
•4
12
.52 M
32.6
315°
23
+ 13-6
+4-9
2
•5
9-7
1 80°
19
•4
13
45°
18.1
•4
11.7
•47 M
32.6
315°
23
+ I3-8
+5
2-3
•45
9-9
1 80°
18.8
•4
14.1
45°
18
•45
12
higher intensities than the 270 illumination units used by Dr. Ives. Thus it seems prob-
able that most of our preceding tests were made with an intensity of light equal to or
greater than that used by him. His claim, it will be remembered, was that one of the two
causes of disagreement between the results obtained by the methods of flicker and
equality of brightness in preceding experiments is the low intensity of the lights used.
(The other was the lack of proper regulation of the size of the photometric field.)
We do not believe that either one of these factors is the fundamental cause of disagree-
ment, as is attested in our experiments by the fact that strong disagreement remains
when both of them have been eliminated, at least, as completely as they were eliminated
by Dr. Ives. A consideration of the functioning of the eye under very short exposures
to light, shows, we believe, a much more fundamental cause of disagreement, namely,
the difference in the way in which the eye responds to light stimuli when presented
for the lengths of time used in the two methods.
144
C. E. FERREE AND GERTRUDE RAND
TABLE VIII
OBSERVER B
.68/1
32.6
3iS°
4i-3
-11.6
-4
3
.8
12
1 80°
44.2
•7
H
45°
45-3
.8
13-4
•57 M
32.6
315°
180°
38
41-5
- 8.9
-4-3
2.9
3
12.5
14.2
45°
42.3
.8
I3-I
.52 M
32.6
315°
234
+ I2.I
+3-9
3-3
.8
12
1 80°
20.5
•7
H
45°
19-5
-7
12.2
•47 /*
32.6
315°
22.8
+ 12.6
+4
3-5
•9
II. I
1 80°
2O
•7
144
4.5°
18.8
.8
12.8
It was stated in the beginning of the paper that disagree-
ment between the results of the method of flicker and equality
of brightness would be shown from a preponderance of the
work done by others who have made the comparison. As a
general case the fact scarcely needs more than the pointing
out. Before the work of Ives, disagreement was pretty
generally admitted. Bell1 says: "That the flicker and equal-
ity of brightness methods do not give coincident results when
we consider the general case of flicker photometers, as com-
pared with equality of brightness photometers, is a fact that
has been too long familiar to photometrists to admit of a dis-
cussion." Comparisons of the two methods have been made
by Whitman, Wilde, Dow, Bell, Stuhr, Luckiesh and Ives.
Whitman2 compared the luminosities of a red and green light
placed 6 ft. apart on a photometer bar. He found that the
setting of the photometer for equality of illumination differed
for the equality of brightness and flicker methods by 1.2 ft.
for one observer, and .8 ft. for another. Wilde3 photometered
a tungsten lamp against a carbon by the methods of flicker
and equality of brightness, and found a difference of 6 per
cent, in the result. Bell4 compared the ratio of lumin-
1 Bell, L., 'Acuity in Monochromatic Light,' Electrical World, Sept. 9, 1911, 58,
P- 637-
'Whitman, F. P., 'On the Photometry of Differently Colored Lights and the
Flicker Photometer/ Physical Review, 1896, 3, pp. 241-249.
•Wilde, L. W., 'The Photometry of Differently Colored Lights,' The Electrician,
July 16, 1909, 63, pp. 540-541.
4 Bell, L., 'Chromatic Aberration and Visual Acuity,' Electrical World, May II,
1911,57, pp. 1163-1166.
FLICKER PHOTOMETRY H5
osities of a mercury vapor lamp with that of a tungsten
lamp by means of the flicker method and found it to be 5.42.
These same lights by the equality of brightness method gave
a ratio ranging from 6.86 to 10.93 for different observers.
Stuhr1 compared red and green lights by several methods
including the method of flicker and equality of brightness.
He found that the mean deviation of the values obtained by
the method of flicker from those obtained by the equality of
brightness method amounted to 14.14 per cent. Luckiesh2
photometered a red against a blue-green light by the methods
of flicker and equality of brightness, and found a difference of
62 per cent, in the ratios of the luminosities of the two lights
by the two methods.
Two factors have in the main been assigned to the cause
of the disagreement: the effect of intensity and of size of the
photometric field. Lauriol, Dow, Millar, Ives, and Luckiesh
have investigated the former factor, and Schenck, Dow, and
Ives the latter. These are both factors which affect the
results of both methods. In comparison little attempt has
been made to find the factors that affect the results of each
method alone. As a general case these, it would seem, might
be more apt to prove a source of disagreement than those
which affect both methods.
With regard to the intensity of the light as a factor,
Lauriol3 and Dow4 claim that the relative shift in the bright-
ness of the different colors at low illuminations is shown by
both methods. The shift for Dow, however, is more pro-
nounced in the equality of brightness than the flicker deter-
minations. For Lauriol the shift for the different colors
varies in magnitude by the two methods and in some cases
in direction. Millar,5 on the other hand, claims that the
1 Stuhr, J., 'Ueber die Bestimmung des Aequivalenzwertes verschiedenfarbiger
Lichtquellen,' Kiel, Philos. Diss., Vol. /p, Okt., 1908, p. 50.
2 Luckiesh, M., 'Purkinje Effect and Comparison of Flicker and Equality of
Brightness Photometers,' Electrical World, March 22, 1913, p. 620.
3 Lauriol, 'Le phptometre a papillotement et la photometric heterochrome,'
Bull. Soc. Intern, des Electriciens, 1904, pp. 647-652.
4 Dow, J. S., 'Color Phenomena in Photometry,' Philos. Mag., 1906, 12, Ser. 6,
p. 131.
6 Millar, P. S., 'The Problem of Heterochromatic Photometry,' Trans. Illuminating
Engineering Society, 1909, 4, p. 769.
146 C. E. FERREE AND GERTRUDE RAND
Purkinje phenomenon is not shown at all by the flicker
method at low illuminations, while Ives1 and Luckiesh2 go
to the other extreme and declare that a reverse Purkinje
effect is obtained by the flicker method. With regard to size
of field as a factor, Schenck3 found that a decrease in size
lowered the mean variation for the flicker method and de-
creased the luminosity value obtained for all the colors.
Dow4 found that as the size of the field was decreased, red
and yellow lightened relatively to green and blue. This
effect was more pronounced for the equality of brightness
than for the flicker method. Ives5 found this effect for the
equality of brightness method, but the reverse effect for the
flicker method.
Ives, admitting the disagreement between the two-
methods and accepting size of field and intensity of the
stimulus as the cause of the disagreement, sought to determine
whether a field size and intensity could not be found for
which the two methods agree. He photometered different
portions of the spectrum against carbon lamps at a number of
intensities and with a number of field sizes. He found in
general for five observers that the luminosity curves obtained
by each method differed. This difference, however, was less
for high intensities than for low.
A table is appended (Table IX) in which is shown in per-
centage the difference in results gotten by the five observers
used by Dr. Ives at the intensity of light which he calls most
favorable to agreement for the two methods (250 Illumination
Units). It will be seen that the disagreement for these
observers is in the average as great, if not greater than was
gotten by our own observers. Percentage of overestimation
by the method of flicker is designated by +, and under-
estimation by — .
1 Ives, H. E., op. cit., p. 171.
* Luckiesh, M., op. cit., p. 620.
* Schenck, F., 'Ueber die Bestimmung der Helligkeit grauer und farbiger Pigment-
papiere mittels intermittirende Netzhautreizung,' Pfliiger's Archiv, 1896, 64, pp. 607-
628.
«Dow, J. S., op. cit., pp. 130-134; 'Physiological Principles Underlying the
Flicker Photometer,' Philos. Mag., 1910, jp, Ser. 6, pp. 58-77.
6 Ives, H. E., op. cit., p. 172.
FLICKER PHOTOMETRY
747
TABLE IX
Showing in percentage the difference in results between the methods of flicker and
equality of brightness for the five observers used by Dr. Ives at the intensity
of light which he calls most favorable.
\
H. E. I.
M.L.
P. W. C.
C. F. L.
F. E. C.
•653M
-12. %
+ 29- %
-18. %
-Si- %
-50. %
.643^
- 3-6
+ 56.
- 7.0
-31.
-23-7
.63 2/t
- 4-3
+ 20.
-iS-5
-45-5
-12.9
.622/1
- 7-3
+ I2.S
- 4.2
-42.
-iS-9
.612^1
— 10.
+ 8.3
- 6.5
- 7-5
+ 0.3
•594M
— i.
~ 0.5
- 0.5
+ 7-5
+ 5-8
•574M
+ 0.5
- 2.4
- 2.5
+27.
- 5-9
•555M
- 0.4
- 8.0
-ii,Q
- 4-8
+ 8.9
•545M
- 3-i
- 8.4
-13.8
+ 3-
+ 8.
.536^
- 1-9
- 4.0
-12.6
— 12.
+ 14-3
.526/11
0
- 8.7
-21.4
- 3-
•4-33-5
•5I7/*
+ 0.6
— 10.8
-13.8
-33-
+30.0
It has been our purpose in general in this part of the
paper to indicate a field of investigation in the department
of physiological optics about which little is known as yet with
certainty, rather than to report a finished piece of work or to
attempt to draw positive conclusions. When functioning
under the conditions imposed by the method of flicker, too
little is known of the characteristics of the eye, we believe,
to render safe its use as a measuring instrument. Our
purpose in particular has been to point out and show the effect
of a factor which we believe to be an important source of
disagreement between the equality of brightness and the
flicker methods, and to suggest that a more careful study be
made of the factors that influence the method of flicker before
it is adopted in its present form as the method for the stan-
dardizing laboratories. Just as one factor has been over-
looked, so there may be others the influence of which should
not be ignored.
APPENDIX
Three other points which may be of interest in connection
with the above work are appended here. The first two were
discussed by Dr. Ives in a series of articles on the method of
flicker in the Philos. Mag., 1912, 24, Ser. 6, pp. 149-188,
352-370, 744-751, 845-853, 853-863. (i) In the third of his
series of articles, he apparently wishes to show that the cause
148 C. E. FERREE AND GERTRUDE RAND
of the disagreement between the results of the methods of
flicker and equality of brightness lies on the side of the latter
method. That is, the difficulty of making the judgment is so
great that not an equalization, only an 'appraisement' is
accomplished. To demonstrate this, he attempts to get rid
of the disturbing factor of color difference in the equality of
brightness method by making his comparisons always between
lights differing only slightly in composition. That is, a
green is compared with a green slightly shifted toward the
yellow or blue, etc. (See his work with the 'cascade'
method, p. 748.) A curve of luminosity for the spectrum
obtained in this way is found to agree more closely with the
flicker curve than one obtained in the ordinary way. The
following things may be said of this demonstration, however.
In the first place, he states that the cumulative errors are so
great in the method that he could not begin at one point in
the spectrum having a given luminosity and work in a given
direction, then reverse this direction of working and obtain
at all a close approximation to the luminosity value for the
point at which he started. For this reason he drops the
point by point procedure of the 'cascade' method, and plots
his curve by taking his observations at twelve points in the
spectrum. From the observations of these points the whole
curve is constructed. In the second place, his method does
not entirely accomplish his purpose of getting rid of all differ-
ence in color quality between the lights compared. In order
to add some further data bearing upon the question whether
the lack of agreement hitherto found between the results
obtained by the equality of brightness and flicker methods
could have been due to the difficulty of making the equality
of brightness judgments of fields differing in color quality,
we have thought it worth while to make the comparison
using an equality of brightness method which for the pur-
poses of this investigation presents, we believe, some points
of advantage over the method used by Dr. Ives.1 That is,
*We do not, however, mean to propose this as an entirely satisfactory method
of heterochromatic photometry for the reason given in the discussion of the relation
of the method to the Talbot-Plateau law (see footnote p. 149). We are using the
method heie merely to show that when the disturbing factor of color difference in the
FLICKER PHOTOMETRY H9
the method we have used offers even less chances for errors
in judgment, is simpler, and entirely eliminates the presence
of a second color in the fields to be compared. The method
is as follows: The sectored disc was adjusted so that its
outer edge bisected vertically the photometric field. A
standard colorless light was moved to the position on the
photometer bar that gave the judgment of equality by the
method of flicker, and the disc was rotated at the fusion rate.
Half of the field was thus of color of the original saturation
and luminosity, and the other half was a fusion of the
colored sector of the original saturation and luminosity and a
gray sector of the luminosity of the color as determined by
the method of flicker. Now, if the luminosity of the color by
the method of flicker were the same as by the equality of
brightness method, the two halves of the photometric field
should match in luminosity (within the limits imposed by the
Talbot-Plateau law).1 That is, the addition of the colorless
fields to be compared is eliminated from the equality of brightness method, there is
still a large, in fact an apparently undiminished characteristic difference between
the results of the equality of brightness and flicker methods, which, so far as one
can see, can in no way be ascribed to the equality of brightness method employed.
The degree to which the influence of color difference on the judgment of the bright-
ness equality of the fields compared is removed by this method is shown by the greatly
increased reproducibility of the judgment. For our observers, the reproducibility
is almost as great as it was for the method of flicker. There was thus but little more
of the element of appraisement in this method than there was in the method of flicker,
while the characteristic difference in the results obtained by the two methods was not,
so far as could be determined, appreciably lessened.
1 A few words are needed to explain what is meant above by "within the limits
imposed by the Talbot-Plateau law." It could scarcely be expected from a considera-
tion of this law that the two fields would match especially under the dark-room condi-
tions under which photometry is done, even when the gray sector was chosen equal in
brightness to the color by the equality of brightness method. That is, when the
colored is mixed with the gray sector by the method of successive impressions, there
is a reduction of the intensity of each impression which is the same as would be
gotten were the intensity of each light to be reduced in proportion to the time of ex-
posure of the eye to each light to the total time of exposure of the eye to both lights.
(See the discussion of the Talbot-Plateau law, p. 121.) That is, if the value of each
sector is 180°, the impression made upon the eye by each light is the same, according
to the Talbot-Plateau law, as if both lights were reduced one-half in intensity. But
in suffering the reduction, the luminosity of the colored sector is not changed the same
in amount as is that of the gray sector. If it is blue or green, for example, its bright-
ness is not reduced so much as is that of the gray sector, and its fusion with the gray
sector tends to lighten that sector and to make the second half of the field lighter than
150 C. E. FERREE AND GERTRUDE RAND
to the colored sector would produce no change in its luminos-
ity, and the two halves of the field would present a fully
saturated color of a given luminosity and a less saturated
color of the same luminosity (within the limits imposed above).
But if there were an underestimation or an overestimation of
the luminosity of the color by the method of flicker, the
brightness of the second half of the field would be modified
in this direction in proportion to the value of the colored and
colorless sector; and if the underestimation or overestimation
were great enough the two halves would not match. In
proportion as the colorless sector is made larger in the second
half of the field, the color of the mixture loses saturation, and
the comparison with the fully saturated half of the fields
becomes more difficult to make. On the other hand, in
proportion as the colored sector is made larger, the effect on
the brightness of the mixture, of the difference between the
flicker value and the true sensation value, if such a difference
exists, is lost. After considerable preliminary investigation
it was decided to use in turn colored sectors of the value of
300°, 270°, and 1 80°. The comparison was made for lights
of the intensities specified in the preceding sections of the
paper. In all cases when the color was red or yellow, the
the first. If, however, the colored sector is red or yellow, it is reduced more in bright-
ness than is the gray sector, and its fusion with that sector tends to darken it and so
to render the second half of the photometric field darker than the first. We have
conducted experiments to determine whether the above effect, which is a direct corol-
lary of the Talbot-Plateau law, actually takes place in observable amounts. When
the light of the spectrum or light of the purity given by the Wratten and Wainwright
filters was used, we found that it did. That is, when the second half of the field was
green or blue and was fused with a gray of the luminosity of the color employed,
determined by the equality of brightness method, this half of the field was observably
lighter than the first half. Conversely, when red or yellow was used, the second half
of the field was darker than the first. The effect, however, was not nearly so great as
it was when the gray sector was made of the brightness of the color as determined
by the method of flicker. That is, if two experiments are conducted, one in which
the second half of the field is made by fusing the colored sector with a gray sector
of the brightness of the color as determined by the equality of brightness method, and
the other in which this half of the field is made by fusing the colored sector with a gray
sector of the brightness of the color as determined by the method of flicker, the differ-
ence in brightness between the two halves of the field is quite appreciably greater
in the second case than in the first. For example, when the colors are green and blue,
the second half of the field is more too light in the second case than in the first; and
when red and yellow, it is more too dark.
FLICKER PHOTOMETRY I51
second half of the field was darker than the first; and when
either blue or green, was lighter than the first half of the field.
Determinations were made also of how much the colorless
light had to be moved to make the two halves of the field
match. These distances were not much different from those
contained in the tables in the preceding sections of the paper
expressing the difference in the estimation of the luminosity of
the colors by the methods of flicker and equality of brightness
(see pp. 136, 137), — certainly not any more than should be
expected when it is remembered that a part of the effect
of the difference is lost by mixing the colorless light repre-
senting the flicker determination with a sector of the colored
light in its true luminosity value. The work was done also
with pigment papers with a similar result. Thus it seems
reasonable to conclude that the cause of the disagreement
between the two methods can not be attributed entirely at
least to the difficulty of making the equality of brightness
judgment due to the difference in color quality between the
fields compared, for in the above cases the color quality of
the lights compared was the same. In the third place, dis-
regarding the results of the above experiments, the writers
scarcely need point out that it would be extremely difficult
to explain such a systematic drift of luminosity in one direc-
tion in one part of the spectrum, and in the opposite direction
in the other part, as we obtained, in terms of errors due to a
false judgment of the sensations actually aroused. More-
over, it would be just as difficult to explain Dr. Ives's own
reverse Purkinje effect in terms of a false judgment of the
actual brightness values presented in sensation; or the closer
agreement he obtains between the results by the methods of
flicker and equality of brightness at high illuminations, in
which case there is the maximum amount of color present
and, therefore, the maximum color difference to disturb the
equality of brightness judgment between colored and color-
less light. Moreover, the kind of errors that one finds as
due to uncertainty of judgment is a deviation on either side of
a mean. This occurs when all other factors are eliminated if
several judgments of the same sensation are made. Such
152 C. E. FERREE AND GERTRUDE RAND
errors are compensated for by taking the average or mean of
the determinations. If it is not conceded that they are com-
pensated for, how, for example, can the average of the results
by the equality of brightness method be taken as a standard
in terms of which to evaluate the results obtained by other
methods? (See Whitman, Schenck, Wilde, etc.).1 Surely
this should not be allowed if there were a consistent deviation
in any one direction from the true brightness value for a given
color due to errors in judgment. Moreover, such a character-
istic drift due to errors in judgment is unknown in all previous
work in psychophysics, and not only unknown, but unsus-
pected.
(2) In the fourth paper of the series,2 Dr. Ives applies as
a test to the method of flicker what he calls two axioms of
measurement. These are (a) things which are equal to the
same things shall be equal to each other; and (b) the whole
shall be equal to the sum of its parts. He finds that the
method of flicker satisfies these axioms better than the equali-
ty of brightness method. We would point out that these
tests would not be expected to reveal to any considerable de-
gree the influence of the factor we are discussing. They are
tests which would apply as a check on the power to make the
judgment of the brightness of the sensation properly, or to
any tendency of this brightness equality to drift in one direc-
tion in any part of the spectrum without a compensating drift
in the opposite direction in some other part of the spectrum;
but they are not tests that could be expected to show whether
or not there is underestimation in one half of the spectrum
and overestimation in the other half. For example, the area
of the curve of the spectrum plotted by the method of flicker
might very well sum up to the value of the reassembled white
light because of the compensating effect of the underesti-
mation of one half of the spectrum and the overestimation of
the other half.
(3) Since the foregoing paper was presented, the writers
1 While Dr. Ives does not explicitly state that he takes the equality of brightness
method as a standard in terms of which to evaluate the correctness of the results by
other methods, the point of view is strongly implied in his first paper (loc. cit.).
*Philos. Mag., 1912, 24, Ser. 6, pp. 845-853.
FLICKER PHOTOMETRY *53
have met with the contention from a prominent advocate of
the method of flicker that the effect of a reduction of intensity
is not given by the method of flicker because each individual
impression is carried over until the next is given, with suffi-
cient intensity to preclude the effect of reduction. Whether
or not each individual impression can be considered as carry-
ing over with sufficient intensity to preclude the effect of re-
duction is an important point and should, lest the issue be in
doubt, be included in a discussion of the principles underlying
the method of flicker. It may not be out of place, therefore,
for us to consider the question here briefly, even though it
has not as yet, so far as we know, been discussed in print.
As evidence that each individual impression should be
considered as carrying over with sufficient intensity to pre-
clude the effect of reduction, it was contended, as the case
was presented to us, that the rate used in the method of flicker
is the fusion rate of the two impressions. Two reasons were
given for considering this rate as the fusion rate, (i) If the
two impressions be red and green, for example, yellow is pro-
duced at the rate of succession used in the flicker method.
Yellow, it was pointed out, is a fusion of red and green, and,
therefore, the rate used must be considered as the fusion rate
for these colors. In answer to this point we would again call
attention to the phenomena (see p. 116) which are produced in
sensation when two impressions differing in color and bright-
ness are given to the eye successively at different rates of
speed.1 When the rate is very slow, the effect of separate
and distinct impressions is given, each in its proper color and
brightness. When a little faster rate is used, the impressions
become confused and a flickering effect is produced both in
the color and brightness components of the sensation. When
the rate is made still faster, the flickering of color dies out,
leaving only brightness flicker; that is, the color components
of the two sensations have been fused. That the brightness
1 We wish at this point to state very emphatically that our account of the fusion
of the color and brightness components of sensation at different rates of speed is not
based on any theoretical conception of a separate brightness and color sense, but upon
actual observation of the phenomena that take place when light impressions differing
in color and luminosity are combined at different rates of succession. These phe-
154 C. E. FERREE AND GERTRUDE RAND
components have not been fused, however, is attested by the
presence of brightness flicker, which is now left outstanding
in a field uniform as to color quality. As the rate of succes-
sion is made still faster, brightness flicker becomes less and
less pronounced and finally disappears.1 The rate at which
this disappearance takes place is the fusion rate for the bright-
ness components for the two sensations, and is much higher
for all the colors than is the rate at whieh the fusion of the
color components takes place.2 (Interpreted in terms of the
nomena may be readily demonstrated by any kind of flicker photometer head if a
sufficiently sensitive control of speed of rotation is had. (We have used for the control
of speed of rotation a rheostat and motor especially constructed to give fine changes.)
It can be very plainly and perhaps most conveniently demonstrated by rotating
sectors of pigment papers at the proper gradations of speed in a good daylight illu-
mination.
!We find that Kriiss (Physical Zeitschr., 1904, 5, p. 67) gives a description of
the phenomena that take place in sensation when two impressions differing in color
and brightness are given to the eye successively at different rates of speed, very
similar to that we have given here. He says : " If we slowly alternate the illumination
from two differently colored light sources, for example, from a Hefner lamp and a gas
burner, we clearly distinguish a succession of reddish and bluish bands with weak
washed-out limits between them. As the rate of succession is increased it becomes
progressively more difficult to distinguish the two colors from each other. At a com-
paratively low rate they begin to lose themselves in each other. At a slightly higher
rate the difference in color disappears altogether and we have a color mixture. In this
mixture, however, a brightness succession, a flicker, is observable which disappears
only by a further increase in the rate of succession. Physiologically, it is of great
interest that the distinguishing of separate colors ceases at a much slower rate of
succession than the rate at which completely continuous sensation begins."
2 The following values will serve to give a rough comparative showing of the
rates at which the phenomena described above take place. The colors used were red
and green. They were obtained from pigment papers of the Hering series of standard
papers and from gelatine filters. Two intensities of color were employed in each case.
The brightness of the Hering green for the lower intensity of illumination was .000814
cp. per sq. in.; of the red, .000594 CP- Per scl' m- The phenomenon of separate
impressions occurred from the lowest speed up to 6.9 revolutions per second. The
impression of an intermingled color and brightness flicker was given from this rate
up to 9.6 revolutions per second, at which rate the color components of the sensation
fused, giving a field uniform as to color quality but with a strong outstanding brightness
flicker. Brightness flicker was present until a speed of 22 revolutions per second was
obtained. At this speed the brightness components in sensation were completely
fused and the rotating disc presented a surface uniform both as to color and brightness.
In making these determinations, the same sized field was used as was employed in our
work with the method of flicker, i. e., the disc was viewed through an aperture 3 mm.
X 3 mm- in a gray screen (Hering No. 24) 20 cm. from the eye. For the higher inten-
sity the green surface was illuminated to a brightness of .00242 cp. per sq. in.; the
FLICKER PHOTOMETRY 155
duration of the impression after the light has been cut off,
this means, of course, that the brightness component in the
sensation does not carry over under these conditions with as
little loss of intensity as does the color component.) It is
evident, then, that the rate of succession which is used in the
method of flicker is at or near the fusion rate for the color
components of the two sensations, not for the brightness com-
ponents; nor is it anywhere near the fusion rate for the bright-
ness components. But it is the brightness components in
which we are interested in photometry. That is, it is in
terms of the brightness component that all photometric judg-
ments are made. The color components, when they differ
in tone, only serve to confuse the judgment. It is, therefore,
our object in all methods of photometry as much as possible
to get rid of difference in the color components. This can be
accomplished in the method of flicker only because of the
fact we have just pointed out, namely, that the fusion of the
color component in sensation comes at a much lower rate of
succession than the fusion of the brightness component. That
is, all color differences, whether sensed as distinct or as flicker-
ing sensations, disappear at a rate of succession that has little
or no effect on eliminating the brightness factor, or in this
case the equivalent of this elimination, the fusion of the bright-
ness components of the two sensations. In fact, if there
were no difference in the fusion rate of the color and bright-
ness components, the flickering color impressions would so
mask the presence of brightness flicker at any rate of succes-
sion that could be used, that the method would doubtless
red, .00167 CP- per sq. in. The phenomenon of separate impressions occurred from
the lowest speed -to 6 revolutions per second, at which rate color flicker began. Color
fusion took place at 12.4 revolutions per second, and brightness fusion at 29.3 revolu-
tions per second. At the lower intensity for the filters, the brightness of the green
was .154 cp. per sq. in.; for the red, .099 cp. per sq. in. As compared with their
brightness these colors were much more poorly saturated than were the Hering pig-
ments. The phenomenon of separate impressions ceased and color flicker began at
6.5 per second. Color fusion took place at 11.5 revolutions per second, and brightness
fusion took place at 35.4 revolutions per second. At the higher intensity for the
filters the brightness of the green was .22 cp. per sq. in.; for the red, .143 cp. per
sq. in. Color flicker began at 7 revolutions per second; color fusion took place at
12.9 revolutions per second; and brightness fusion was complete at 38.3 revolutions
per second.
156 C. E. FERREE AND GERTRUDE RAND
have little if any greater sensitivity than the equality of
brightness method. (2) The second point that was cited
in support of the contention that the rate used in the method
of flicker is the fusion rate for the two sensations aroused, is
that no brightness flicker is present when in terms of the
method the two impressions are adjudged of the same bright-
ness. This to the present writers seems indeed a strange
confusion of meanings. Fusion is a term used to represent
what takes place when two impressions or sensations differing
in quality are combined into one, the same or homogeneous
as to quality. This combination may be obtained in case of
light stimuli, for example, by mixing two lights evenly and
allowing them to act simultaneously on the eye; or it may be
obtained by giving two lights to the eye in succession at such
a rate that the sensation aroused by the one lasts over until
the next one is set up with a sufficient degree of intensity to
give the effect of continuity or homogeneity of quality. It
may add to the clearness of our discussion, then, to consider
what takes place in this regard when two impressions differing
in brightness are given to the eye at the different rates of
succession mentioned in the preceding paragraph. At the
rate at which distinct and separate impressions are given,
each sensation obviously dies away completely before the
next one is aroused. If a rate slightly faster than this is
selected, the sensation does not die away completely before
the next one is set up.
There is a slight lasting-over from one impression to the
next. This when the two impressions differ in brightness
gives the effect of a wavering or flickering sensation. At the
lowest speed at which flicker is produced, the effect of this
lasting-over has its minimum value. As the speed is further
increased it becomes greater and attains its maximum value
at the rate of complete brightness fusion.1 (See discussion of
Talbot-Plateau law, pp. 121.) It is obvious, then, that the
rate of speed employed in the method of flicker, which is,
roughly speaking, the lowest rate at which brightness flicker
1 At the fusion rate neither sensation rises to its maximum value, for example, nor
has a chance to die away until the next one develops. The effect is that of a con-
tinuous sensation homogeneous as to color and brightness.
FLICKER PHOTOMETRY 157
can be obtained unmixed with color flicker, is not the fusion
rate for the brightness components in sensation nor is it
anywhere near' this rate.1 It is equally obvious also that the
absence of flicker when the final adjustment of the lights has
been made for a photometric balance, can not be adduced as
any evidence that this rate is the fusion rate for the brightness
component of the two sensations, or, what is more significant
in relation to the above mentioned claim, that it is a rate to
which more than a minimum of lasting-over effect from
impression to impression can be ascribed. Flicker is absent
merely because, in accord with the purpose of the method,
such an adjustment of the distance of the lights from the
photometer head is made that the sensations aroused by the
two lights are of equal brightness. Such sensations do not
flicker whatever may be their rate of succession. It can,
therefore, be considered as little more than absurd to adduce
the absence of flicker when the photometric balance has been
attained as evidence that the rate used is the fusion rate for
the brightness components of sensation, and to pass from this
to the conclusion that the same amount or anywhere near the
same amount of carrying-over effect is present for this rate
as obtains when the fusion rate is used. In fact, if this
carrying-over effect were present to any considerable degree,
the whole point of the flicker method would be lost. That
is, it is the purpose in the method of flicker to select a rate of
succession that will give the eye the maximum of sensitivity
to brightness difference (or flicker), namely, the lowest
rate at which flicker can be produced, rather than a rate that
will fuse out this difference in sensation.
But supposing it could be established, as was contended,
that we have in the rate used in the method of flicker a com-
plete color and brightness fusion of the sensations aroused
by the two lights, little would be gained for the claim that
there is no reduction in the effect on sensation of the two
lights employed, if it be granted, for example, that the
Talbot-Plateau law is true. In substance this law is as
1 Flicker and fusion are in fact antithetical terms, and the rates of succession
which are favorable for each are widely separated in the scale of frequencies.
158 C. E. FERREE AND GERTRUDE RAND
follows. When once the rate of rotation is sufficient to give
a uniform sensation, the color and brightness of the disc are
the same as they would be if all the light reflected from the
sectors were evenly distributed over the surface of the disc;
and no further increase in rapidity produces any effect on its
appearance.1 In terms of this law it is seen that the effect on
1 See H. F. Talbot, 'Experiments on Light/ Philos. Mag., 1834, Ser. 3, 5, pp. 321-
334-
Talbot phrases this law as follows (pp. 328-329): "Since then these two things —
the intensity of light and the time of the body's remaining in any given part of the
circle — are each inversely proportional to the circumference of the circle it describes,
it follows that they must be directly proportional to each other; that is to say, an
irregular intermittent luminary whose observations are too frequent and too transitory
for the eye to perceive, loses so much of its apparent brightness from this cause as is
indicated by the proportion between the whole time of observation and the time
during which it disappears." "The rapidity of the rotation does not affect the argu-
ment." To verify this reasoning, Talbot conducted experiments with reflected light
using pigment surfaces and mirrors to send the light to the eye; and with transmitted
light using sectored discs to cut down the time of exposure of the eye to various luminous
sources.
In 1835 Plateau repeats and verifies Talbot's experiments. (* Betrachtungen
fiber ein von Hrn. Talbot vergeschlagenes photometrisches Princip,' Poggen. Annal,
1835, 35, pp. 457-468). He concludes from his experiments as follows (pp. 462-463)
"Nun muss zufolge des am Anfange dieses Aufsatzes dargelegten Princips die schein-
bare Helligkeit der Scheibe sich zu der des Papiers verhalten wie die Voriibergangsdauer
eines weissen und eines schwarzen Sectors; odor was dasselbe ist, wie die Winkelbreiten
eines weissen Sectors zur Summe der Winkelbreiten eines weissen und schwarzen
Sectors, oder endlich, was auch noch dasselbe ist, wie die Breite sammtlicher weisser
Sectoren zum ganzen Kreisumfang."
Swan, apparently working in ignorance of the writings of Talbot and Plateau,
in substance formulates the law anew in 1849 (see W. Swan, 'On the Gradual Produc-
tion of Luminous Impressions on the Eye and Other Phenomena of Vision,' Trans.
Roy. Soc. Edinb., 1849, 16, pp. 581-603. See also F. Boas, 'Ein Beweis des Talbot'-
schen Satzes und Bemerkungen zu einigen aus demselben gezegonen Folgerungen/
Wiedem. Ann., 1882, 16, 359-362; A. M. Bloch, * Experiences sur la vision,' Compt.
Rend, de la Soc. de Biol, 1885, 2, p. 495; A. Charpentier, 'Loi de Bloch relative aux
lumieres de courte duree,' ibid., 1887 4, p. 5; etc.
For a more modern statement of this law and one also more consistent with the
relation of changes in light energy to changes in sensation, see Helmholtz, 'Handbuch
der physiol. Optik,' zw. Aufl., 1896, p. 483, "Wenn eine Stelle der Netzhaut von
periodisch veranderlichem und regelmassig in derselben Weise wiederkehrendem Lichte
getroffen wird, und die Dauer der Periode hinreichend kurz ist, so entseht ein continuir-
hcher Eindruck, der dem gleich ist, welcher entstehen wiirde, wenn das wahrend einer
jeden Periode eintreffende Licht gleichmassig uber die ganze Dauer der Periode
vertheilt wiirde"; or E. C. Sanford, 'Experimental Psychology,' 1898, p. 146, "When
once the rate of rotation is sufficient to give a uniform sensation, the color and bright-
ness of any concentric ring are the same that they would be if all the light reflected
FLICKER PHOTOMETRY
sensation is the same as is gotten by reducing the intensity of
each light by an. amount proportional to the ratio of the
exposure time of that light to the total time of exposure to
both lights; or in case the photometer head is a sectored disc,
in proportion to the value of the given sector or set of sectors
to 360°. That is, with a total value of each sector or set
of sectors of 180°, the effect on sensation is the same as if each
light were reduced one-half in intensity; if the total value of
one sector or set of sectors is 90°, the effect on sensation is the
the same as if the light illuminating that sector were reduced
to one-fourth of its intensity; if the total value were 45°, the
same effect is produced as if the light were reduced to one-
eighth of its intensity; etc. Thus, even if the rate of suc-
cession that is used in the method of flicker could be con-
sidered as the fusion rate for the brightness component of the
sensation aroused, little advantage could be gained for the
position in question. For the conclusion most certainly could
not be avoided that the effect on sensation would be the same
as if the lights were reduced in intensity, and by an amount
proportional to the ratio of exposure time of each light to
the total time of exposure to both lights.
The position under discussion seems also to involve to
some extent a confusion of principle of the method of flicker
with the method of critical frequency. For example, in the
method of critical frequency, the impressions are given to the
eye at the fusion rate. We need scarcely call to mind the
procedure. One sector or set of sectors of the disc is illu-
minated by one of the lights to be compared and the other is
black or of a very low luminosity. The disc is rotated at a
rate which completely fuses the sectors in sensation. This
light is then removed and the other light to be compared is
substituted for it. The distance of this light from the disc
is then adjusted until the rate of rotation required to produce
fusion is the same as it was in the previous case. When this
adjustment is obtained the intensity of illumination of the
disc by the two lights is said to have been the same, and the
from it were evenly distributed over its surface, and no further increase in rapidity
produced any effect on its appearance."
160 C. E. FERREE AND GERTRUDE RAND
relative brightnesses of the lights themselves are calculated
by the law of inverse squares. The situation is, however,
quite different for the method of flicker. Both sectors or
sets of sectors of the disc are illuminated by the lights to be
compared, and the rate of rotation is to be made such that
if there were any brightness difference between the sectors,
the maximum of flicker, not fusion, would be produced. If a
rate were used that would produce fusion, for example, for
any given amount of brightness difference, it is obvious that
no difference in brightness equal to or less than this amount
could be detected by the method. That is, the whole point
of the method is to use a rate of speed that could not possibly
be the fusion rate for any appreciable amount of brightness
difference between the impressions to be compared; and in so
far as this purpose can be realized in the different cases in
which the method is employed, sensitivity for the method is
obtained.
What our critic really needs to establish in order to support
his position is that summation instead of fusion takes place.
That is, if the total effect of each light on sensation is to rise
to a higher level than is given by each individual impression,
the individual impressions must in proportion to the rise
summate or add their individual intensities. To produce this
effect of summation, each individual impression would have
to last over in sensation until the next impression of its kind
is received, which, since the impressions alternate, could be
the next impression but one. For example, when red and
green lights are being compared, if the value of the red sensa-
tion is to rise to a higher level than that given by a single
impression, the sensation aroused by one exposure to red
would have to last over until sensation is aroused by the next
exposure to red; that is, would have to last through the inter-
val of exposure to green and into and wholly or partly through
the succeeding interval of exposure to red. How highly
improbable it is that this could happen to any degree that
would be of saving consequence to the method, is shown by the
following two considerations, (a) The wavering character of
the sensation which we call flicker is due to the fact that a
FLICKER PHOTOMETRY 161
given sensation does not carry over without a great loss of
intensity through the next succeeding interval, let alone through
the next interval but one. And (b) even at the rate at which
complete color and brightness fusion takes place, there is
according to the Talbot-Plateau law no effect of summation
great enough to cause each individual sensation to attain to a
higher intensity than that fixed by the ratio of the time of
exposure of its stimulus light to the total time of exposure of
both lights, nor to produce a noticeable change in this inten-
sity, however great is the speed of the succession. That is,
we have a reduction of the intensity of the sensation aroused
by each light which is the same as would be "gotten were the
intensity of each light to be reduced by an amount propor-
tional to the ratio of the time of exposure of that light to the
total time of exposure of both lights, and no further increase
in the rapidity of the succession produces any change in this
effect.1
With regard to the method of flicker, then, the case
apparently stands as follows. The individual impressions
are so short that the eye is very much underexposed to its
stimulus, and the rate of succession is so slow that there is
1 If one were permitted to interpret the Talbot-Plateau law with regard to what
takes place when a rate of succession is employed greater than the fusion rate for both
the colored and brightness components of sensation, two possibilities would be opened
for explaining why no change in sensation is produced as the rate of succession is
increased, and the length of each individual exposure is correspondingly decreased,
(i) Either the increase in the reduction of the exposure-time causes no further reduction
in the sensation aroused by the individual exposures; or (2) there is, owing to the
increased rate of succession, a summation effect which just compensates for the reduc-
tion of the individual impressions. Now even if we were to accept as true the one of
these alternatives which is the more favorable for the case of flicker, namely, that a
compensating summation action takes place, and assume that this compensating
summation obtains clear down to the rate of succession that is used in the method of
flicker, we would have to expect as much reduction in the sensation aroused by each
of the lights as is expressed by the Talbot-Plateau law. That is, the reduction for
each would be the same as would be gotten were the intensity of each light to be re-
duced in proportion to the exposure-time of each to the exposure-time of both. As we
have already pointed out, however, it is extremely improbable that there could be a
compensating summation action at the flicker rate great enough to be of any consid-
erable consequence to the method, because the wavering character of the sensation
which we call flicker is due to the fact that a given sensation does not carry over
without great loss until the next one develops, let alone until the next but one
develops, which it would have to do to produce any summation effect.
1 62 C. E. FERREE AND GERTRUDE' RAND
not enough carrying-over from impression to impression to
produce fusion, let alone the summation effect which is
needed to cause the intensity of the sensation to rise to its full
value or perhaps even to a higher level than would be given
by a single exposure. Moreover, according to the Talbot-
Plateau law a summation effect great enough to cause the sen-
sation to rise to its full value is never produced, however fast
is the rate of succession; for once the fusion rate is obtained,
there is a reduction of the intensity of the sensation aroused by
each light which is the same as would be gotten were the
intensity of each light to be reduced in proportion to the
time of exposure of that light to the total time of exposure
to both lights, and there is no change in this effect however
much the rate of succession is increased.
1 Since the above discussion was presented to the Illuminating Engineering
Society, Ives in collaboration with Kingsbuiy, has published a sixth article on the
method of flicker (Philos. Mag., Nov., 1914, 28 (167), pp. 708-728) in which a theory
of flicker photometry is developed based on an analogy drawn between the response
of the eye under successive stimulation to the action of incandescent lamp filaments
under a fluctuating current. The gist of the article is that if the eye behaves under
the conditions obtaining in flicker photometry as do lamp filaments (subject to cer-
tain modifications which are not in accordance with what is known of the function-
ing of the eye) under a fluctuating current, the method of flicker should give with
high intensities of light at the photometer screen the same results on the average as
the equality of brightness method. It is our purpose here merely to note the article,
not to give a detailed discussion. The theory will be discussed in a later paper in con-
nection with further experimental data. It may not be out of place to state at this
time, however, that the analogy of the eye and the incandescent lamp filament is not
based on experimental examination of the eye's manner of response, but is assumed.
Moreover, considerable evidence is offered in the present paper, we think that the
eye does not react to its stimulus given to it in succession at the flicker rate according
to the laws which govern the temperature response of lamp filaments, more especially
when the impressions differ widely as to wave-length. It has not been claimed, for
example, that the flicker method does not give the same results as the equality of
brightness method when the lights compared do not differ as to wave-length.
DISCUSSION
THE FUNCTION OF INCIPIENT MOTOR PROCESSES1
There is no doubt that such a theory as the author discusses is
of important advantage, yielding a base for a fair understanding
of nervous functions. In regard to the assumption that the dis-
charge of a motor center may induce the discharge of a cortical
center that is tributary to it, there is evidently something left to
the imagination of the reader. It is possible that from the point
of view of practical or quantitative science, so to speak, some other
hypothesis may be found more defensible.
With this in mind the writer will venture to describe the nervous
mechanisms that produce the image, holding to the author's first
assumptions but departing from the induced discharge assumption.
To bring out the point quickly, let us begin by considering the
following case that is easy of explanation, and lead up gradually
to the functions under controversy.
If a child sees a red ball and utters the word ball, and then
makes a forward movement, certain associations will be formed.
At a later time the child is prompted to utter the word ball but
the movement is only partially carried out and the word is inaudible.
An image of the red ball appears in the child's mind. Now in the
first occurrence which we may term the experience, we may say
that there are afferent impulses due to the sight of red, to the shape
of the ball and to the sound of the uttered word. These go to the
cortex, or at least a part of each kind does so. There are also
kinsesthetic impulses from the eye muscles and from the throat
and lips, and a part of these causes excitation in the cortex.
In the second occurrence, which we may term the recall, there
are kinaesthetic impulses from the muscles used for the word and
no doubt some of these will be just the same as if the word had not
been suppressed but uttered aloud. Now if conditions are right,
these latter excitations will reach the cortical centers which were
excited in the experience. The author's first assumptions and
theory of association are the explanation. Take the color red for
example. In the experience, the excitation starts in the retina,
1 M. F. Washburn, PSYCH. REV., Vol. XXI., No. 5.
163
164 S. BENT RUSSELL
thence goes to the cortical center for red, thence it flows to the motor
centers in activity. Within a moment the excitations from the
speech muscles pass through the cortex and perhaps follow the
identical neurons just stimulated by the color. By the rules,
the common pathway will have its conductivity increased by the
experience.
When the recall comes, the flow from the receptors in the
muscles will follow this line of increased conductivity, pass through
the cortical center for red and hence an image of red will arise.
The flow will proceed from the cortical center by the pathways
that are open to some motor center or centers. Thus we see that
the suppressed utterance of the word ball has brought up an image
of the color red and yet there has been no induced discharge such
as the author describes. For the above demonstration, it is essen-
tial that in the recall some of the muscles be partially contracted
so as to cause the kinsesthetic excitations which we assume to
follow from muscular contractions.
Let us now go a step further and suppose that some time later
there is an occurrence we will term the secondary recall. The child
is prompted to utter the word but there is no movement and no
real contraction nor even a noticeable change of tone in any muscle.
Again an image of the red ball appears. It is probably fainter than
in the previous case but it is still clear.
To explain the secondary recall, we will advance the theory of
strain signals.
Beginning with the motor discharge which prompts the utter-
ance but is not of sufficient intensity to cause muscular contraction,
we are brought to the motor terminal in the muscle. Let us here
make the following assumption:
When a nervous discharge to a motor terminal is too weak to
cause contraction it will produce a chemical or molecular change in
the muscle substance which spreads to the sensory terminals,
causing an excitation of certain neurons, which we will term a
strain signal. This change in the muscle substance requires about
the same time as a muscular contraction.
With the aid of this assumption our explanation of the secondary
recall will follow the same course as for the other recall. The strain
signal acts upon the cortex just as a kinsesthetic excitation would
and stimulates those very sensitive cortical neurons which give
rise to the image. Thus we see again that the incipient utterance
of the word ball has brought up an image of the color red. It is
MOTOR PROCESSES 165
worth noting that between the two types of recall that we have
discussed, there are possible stages where kinaesthetic impulses from
some muscles are joined by strain signals from others to arouse the
image.
We find that the theory of strain signals is in some conformity
with the good old rule that what is true approaching a limit is true
at the limit, for the image arises from a discharge coming from the
muscle to the cortex, both when there is some contraction occurring
and when the contraction is incipient only. Moreover the theory
appears to be borne out by introspection as when you recall a
song, the words seem to sound in your ears at about the same rate
of succession as if you were singing them. Again, the intimate
relation of nerve to muscle would indicate that a disturbance of
the motor nerve, however faint, would cause a change of some kind
in the muscle as the theory requires. It may be only a sort of
ripple like a sound wave that traverses the muscle. If the reader
has given much thought to such matters, he will be able to find
other arguments in favor of the theory of strain signals.
It would take too long to discuss fairly the matter of "imageless"
conscious processes or degrees of clearness or faintness of images.
We may briefly note, however, that in considering these matters,
one should keep in mind the rules for the formation of associations
and the changes that occur during the development of a movement
system or performance. As the performance is being perfected by
practice, unnecessary movements are dropped. But the dropping
of movements means the elimination of kinaesthetic impulses and
also certain changes in the excitations due to the reaction of the
environment. These eliminations and changes will naturally result
in the fading and disappearance of images and in the formations
of new associations. By way of illustration, remember that the
images of Tuesday reflect the movements of Monday and prevail
over the faded images that reflect the movements of Sunday.
In the final stage when the performance has become automatic,
the only paths of high conductivity will be those connecting the
movements or essential to the performance.
Finally it is submitted that the considerations brought out by
the author regarding attention would be met by the theory of
strain signals equally well as by the author's theory of incipient
motor processes involving induced discharge.
Comparing the two theories, we observe that the author's
theory assumes that a motor discharge that is too faint to cause
1 66 5. BENT RUSSELL
contraction of the muscle is strong enough to induce a discharge in
the extremely sensitive tributary cortical center. The theory of
strain signals assumes that a motor discharge that is too faint to
cause contraction is strong enough to excite certain sensory terminals
in the muscle which have communication with cortical centers.
S. BENT RUSSELL
VOL. XXII. No. 3 May, 1915
THE PSYCHOLOGICAL REVIEW
THE THEORY AND PRACTISE OF THE ARTIFICIAL
PUPIL
BY LEONARD T. TROLAND
Harvard University, Cambridge, Mass.
So far as the writer can ascertain the references in the
literature to the theory and application of artificial pupils,
although not infrequent, are quite unenlightening. Yet in
all work upon the visual processes in which the amount of
light energy striking the retina has to be controlled the
artificial pupil would seem to be an indispensable accessory.
Its value in exact studies upon visual acuity is also self-
evident.
The intensity of the light which strikes the retina at any
time is determined not only by the intensity and distance
of the source, but also by the size of the pupil. It is useless
to experiment upon the effects of lights of different objective
intensities upon the retina if the reaction of the pupil to these
lights is disregarded, for as soon as the objective intensity is
increased the pupil contracts, and vice versa, so that there is
a tendency for the retina to receive a constant illumination,
independent of changes in the intensity and distance of the
stimulus. This tendency fails to be effective only for very
bright or for very dim lights, for which the pupil has attained,
approximately, its minimum or maximum opening.
In addition to the compensating effect just mentioned
the behavior of the natural pupil offers another difficulty to
the student of retinal physiology in the continued fluctuations
in opening which it exhibits even for a constant illumination.
These fluctuations are periodic in character, but they follow
no definite law, and the average aperture about which they
167
1 68 LEONARD T. TROLAND
hover varies for different persons and for the same person at
different times. Such variations are sufficient to render
impossible accurate comparative tests of retinal sensitivity
without the introduction of some further artifice.
A partial solution of the difficulty lies in temporarily
disabling the pupillary reflexes by the use of. such drugs as
homatropin and pilocarpin. This procedure, however, is not
feasible in extensive researches on account of the discomfort
which it entails for the subjects. Moreover, it does not
insure the same pupillary opening for different persons, or
even, at different times, for the same person, so that if
results are to be made comparative the size of the pupil
must be measured for each series of observations.
The simplest and surest way in which to eliminate the in-
fluence of the pupil upon retinal measurements would seem to
be to place in front of the natural pupil a diaphragm the
aperture of which is smaller than the smallest aperture of the
natural pupil, and which is concentric with the latter. It is
the purpose of the present paper to discuss the theory and
practice of such an "artificial pupil."
The three important problems which are involved in the
use of the artificial pupil consist in the determination of the
proper size of the diaphragm, of the distance from the eye
at which it must be placed, and the invention of some means
of insuring the coincidence of the axis passing through the
center of the stimulating field and that of the diaphragm with
the center of the natural pupil. The geometrical and optical
conditions under which the artificial pupil must be employed
make it necessary for the stimulus to be distinctly limited in
angular size, and thus make impossible its application in the
study of vision in the extreme periphery.
The necessary size of the aperture of the artificial pupil
depends upon four variables: (i) the diameter of the surface
used as a stimulus, (2) the distance of this stimulus from the
eve> (3) tne minimal size of the natural pupil under this
sort of illumination, and (4) the distance between the artificial
and natural pupils. The main condition for the successful
use of the artificial pupil is that it should be so adjusted that
THE ARTIFICIAL PUPIL
169
none of the light from the stimulus is intercepted by the
iris of the eye itself.
Although the refraction which occurs as the light passes
through the surface of the cornea narrows the pencil of rays,
practical considerations nevertheless demand that the arti-
ficial pupil be smaller than the natural one at any time. In
the ensuing discussion we shall neglect the effect of refraction
at the corneal surface since the error which such neglect
introduces into our calculations merely contributes to the
large margin of safety which is necessary, at all events in the
use of the artificial pupil. The argument becomes more
rigid, and also more immediately applicable to practise, if for
o, below, the diameter of the so-called Eintrittspupille is em-
ployed, in place of the actual pupillary opening, the value of x
being taken to correspond. This means using the size and
distance of the apparent pupil rather than of the actual. Prac-
tically, however, the difference between these two cases may be
neglected.
The accompanying diagram, Fig. i, represents in cross-
section the arrangement of the artificial pupil with respect
to the natural pupil and the stimulating surface, a is the
diameter of the artificial pupil, o that of the natural pupil,
w that of the stimulus, d is the distance from the plane of
170 LEONARD T. TROLAND
the stimulus to that of the iris, while x is the distance between
the iris and the artificial pupil. The lines AB and CD
represent those light rays which form the critical boundaries
of the rays passing through the artificial pupil. (It may not
be immediately obvious why these lines are the ones which
it is important for us to consider, rather than the external
boundaries of the whole pencil of rays, but a study of the
diagram will make this clear.) The point of intersection, E,
of the lines in question has critical significance, and may be
called the crossing-point, the distance of this point from the
plane of the natural pupil being /.
Inspection of the diagram shows the following relations
to be true:
w -r;- •
If we solve for / in (i), and for a in (2), and then eliminate /,
we get:
o(d — x) — wx
(3)
d
which gives us the maximum diameter of the artificial pupil
which will satisfy the prescribed conditions.
On account of the relative convergence of the pencil of
rays after it has passed through the cornea the diameter
calculated by the above formula would not, strictly speaking,
be the largest available opening. However, in practise it
would be very unsafe to utilize an opening closer to the maxi-
mum. There are two reasons for this: first, the fact that
slight accidental movements of the eye and head are un-
avoidable, even with the best of head-rests and fixation, and,
second, the fact that the natural pupil is subject to constant
fluctuations in size. Ordinarily it is advisable to work with
an opening at least two millimeters smaller than the maximum
for the smallest aperture of the natural pupil which is to be
expected in the course of the observations.
THE ARTIFICIAL PUPIL
The distance, /, of the crossing-point from the plane of the
iris may be calculated from the formula:
which follows from (i), above. The position of this point
depends upon the size and distance of the stimulus and upon
the aperture of the natural pupil. A knowledge of it is
important, since if the artificial pupil is placed in front of
the crossing-point it will necessarily fail in its purpose, no
matter how small it is made. The distance of this point
from the eye is, in general, of sufficient magnitude so that
the artificial pupil may be placed in a position comfortable
to the observer. For example, when the diameter of the
stimulus is 5 centimeters, its distance I meter, and the natural
pupil 3 millimeters, / is approximately 6 centimeters.
To determine the maximum admissible diameter of the
stimulus under given conditions, the original equations may
be solved for zv, with the result:
, N d(o — a) — ox
(5) m = -s - -f - .
•V
With, for example, a pupil aperture of 4 millimeters, an
artificial pupil of I millimeter, and the values of x and d
used in the example above, we find w to be approximately
30 centimeters. The corresponding angular aperture is 17°.
It is clear that if we wish to increase the angular size of the
stimulus we must decrease a or x or both. The limit for x
is the distance from the iris to the cornea, viz., a little under
4 millimeters, that of a is zero. Substituting these values in
place of those first employed we get: w = (approximately)
I meter. This corresponds to an angular opening of about
53°. It would appear to be the maximal size of stimulus in
connection with which the artificial pupil can be used under
ordinary conditions.
Evidently, however, this maximum is of such a character
as to make the artificial pupil available in the study of all of
the color perceptive regions of the retina, since these do not
extend, in general, beyond 50°. In practise, of course, it
172 LEONARD T. TROLAND
would be necessary for the pupil to have a finite aperture
and to be removed from the cornea by a distance somewhat
greater than that allowed in the above calculation (viz.
.5 mm.).
The best possible conditions for the use of the artificial
pupil would be those which go with complete mydriasis.
Under these conditions the aperture of the natural pupil is
about 7.5 millimeters. With an artificial pupil of I millimeter
aperture, and a distance from the iris of 8 millimeters, the
angular size of the largest available stimulus would be about
46°. The worst possible conditions are those of complete
miosis, which give a natural pupil of about 1.5 millimeters.
The maximum angular size of the stimulus for the latter
conditions is approximately 11.5°. These maxima include
the margin of safety, introduced by corneal refraction, which
was mentioned at the outset. In general, of course, the
artificial pupil would not be used in connection with the drugs
which are customarily employed to produce mydriasis and
miosis, and consequently it is necessary to base one's calcula-
tions upon the so-called physiological pupil, which lies
between 3 and 4 millimeters for a considerable range of
intensities of the stimulus.
The final, and perhaps the most difficult problem which
must be solved in the use of the artificial pupil is that of
securing what may be called register between the natural
and the artificial diaphragms. Perfect registration may be
defined as a disposition of the eye with reference to the
artificial pupil such that the axis passing through the centers
of the stimulus field and the artificial pupil, and perpendicular
to the planes of these, also passes through the center of the
natural pupil, and is perpendicular to the plane of the iris.
Under these conditions the projection of the natural pupil
on the plane of the artificial pupil is concentric with the
latter.
The conditions accompanying the use of the artificial
pupil are such as to make it difficult, if not impossible, to
secure or test registration by objective observations. Ap-
proximate registration can be obtained by moving the head
THE ARTIFICIAL PUPIL 173
slightly with the eye in position, since when the artificial
and natural pupils do not coincide the intensity of the
stimulus appears to be reduced. However, in most work in
which the artificial pupil is helpful it is desirable that registra-
tion should be secured before the eye is exposed to the action
of the stimulus.
One method of securing accurate registration would be
to place two dimly illuminated diaphragms concentrically on
the axis passing perpendicularly through the center of the
stimulus field, these diaphragms to be at different distances
from the latter and of such size that, when the eye is in
position, the edge of the farther one can be seen framed in
that of the nearer. If the artificial pupil now be placed
concentric with the diaphragms the eye will be in register
when the framing of the one diaphragm in the other appears
concentric.
A neater and somewhat simpler method of insuring
registration — the one now in use by the author — is the
following:
It is a well-known fact that if a small source of light be
held very close to the eye it will be seen not in its true form,
but as a relatively large " diffusion circle,"1 fluctuating in size
with the constant changes which are occurring in the size of
the pupil. Such a circle is in reality a luminous shadow of
the natural pupil, and if the point of light be on the line of
sight the diffusion circle which it produces will be concentric
with the fovea, provided, of course, that the pupil itself is not
eccentric. If, now, a true circle be placed concentric with
the same axis, but at a greater distance, so that a more or
less distinct image of it can be formed on the retina, this
image will be seen to be concentric with the diffusion circle,
but if it is displaced it will become eccentric with respect to
the latter.
If an adequately small artificial pupil be placed in front of
the eye and in register with the natural pupil the size of the
diffusion circle will be determined by the former instead of
^n the theory of the " Zerstreuungskries" see: Helmholtz, "Handbuch der
physiologischen Optik," 3d ed., 1909, Vol. I, pp. 101-120.
'74
LEONARD T. TROLAND
the latter. With perfect registration the Zerstreuungskreis
in question will be seen to be concentric with the image of
the second circle mentioned above, but when the registration
is imperfect the two will be eccentric with respect to each
other. We are thus provided with a very accurate means of
securing and of testing registration.
Fig. 2 represents, somewhat diagrammatically, an element
of the artificial pupil apparatus which the writer has been
using in his investigations concerning retinal fatigue, and
FIG. 2.
which embodies in a general way the principle just described
together with the others previously discussed. The pupillary
diaphragm, P, is held at the end of a small telescoping tube
before the other end of which a small electric light, Z, carrying
in front of it a piece of opal glass, G, can be let down by the
movement of a shutter, S. When this light is in position
and the eye is held opposite the artificial pupil, the reflection
of the light from the interior walls of the tube near to the eye
THE ARTIFICIAL PUPIL 175
produces a large diffusion circle upon the retina. Within
this circle of light is seen a smaller, dark, circular ring which
is constituted by the image of the circular diaphragm, D, at
the far end of the tube. When these two circles are seen to
be concentric the line of vision must coincide with the axis
of the artificial pupil and the diaphragm in question. The
conditions described are those of approximate registration.
In practise, registration is secured by a brief exposure of the
eye of the subject to the test light — which can and should
be very dim — during which exposure he adjusts his head so
that the two circles are seen as concentric, fixation being
directed to the center of the inner one. The head is then
held firmly in position, the test lamp is extinguished, and the
eye is given a period of rest sufficient to remove the effects
of the faint stimulation which it has thus received. When
the absence of such effects is insured the shutter is raised
and the stimulus proper is exposed. When the observation
is completed the shutter is again dropped, the test lamp
lighted, and the state of concentricity or eccentricity of the
circles is again noted. The author has found that his sub-
jects have little difficulty in maintaining practically perfect
registration in this way during periods of several minutes'
duration, a simple head and chin rest being employed.
The diaphragm, Z), acts to prevent the relatively small
stimulus field from illuminating the walls of the tube, al-
though it does not interfere with such illumination by the
registration lamp. The interior of the tube should, of course,
be painted black, and the current supplied to the lamp should
pass through an adjustable resistance so that the intensity
of the light may be easily reduced to a minimum. The
apparatus as figured is not applicable to maximally large
stimulus fields.
Neither of the two methods described above can be
relied upon to give perfect registration unless a possible
anatomical eccentricity of the natural pupil is taken into
consideration. According to Gullstrand,1 such eccentricity
is the rule rather than the exception. Usually, however,
1 A. Gullstrand, Appendix to Helmholtz's " Handbuch der physiologischen
Optik," 3d ed., Vol. i, 1909, pp. 270-272.
176 LEONARD T. TROLAND
the lack of concentricity is not marked, so that if the artificial
pupil employed is relatively small — say one half the diameter
of the natural pupil — registration of the line of sight, which is
obtained by the methods in question, may be relied upon for
practical purposes.
In careful work, however, the pupils of the subjects should
be measured to determine their eccentricity. Registration
may then be effected by having each subject so place his
eye that the inner circle is eccentric with respect to the
diffusion spot in such a direction and to such a degree as to
correct for the eccentricity of the pupil.
One way of testing the concentricity of the natural pupil,
which will at the same time educate the subject in the amount
of eccentricity necessary to correct the registration of each
eye, is the following. A series of trials may be made in which
the stimulus is viewed through the artificial pupil. Each
time the head is adjusted by trial and error so that the stimu-
lus field appears as bright as possible, a position being found
which is as close as may be to the center of the range of
maximum brightness, then this range has an appreciable
magnitude. When this position has been secured the head
is held rigidly against the head-rest and the registration
lamp is dropped into place, the degree of eccentricity of the
two circles being noted and recorded. The average of a
number of such determinations may be used for correcting
the registration.
The diffusion-circle method of securing registration per-
mits a very simple qualitative test as to the adequacy of the
register in a given instance, since the outline of this circle is
determined by the effective pupil. If the registration is
inadequate, fluctuations in the outline of the circle will be
apparent, due to the pulsating contractions and expansions
of the natural pupil. The writer has found that with an
artificial pupil of two millimeters (diameter), using a small
but bright stimulus in a dark room, registration of the line
of sight is adequate for most subjects, although in no subjects
which he has .examined has such registration proven itself
perfect.
THE TEMPORAL RELATIONS OF MEANING AND
IMAGERY
BY THOMAS VERNER MOORE
Catholic University of America
I. THE PROBLEM
The experiments here reported constitute a part of a more
extensive study of memory and perception, which will prob-
ably be made public in the future. The work was done in
the laboratory of Professor Kiilpe at Munich. The part
now published cannot, however, be properly evaluated with-
out some indication of the nature of the results obtained in
the first section of the more extensive study. This first part
consisted in an introspective investigation of the mental
processes involved in perception and recall.
The material for experiment in the unpublished section
consisted of spoken words, printed words, printed pictures
and real objects. A series of eight words, pictures or objects
were presented to the subjects. Their task was to repeat
what they had seen or heard and then to give an introspective
account of the mental processes they had experienced during
the perception of the series and during their attempts to
reproduce the same from memory.1 The subjects were asked
particularly to give an account of the temporal sequence of
events as they had experienced them.
It was rather remarkable that in perceiving, the first
thing in consciousness was reported as meaning the second
some kind of imagery. Whereas in repeating the first thing
was often an image whose meaning was understood and then
designated by a word.
A few introspections will bring out more clearly what is
meant by this assertion.
1 A fuller description of the details of the^technique will be given when the entire
work is made public.
177
178 T. V. MOORE
PERCEPTION OF PRINTED WORDS
"I notice now a certain regularity in this process. With
the first word, the meaning appeared with the reading, with-
out any clear visual image of the object thereby designated.
The same process takes place on the continuation of the
series of words. Gradually it goes on so rapidly that during
the period of exposition (2 seconds) there is time to apprehend
a goodly number of apperceptive complexes, which become
associated with the imaged object. The steps in the process
— so far as I can notice them — are:
"i. Apprehension of the meaning.
" 2. Imagery of the object — generally by means of memory
images.
"3. Associations which are connected with the object."
Subject Lehner, Nov. 17.*
PERCEPTION OF PICTURES
"I look at the picture and generally have its meaning at
once. Often I am not entirely certain, e. g., spoon or trowel.
When I have the meaning, its naming follows immediately."
— Griininger, Dec. ly.2
"In the perception of the several pictures, I notice that I
experienced auditory-motor words in immediate connection
with them, and that these words followed with varying
rapidity the individual pictures. It lasted some time till I
got the word l Mitre.' In this experience it appeared to me
that the rapidity with which the word comes, does not depend
as much upon the finding of the words as it does upon the
1 Ich merke jetzt eine gewisse Gesetzmassigkeit des Prozesses. Bei dem ersten
Wort tritt mit dem Lesen die Bedeutung bewusst auf, ohne deutliches Gesichtsbild
des darin fixierten Objektes. Derselbe Vorgang vollzieht sich bei der Fortsetzung der
Reihe, allmahlich mit so grosser Schnelligkeit dass wahrend der Exponierungszeit
noch Zeit bleibt eine ganze Fiille von Apperceptionsmassen bewusst zu erfassen, die
sich an das vorgestellte Objekt noch kniipfen. Die Stufen so weit ich sie bemerken
kann sind: i. Erfassung der Bedeutung. 2. Vorstellung des Objektes, gewohnlich
durch Erinnerungsbilder. 3. Associationen die sich an das Objekt kniipfen. (lyten.
Nov.)
2 Ich sehe das Bild and und meistens habe ich sofort die Bedeutung. Manchmal
bin ich nicht ganz sicher z. B. Loffel oder Kelle. Wenn ich die Bedeutung habe, folgt
sofort die Benennung.
MEANING AND IMAGERY i?9
recognition of the picture. It is on this account that I would
willingly have looked longer at the pictures. The words
served as designations for the pictures or if you will the
objects represented by the pictures, and had another sense, a
more general meaning than their relation to the individual
pictures or their objects." — Subject Kiilpe, Nov. I4.1
REPETITION OF OBJECTS
"On repeating, there comes to me all of a sudden a visual
image. When this image comes promptly it is usually
complete. But when I must think awhile, there comes to me
first of all something striking in the object. Then come
further qualities, e. g., to the color the form. As soon as this
process of supplementing has developed to a certain point,
the meaning is all of a sudden present. As soon as I have
the meaning, the object seems to become still clearer. E. g.:
All of a sudden I see the typical lustre of a pearl. Then
there comes to me the round form and then all at once I
know what it is." — Subject Griininger, Dec. io.2
Were meaning in some manner identical with imagery,
or were it produced by imagery or the imaginal context of a
sensation as Titchener suggests is often the case, we should
expect just such introspections as this from our subjects —
not however for memory but for perception. That they are
1 Ich bemerke dass ich bei der Wahrnehmung der'einzelnen Bilder sofort akustisch-
motorische Worter in Anschluss an sie erlebt habe, und dass diese Worter in verschie-
dener Geschwindigkeit sich an die einzelnen Bilder anschlossen. Bei dem Wort Bischofs-
miitze, z. B., dauerte es ziemlich lang bis ich es fang. Dabei schien die Geschwindig-
keit des Auftretens der Worter nicht sowohl in der Wortfi ndung selbst als vielmehr in der
Erkennung des Bildes begriindet zu sein. Damit hangt es zusammen dass ich einige
Bilder gerne langer betrachtet hatte. Die Worter galten als die Bezeichnungen fur
die Bilder bzw. die Gegenstande die in ihnen dargestellt waren, und hatten einen an-
deren Sinn, eine allgemeinere Bedeutung als die Beziehung auf die einzelnen Bilder oder
ihre Gegenstande. (i4ten Nov.)
2 Beim Hersagen taucht einfach ganz plotzlich ein optisches Bild auf. Wenn das
Bild schnell auftritt, dann ist es meistens vollstandig. Wenn ich einige Zeit suchen
muss, dann taucht zuerst etwas besonders auffalliges am Gegenstand auf. Dann
kommen weitere Qualitaten, z. B. zur Farbe die Form, und sobald diese Erganzung
einen grosseren Grad erreicht hat ist die Bedeutung auf einmal da, und sobald ich die
Bedeutung habe, scheint mir der Gegenstand noch deutlicher zu werden. Z. B. Ich
sehe auf einmal den eigenartigen Glanz der " Perle." Dann kommt mir die runde Form,
und dann auf einmal weiss ich was es ist. (loten Dez.)
i8o T. V. MOORE
found in memory and not in perception is strong evidence
against any such theory. Here the nature of the occurrence
points to the fact that an image as such means nothing just
as Professor Titchener himself claims. It must be inter-
preted. It can be interpreted only when sufficient data is
present. When this is the case, the subject knows what it is.
This knowledge of what the image represents is not reported
as a sensory element added to the elaboration of the image.
A new image would itself have to be recognized. The
interpretation of the image is a knowing. It is something
which follows the awareness of the image just as understand-
ing follows the sensations involved in perception.
REPETITION OF PICTURES
"The repetition took place in this manner: First I thought
of the first member of the series. Then without holding more
strictly to the order of perception each word was spoken
following an imaginal representation of the pictures. When
I stopped, I attempted to bring up to myself the series. Only
by the rising up of a visual image did I obtain a new word."
—Subject Kiilpe, Nov. I4.1
Such introspections as these suggested a further investiga-
tion. The subjects had noticed a certain sequence of events in
the process of perception. Would it be possible to react to
the events that had been noted? If meaning comes before
imagery in the perception of printed words, would it be
possible for the subject to react, now to imagery and now to
meaning? And if so, what would be the quantitative results?
In the experiments here reported this problem was at-
tempted, to investigate, namely, by means of reaction time
the temporal relations of meaning and imagery in the per-
ception of printed words and pictures. The experiments
were made in the psychological laboratory of Professor
1 Das Hersagen geshah so dass ich mich zunachst auf das erste Glied der Reihe
zuriickbesann. Danach wurden die einzelnen Worter im Anschluss an die anschau-
lichen Vorstellungen der Bilder ohne die Ordnung der Wahrnehmung strenger einzu-
halten ganannt. Wenn ich stockte, suchte ich mir die Reihe wiederzuvergegen-
wartigen und bekam erst durch eine neue Auftauchung des Vorstellungsbildes ein neues
Wort (Nov. I4th).
MEANING AND IMAGERY 181
Kiilpe in Munich during the winter semester of 1913-14
and the summer semester of 1914. The author wishes to
take this opportunity to thank Professor Kiilpe for his great
kindness, for his interest and suggestions, and for the sacrifice
of his time as subject.
II. METHOD OF RESEARCH
The words and pictures used in these experiments desig-
nated simple familiar objects — all capable of being visualized;
e. g., tree, lamp, knife. Abstract words, prepositions, etc.,
were not used, in order that conditions might be as favorable
as possible for the development of imagery. Had such words
been used the difference that was found in reaction time for
meaning and imagery would have been much greater. The
use of such words would indeed have been justified. For if
sensations and images must explain all meanings they must
be involved, and exclusively involved, not merely in the
perception of things that can be immediately sensed, but also
in more abstract mental content. In order, however, to test
the theory on the ground where it is best able to stand, it
was concluded to forego the use of any words except those
that represented familiar sensory objects.
The accompanying plates give an insight into the material
used in these experiments. Most of them represent objects
that can be named by a one or two syllable German word.
The words used were printed on cards in a large legible type.
The use of control words and drawings enabled one to be
sure that the subjects were actually reacting to meanings.
The controls used in the series of words were nonsense
combinations of letters forming one or two syllables. The
controls used in the series of pictures were meaningless
drawings. In general, the subject was instructed to react
(by releasing a telegraph key) in case the word or the drawing
represented some real object. The words were exposed by a
combination memory and tachistoscope apparatus. The re-
action times were measured by a Hipp chronoscope. This
was controlled by a pendulum constructed in accordance with
a design by Professor Kiilpe. The variable error in the
182
T. V. MOORE
MEANING AND IMAGERY
183
i84
T. V. MOORE
MEANING AND IMAGERY
1 86
T. V. MOORE
MEANING AND IMAGERY
187
1 88
T. V. MOORE
MEANING AND IMAGERY 189
chronoscope was negligible — averaging less than 3 cr. The
constant error. was about 700-. Nine subjects took part in
the experiments. A preparatory signal (i->£ sec.) was given
verbally with the aid of a stop watch.
III. SiMPLE1 MEANING AND VISUAL IMAGERY
(a) Quantitative Results
The instructions to the subject in this experiment will
indicate the precise nature of the problem. They are repro-
duced without translation. The subject read them over at
the beginning of each period. A few trial periods were
necessary for some subjects in order that they might learn
not to react to the control word. These preparatory series
were not included in the final results. One of our subjects
(Gl.) never did get free from erroneous reactions and his
results show a marked difference from the others.
Sie werden nach einem Signal ein Wort zu sehen (bzw.zu
horen) bekommen. Ich bitte Sie zu reagieren wenn Sie das
Wort verstanden oder seine Bedeutung erfasst, bzw. wenn
Sie eine Gesichtsvorstellung von dem durch das Wort bezeich-
neten Gegenstand gehabt haben.
Die Worte ' Bedeutung' und 'Vorstellung' werden Ihnen
angeben ob das eine oder das andere verlangt wird. Nachher
bitte ich mir kurz das Erlebnis zu charakterisieren, and dabei
anzugeben, ob die aufgetauchte Vorstellung an die Stelle
der Bedeutung gesetzt werden konnte, etwa bloss die konkrete
anschauliche Erfullung dessen war, was in der Bedeutung
abstrakt intendiert wurde.
In this series, therefore, the subject reacted either (a) To
the awareness that the word had a meaning, or (b) To the
awareness of the visual image of the object.
If there is no difference between meaning and the visual
image of an object represented by a word the average of the
two series should be approximately the same. The subject
ought not to be able to distinguish meaning and imagery
1 By 'simple meaning' is not meant an absolute simplicity. The word is used to
contrast this set of experiments with a later one where the more complex consciousness
of purpose was required.
J90
T. V. MOORE
and this should manifest itself in averages for the two sets of
reactions that approached each other within the limits of
experimental error. If meaning, however, is produced by
or is identical with the visual image which accrues to the
sensations involved in the perception of the word, the image
series should be shorter if anything than the meaning series.
The results are given below. The tables are clear without
any explanation, except perhaps, that in column T is given the
SUBJECT Gl
Words
Date
Visual Image | T
v
Date
Simple Meaning
T
V
Zange
1,182
628
Rechen
368
154
23m
Fernglas
Pfeil
563
233
9
321
23/VI
Besen
Gabel
563
751
229
Messer
639
85
Truthahn
686
164
26m
Lampe
Sichel
640
660
86
106
26m
Esel
Trichter
676
550
154
28
Sabel
472
82
Dreick
602
80
7/VII
Stuhl
Eimer
527
429
27
125
Nest
Aal
662
558
140
36
Baum
472
82
Geige
338
184
A7TT
Trichter
346
208
Fernglas
597
75
IO/VII
,
Korb
482
72
7/VII
Besen
424
98
12)6,645
1,831
Korb
462
60
554
152
Maske
249
273
io/VII
{ Meissel
171
15)7,837
1,887
Mean =522
126
Median = 563
Reactions to visual imagery equal or below median = 8.
SUBJECT GRUNINGER
Words
Date
Visual Image
T
v
Date
Simple Meaning
T
V
II/II
Auge
Ballon
902
,092
341
151
II/II
/ Schuh
\ Nase
682
787
18
87
Kuh
,660
4J7
Fass
661
39
I6/II
Bohrer
,277
34
I6/II
Schwan
659
41
Sofa
,261
18
Schaf
781
81
Stiefel
,229
H
[Ring
531
169
25/11
Schuh
Fass
,I03
1,462
140
219
2SAI
Auge
Ballon
680
705
20
5
Nase
1,200
43
Fass
660
40
,Kuh
856
156
9)11,186
i,377
10)7,002
656
1,243
153
Mean = 700
6<>
Median = 68 1
Reactions to visual imagery equal to or below median =o.
MEANING AND IMAGERY
191
SUBJECT KULPE
Words
Date
Visual Image
T
V
Date
Simple Meaning
T
v
9/n
{ Lowe
1,690
790
9/n
Kerze
517
H
13/11
Ballon
Auge
944
675
44
225
13/11
Schuh
Ring
538
434
7
97
Rose
1,097
197
Nase
799
268
16/11
Fliege
Veilchen
802
753
98
H7
i6/II
Dampsfchiff
Fass
540
573
9
42
Kuh
848
52
Ballon
708
177
23/11
Kerze
Schuh
839
897
61
3
23/II
Fliege
Schuh
420
577
in
46
Nase
. 813
87
Lowe
607
76
18/5
Ring
Dampsfchiff
800
649
100
251
i8/V
|
Veilchen
Rose
375
460
156
7i
12)10,807
2,055
Kuh
363
168
900
171
13)6,911
1,242
Mean =531
95
Median = 540
Reaction to visual imagery equal to or below median =o.
SUBJECT LEHNER
Words
Date
Visual Image
T
v
Date
Simple Meaning
T
V
f Rechen
38i
263
Schwan
445
24
I6/II
1 Buch
| Rettich
1,131
611
487
33
I6/II
Bohrer
Sofa
1,007
663
538
194
[Nase
604
Rose
317
152
Tnchter
584
60
Brief
691
222
Dreieck
490
154
' Palme
500
31
Blatt
597
47
Baum
399
70
30/VI
Bar
Krone
606
479
38
Hirsch
Spaten
389
406
80
63
Veilchen
488
156
30/VI
.
Aal
407
62
Treppe
631
13
Krug
395
74
Ofen
572
72
Tasse
489
20
Sabel
588
56
Schadel
426
43
Geige
461
183
Ohr
435
34
3/VII
Pfeil
Meissel
Fernglas
Flasche
874
758
193
129
230
114
3/VII
'
Spaten
Hirsch
Baum
Palme
371
483
524
49
98
14
55
Krug
691
47
'Bar
368
101
7/VH
.
Tasse
Schadel
736
918
92
274
7/VII
Krone
Ofen
384
429
85
40
Ohr
844
200
Veilchen
344
125
Ballon
803
159
Treppe
398
23)14,813
3,205
23)10,788
2,245
644
139
Mean =469
97
Median =429
Reaction to visual imagery equal or below median =i.
I92
T. V. MOORE
SUBJECT MAREZOLL
Words
Date
Visual Image
T
V
Date
Simple Meaning
T
v
ii/V
Ohr
Zwicker
1,308
1,633
221
104
ii/V
Kafer
Kerze
841
638
254
051
Eule
1,598
69
' Storch
831
244
I2/V
Schnecke
Pfau
1,374
2,721
155
1,192
I2/V
Katze
Kuh
576
822
on
235
Kafer
3,868
2,339
Schlange
761
174
I4/V
Maus
Schaf
2,610
1,309
i, 08 1
220
I4/V
Lowe
Stier
774
458
187
129
Hirsch
1,222
307
Geier
325
262
Zirkel
I,2IO
319
Ohr
508
079
I6/VI
Eimer
Stuhl
1,021
947
508
582
i6/VI
Eule
Schnecke
398
676
189
089
Sichel
870
659
Schlange
582
005
Kerze
1,187
342
Storch
693
106
Geier
1,367
162
i
Kuh
650
073
Stier
988
541
I9/VI
Kerze
465
122
I9/VI
Lowe
Schlange
1,607
1,292
78
237
Katze
Maus
684
456
097
Kuh
*,57!
42
Pfau
436
151
Katze
1,146
383
Schaf
697
1 2O
Storch
1,258
271
Hirsch
384
203
30/VI
Zirkel
Sichel
406
594
181
007
Stuhl
425
162
Eimer
557
030
21)32,107
9,812
25)14,687
3,292
1,529
467
Mean =587
152
Reactions to visual imagery equal or below median = o.
SUBJECT MOORE
Words
Date
Visual Meaning
T
V
Date
Simple Meaning
T
V
Schuh
839
330
Finger
5l6
53
Sofa
1,685
516
Buch
337
126
9/VI
Rechen
Auge
1,112
1,448
57
279
9/VI
Schere
Biirste
444
572
19
109
Lampe
Dampsfchiff
1,145
857
24
312
Fernglas
Uhr
563
401
100
62
Ochse
J,553
384
Schuh
484
21
I2/VI
Frosch
1,009
Tiger
758
295
Kamm
794
375
I2/VI
Vogel
259
204
Biirste
8i5
354
Rechen
386
77
,
Pinsel
1,091
78
Fernglas
396
67
Besen
1,108
61
Sofa
245
218
Handbeil
1,188
19
Kamm
572
109
30/VI
Vogel
Tiger
i'27S
I4/VII
Auge
Frosch
583
436
120
27
Uhr
1,209
40
Biirste
44°
23
I4/VII
\
Schere
f Schuh
( Rechen
937
i,44S
i,544
232
276
375
Dampfschiff
iv/r
478
IS
17)7,87°
ao r\ A A'J
i,645
ofi
19)22,209
3,992
.L > .L V_ 0. 1 1 H"V/J
Median = 444
9°
1,169
210
Reactions to visual imagery equal or before median = o.
MEANING AND IMAGERY
'93
SUBJECT SCHERREN
Date
Visual Image
T
V
Date
Simple Meaning
T
v
22/VI
Krebs
Stern
9,807
4,537
5,i40
130
I3/VII
Fliege
Facher
656
988
104
228
Herz
3,370
1,297
Finger
846
86
I3/VII
Trommel
Spinne
6,505
4,587
1,838
80
Herz
Fliege
788
1,034
28
274
Koffer
2,53i
2,136
Stern
i, 080
320
Flasche
3,159
1,488
Krebs
963
203
Hirsch
3,4H
1,256
Koffer
552
208
20/VH
Sabel
Traube
5,354
5,402
687
735
2I/VII
Spinne
Trommel
697
561
63
199
Engel
5,340
673
Sabel
631
129
Facher
2,007
2,660
Flasche
969
209
Engel
378
382
Traube
501
259
12)56,010
18,120
14)10,644
2,692
4,667
i,5io
Mean =760
192
Median = 742
Reaction to visual imagery equal or below median = o.
SUBJECT STAPPEN
Words
Date
Visual Image
T
V
Date
Simple Meaning
T
V
Ring
572
87
Veilchen
604
63
12/11
Auge
Ballon
6l3
584
46
75
12/11
Fass
Kerze
743
5H
202
27
Stiefel
656
3
Lowe
509
32
Schaf
Rettich
634
1,076
25
417
17 II
Rose
Kuh
497
423
44
118
I7/H
Nase
Fass
!6s
600
94
59
Haken
Lilie
573
507
32
34
Rechen
^Nest
657
874
2
215
I9/H
Storch
Sichel
I11
614
30
73
Sage
707
48
Lampe
516
25
I9/H
•
Maske
632
27
Sofa
485
56
Uhr
Sl8
141
Esel
537
122
14)9,225
I,36l
12)6,496
736
659
97
Mean =541
61
Median = 5 13
Reactions to visual imagery equal or below median = o.
reaction times in thousandths of a second. Under V, the
variations from the mean.
At the bottom of each column the mean reaction times
and mean variations have been calculated. The median for
the reaction times to meaning have also been determined.
With but one exception, our nine subjects show a marked
difference in their reactions to meaning and imagery. The
i94
T. V. MOORE
SUBJECT T
Words
Date
Visual Image
T
v
Date
Simple Meaning
T
V
Kette
1,129
58
Puppe
747
255
Nest
1,162
25
Bretzel
794
302
Trichter
i,774
587
Mond
754
262
Stuhl
890
297
Schadel
546"
54
Anker
1,306
119
Taube
884
362
27/VH
Ballon
Auge
i,257
i, 216
70
29
27/VII
Zirkel
Hirsch
753
520
261
28
Apfel
905
282
Engel
569
77
Eimer
1,295
108
Besen
499
7
Ofen
1,109
78
Pferd
517
25
Kuh
995
192
Hund
498
6
Tiger
1,260
73
Geige
464
28
Fass
1,132
55
Schiff
667
175
Strumpf
879
308
Tiger
347
H5
Schrank
1,263
76
Schnecke
453
39
Schlange
866
321
Pfau
554
62
Facher
1,259
72
Fliege
470
22
Herz
i,845
658
Fahne
441
51
Hahn
947
240
Finger
793
301
28/VII
Krebs
2,176
989
28/VII
Stern
369
123
Spinne
329
Trommel
335
157
Hase
1,072
Koffer
525
33
Schlitten
906
281
Nase
527
35
Feile
354
Schliissel
556
64
Hand
900
287
Hammer
479
13
Handschuh
871
316
Spaten
411
81
Eule
999
188
Bar
401
91
Kirsche
921
266
' Klavier
663
171
Lowe
802
385
Kerze
419
73
Leiter
1,415
228
Bleistift
38i
in
Fernglas
1,677
490
Nest
140
Wiege
1,336"
149
Wage
367
125
Wurst
993
194
Bohrer
307
185
29/VII
Loffel
806
38i
29/VII
Fliege
304
188
Haken
2,115
928
Trommel
376
116
Veilchen
1,315
128
Brille
362
130
Treppe
919
268
Lilie
369
123
Bretzel
779
408
Blatt
303
189
Turm
894
293
Biirste
312
1 80
Trichter
1,030
157
Truthahn
301
191
40)47,472
10,782
40)19,689
4,98i
1,187
269
Mean = 492
122
Median = 474
Reactions to visual imagery equal or below median = o.
one exception is not to be explained by individual difference
in mental type, but rather by an anxiety to react as quickly
as possible. At first he reacted to every nonsense word.
He was then tried with pictures. Here again, every meaning-
less drawing elicited a reaction in spite of instructions. The
reaction times at first varied around 100 a. Later he was
MEANING AND IMAGERY
'95
asked to wait each time and make a judgment that he had
fulfilled the task given him. Even under these instructions,
he continued to react occasionally to nonsense words — the
following reaction being very much retarded. He finally
gave up the experiments. What would have resulted had he
by practice become entirely free from erroneous reactions
one cannot say. It would seem, however, more fair to a just
conclusion to exclude rather than include his results in our
summary. Leaving aside the results of this subject, the
reaction times to visual imagery were all but one above the
median of the reaction time to meaning. It is worthy of note
that this single exception is the first recorded reaction of this
subject. (He had made several practice series before.) Our
subjects made 150 reactions in all. Were it merely a matter
SUBJECT MAREZOLL
Worter
Visual Image
Simple Meaning
D
Ohr. .
1,308
508
+ 800
Eule
i 508
398
+ I,2OO
Schnecke
J -274.
676
+ 698
Pfau
2,721
436
+ 2,285
Kafer
3,868
841
+ 3,O27
Maus
2,610
T*
456
+ 2,154
Schaf .
i, 300
J3
697
+ 612
Hirsch .
1,222
384
4- 838
Zirkel .
I,2IO
406
+ 804
Eimer
I, O2 1
cc7 •
~i~ 464
Stuhl
Q4-7
42 C
+ 522
Sichel
870
CQ4
+ 376
Kerze
1,187
638
ji
4- 1:40
Geier
1,367
325
3T7
T 1,042
Stier
088
4C8
4- C-2Q
Lowe
yoo
I 6O7
m^mrcw
->J
+ QT?
Schlange
I 2Q2
f
76l
+ 531
Kuh.
1, 571
822
4- 7cn
Katze. . .
1,14-6
576
+ C7O
Storch
1,258
y
693
+ 565
20)30,474
",425
19,259 '
15237
571.2
962.9
of chance that reaction times to imagery should be longer than
those to meaning we could find about 75 longer and 75 shorter.
As a matter of fact, we find 149 longer and only one shorter.
In spite then of the rather small number of reactions (condi-
tioned by taking the introspective reports) there is over-
196 T. V. MOORE
whelming evidence to show that something more than chance
has to do with the difference in reaction time to meaning and
imagery. This difference is not due to the words used for
imagery and meaning. Not only were the words in both
cases representative of sensory objects but care was taken to
repeat the same words in the two series. A table is given
above comparing the reactions of one subject for meaning
and imagery to the same words. Under D is given the
difference between the two. The reaction time to imagery is
always longer than to meaning. With some of our subjects
the results are not so unanimous, the meaning reaction being
occasionally longer. This is to be explained mainly by the
effects of practice, though something is no doubt due to
accidental variation.
(b) INTROSPECTIVE DATA
From the quantitative results that have just been given,
it is evident that the subjects give a different response when
told to react to meaning or imagery. Were we to stop with
the quantitative results we would not know very much about
the nature of that difference. Is meaning simply an early
stage in the development of the image? Is it a vague con-
fused image? Is it merely the realization of the power to
visualize the object? Is it the tendency of a number of
images to crowd into consciousness? What is the difference?
There can be no doubt that a considerable difference exists
and it is of great importance to find out precisely what it
is. This can be done by an examination of the subjects'
introspective reports of what they experienced during their
reactions. The reports were taken down by dictation imme-
diately after the reaction and then re-read to the subject to
insure their accuracy. The originals are in German and
will be given in German and in English when a complete
account of all the experiments is published.
MEANING AND IMAGERY
197
CONSCIOUSNESS OF MEANING
(1) The meaning has a general character.
Kerze: " There came to me at once the
word 'Light.' This was not a determina-
tion of the meaning, but only another
word for it. The meaning was entirely
general, as if I should say a candle, that
is, any candle — every possible candle."
-Kiilpe, 9/H.
(2) The universality of the meaning is
not always absolute.
Ring: "As soon as I saw this word, I
experienced an auditory motor stimulus,
and immediately in connection therewith
the understanding of that which the word
signified. This was quite universal with-
out being related to anything in partic-
ular— except the limitation to 'finger-
ring.' I am distinctly conscious that a
finger-ring was intended. I cannot re-
member an image of any such ring."
—Kiilpe, 13/11.
(3) The meaning is at times felt to be in-
complete, because of an unanalyzed con-
sciousness of what the word signifies.
Schere: "At first, a feeling of familiar-
ity was present and then a feeling of cer-
tainty that I know what the word sig-
nifies without having analyzed its mean-
ing any further. First, during the reac-
tion itself there came the further thought
'something with which one cuts.'"
— Moore, 9/VI.
Eule: "I knew that the word was some-
thing with which I am familiar and knew
that from this point I could, at any time,
go on and find its more specific meaning.
Thereupon I reacted. In the word itself
there was something presented to con-
sciousness (mir gegeben) that I cannot
further describe."— Frl. Marezoll, io/VI.
CONSCIOUSNESS OF VISUAL IMAGERY
(1) The image is particular.
Sofa: " I have a rather good image and
I did not pronounce the word. I see with
great precision the brown cobr and the
form of the object — but not of the entire
object. I could derive several concepts
from this one image. It looks like a large
reclining chair. The image would not do
for all sofas." — Griininger, i6/II.
(2) The image is at times schematic.
Herz: "I read the word 'heart' and
apprehended its meaning. I remembered
my task and sought after an image. I
projected over the place of the card a
heart of regular mathematical propor-
tions. Only the contours were imaged,
and these by such an airy line that I
question myself whether I had a visual
image at all or whether it was an ideal
construction, such as one ca'rries out
in mathematical thinking." — Scherren,
I3/VII.
(3) The image is at times incomplete in
a different way.
(i) It is partial.
Rechen: "First, the meaning, then the
image. Nevertheless, I reacted before the
image was clear. I imaged a part of a
rake. Already I have noted several
times that I image the left lower parts of
objects. Here I imaged a wooden rake."
— Stappen, 17/11.
(ii) It represents a single definite char-
acter of an object.
Kuh: "I have the meaning and now I
must have an image. I then look at an
empty spot — no longer at the word.
Then there appears the color of the ani-
mal. I see 'brown.' But a satisfactory,
complete image, I do not obtain. I must
exert myself even to obtain the color. I
could not take the image for the meaning.
I cannot read anything more out of the
image than 'brown' — never the meaning
'cow.'" — Griininger, i6/II.
i98
T. V. MOORE
CONSCIOUSNESS OF MEANING
(4) The meaning never has sensory
characteristics but rather a conceptual de-
termination.
Geier: "A moment passed before I
found the meaning. No auditory-kinses-
thetic image was present. I knew that it
was something that hovers over moun-
tains in the air — even though I did not
see the mountains. Visually I imaged
only a pair of extended wings and knew
that something belonged between them."
— Frl. Marezoll, I4/V.
Veilchen: Immediately after the word
appeared, I had an auditory-kinaesthetic
image of it — as I pronounce it. 'Veil-
chen,' and in connection therewith a
knowledge of its meaning (Ein allge-
meines Bedeutungswissen), that I can
thus explain: a definite species of flower.
I dare say that it is this which makes up
the content of the meaning — what I actu-
ally know about this object during the ex-
periment.— Kiilpe, i8/V.
(5) The meaning is often expressed in
terms of a definition of general applica-
tion.
Dampfschiff : Immediately on the expo-
sition of the word, auditory-kinsesthetic
image thereof, and a realization of the
meaning in the sense of 'a means of trans-
port by water.' This time there was no
trace of any image. — Kiilpe, i6/II.
(6) The meaning is never localized.
CONSCIOUSNESS OF VISUAL IMAGERY
(4) The image manifests degrees of
brightness, color and clearness.
Rose: "Immediately after the word
came I had the auditory-kinsesthetic
image of the word and thereupon an
understanding for its general significa-
tion. Then first came an image — the
image of a blossom. Almost nothing of
the stem was seen. Colorless, mere dif-
ferences of brightness in the blossom
and the leaves were perceived. A full
blown rose. The common form. Image
and meaning did not cover each other."
—Kiilpe, I6/II.
Krug: Meaning then the visual image.
It was an earthenware jug, bellied out in
front — antique as if it had just been dug
up. — Lehner, 7/VII.
(5) The image is often described in sen-
sory terms that would fit only a very definite
object.
Rettich : This time there came to me the
image of a radish of medium size. I saw
clearly the little hollows in its skin filled
with dirt and myself in the attitude in
which I cultivate this beautiful variety in
my wife's garden. All at once there
came to me a poem of Morike. It is en-
titled 'The Radish.'— Lehner, i6/II.
(6) The image has often a definite
position.
Schuh: I had an indistinct image of a
laced shoe — the point to the right some-
what behind the plane of the word. A
confused consciousness of meaning was
also immediately present, which did not
coincide with the image. The meaning
was even more general than foot covering.
It had somewhat the sense of a piece of
clothing without relation to a part of the
body. — Kiilpe, 23/11.
MEANING AND IMAGERY
199
CONSCIOUSNESS OF MEANING
(7) The meaning is always pertinent to
the word.
(8) The meaning is never looked upon as
superfluous.
(9) The meaning is always present.
(10) The meaning leads regularly to the
image.
Nearly always the subjects report the
meaning as coming first.
• CONSCIOUSNESS OF VISUAL IMAGERY
(7) The image is sometimes recognized
as not strictly pertaining to the word.
Rettich: The word appeared very
strange to me. I think I read something
like 'Bettish.' Only later did I get the
correct meaning. There came a visual
image. The image did not really repre-
sent a radish but rather a kind of turnip.
(8) The image is often regarded as un-
necessary and of secondary importance.
Ochs: I first understood the word as
something familiar, as something that I
knew what it was. A further analysis of
the meaning did not take place. Under
the influence of the task, my attention
was directed to experiencing an image and
then arose the head of an ox with his
horns as drawn in the pictures for these
experiments. — Moore, I2/VI.
Fass: Immediately after looking at the
word an auditory-kinsesthetic representa-
tion and understanding of its general sig-
nification in the sense of a spatial meas-
ure. There came also — altogether fleet-
ingly a weak image with a pair of hoops
lying on the ground — wholly accessory as
if a schema. — Kiilpe, i6/II.
(9) The image is often lacking.
Such cases could be multiplied indefi-
nitely. Some have already been given.
(10) The image is only occasionally
present before the meaning.
Kuh: I believe1 the image came first —
wholly undefined. Very soon thereafter
the meaning, immediately after which the
reaction. — Stappen, 17/11.
(n) The meaning comes spontaneously.
Cases enough have already been cited
to make this evident. Only occasionally,
where the word is read incorrectly, is any
effort required to bring out this meaning.
The image must often be sought.
Flasche: I had the feeling of a consider-
ably retarded flow of imagery, and per-
ceived clearly that I was sharply concen-
strated upon my task. I then imagined
that I went through Amalien Street and
1 These cases are very rare. I have found them only with this subject and when
he does mention them, it is always with reserve. He says 'I believe' indicating that he
is not sure of the observation.
200 T. V. MOORE
CONSCIOUSNESS OF MEANING CONSCIOUSNESS or VISUAL IMAGERY
had the task to represent to myself a
bottle. The representation succeeded
but rather poorly. Only the image of
the material (glass) and the long form
was clear. — Lehner, 3/VII.
A careful consideration of these results will show that
the difference between meaning and visual imagery does not
consist in any possible difference in the original imagery itself.
If meaning were an early stage in the development of the
visual imagery, it might be possible to explain in this way
the difference in the reaction times to the two events. A
candid consideration of the introspections shows that this is
not the case. The universality of the meaning cannot be
pictured and is something quite different from the schematism
of the image. The incompleteness of the image with a
fragmentary character and washed out coloring differs pro-
foundly from the imperfect unanalyzed embryonic stage of
the meaning. The image has sensory characters which
cannot be ascribed to the meaning — the meaning cognative
characters which are utterly foreign to the image. The
meaning is a 'knowing' sui generis; the image is a sensational
element with its own specific character.
The meaning is not the potentiality to visualize. It may
have an element of potentiality about it, but it always has
an element of actuality which extends from the unanalyzed
knowledge expressed by the phrase: "I know what that is"
— to the more perfect conception expressed by a definition.
The potentiality of the meaning when present is not the
same for all meanings. It is a definite potentiality in which
the elements of a definition of the object are in subconscious-
ness. It is not the potentiality to visualize, for the poten-
tiality to visualize (i) depends on a meaning to determine
what is to be visualized; (2) results in something different
from the actualization of the meaning. The actualization of
the meaning leads to the consciousness of a definition which
may not even be accompanied by imagery of any kind
whatever.
Nor is imagery the tendency of a number of images to
MEANING AND IMAGERY 201
crowd into consciousness. That tendency is sometimes pre-
sent especially with one of our subjects, but by him it was
recognized as something that came after the meaning.1
Meaning is often present and one is definitely conscious
of it without being conscious of a tendency of images to crowd
into consciousness. Meaning is a consciousness of knowledge
that has definite characters foreign to the images that tend
to crowd into consciousness. Furthermore, where images do
crowd into consciousness they have to be known. This
knowledge of what the image represents cannot be explained
by another image which would itself have to be known.
Meaning, therefore, appears to be a conscious process
sui generis distinct from imagery.
IV. CONSCIOUSNESS OF PURPOSE AND KIN^ESTHETIC
IMAGERY
(a) Quantitative Results
The instructions to the subject indicate sufficiently the
nature of the investigation in this section of the work. These
instructions were as follows:
Sie werden nach einem Signal ein Wort zu sehen bekom-
men. Ich bitte Sie zu reagieren wenn Sie die Bedeutung des
Wortes im Hinblick auf den Gebrauch oder die Funktion des
damit bezeichneten Gegenstandes erfasst, bzw., wenn Sie
eine kinaesthetische oder kinaesthetisch-optische Vorstellung
davon gehabt haben.
The words chosen for reaction stimuli in this set were not
merely capable of being visualized but represented objects
that most of us have often handled as: brush, bell, hat. To
represent a word like 'lion' by a kinsesthetic image is to some
subjects a very difficult task. Consequently a more appro-
priate set of words was chosen. Even under the most favor-
able conditions the kinaesthetic image comes far too late to
account for the meaning. It might, however, be claimed
that such an image is identical with the consciousness of the
purpose of an object. Accordingly the comparison was made
between reactions to the consciousness of purpose and those
1 Cf. supra, p. 178. Subject Lehner.
2O2
T. V. MOORE
to the awareness of a kinaesthetic image which concerned
the object itself. Mere verbal images were excluded. Seven
subjects took part in this set of experiments.
The results are shown in the accompanying tables:
SUBJECT GRUNINGER
Words
Date
Motor Image
T
V
Date
Concept of
Purpose
T
V
Anker
2,084
210
f Stiefel
2,247
699
2/III
Rechen
1,647
647
2/III
Koffer
,389
159
Gabel
1,929
365
Biirste
,292
256
Trichter
2.379
85
' Flasche
,671
123
Apfel
Wiege
2,307
2,448
13
154
Wurst
Birne
7H
1,255
834
293
Eimer
1,690
604
Brief
,474
74
4/III
<
Kette
Zwicker
2,398
2,101
I04
193
4/111
Leiter
Fahne
,4H
2,255
134
707
Treppe
2,506
212
Finger
i,458
90
Hund
3,745
1,451
Kuh
i,474
74
11)25,234
4,038
Ente
1,942
394
2,294
367
12)18,585
3,837
Mean = 1,548
320
Median = 1,466
Number of times median for concept exceeds reaction time for imagery = o.
SUBJECT KULPE
Words
Date
Kinsesthetic
Image
T
V
Date
Concept
of Purpose
T
V
27/H
2/III
Ring
\ Sichel
Ring
Rechen
Bleistift
Wurfel
Lampe
Pickel
,100
,416
,332
,823
,227
,579
,298
,638
45i
135
219
272
324
1,028
253
87
27/H
2/III
•
' Stiefel
Auge
> Haken
Hammer
Biirste
Gabel
Feile
Uhr
Stiefel
Mea
Med
804
1,180
771
1,822
1,346
1,165
1,153
808
944
306
70
339
712
236
55
43
302
166
8)12,413
1,55!
2,769
346
9)9,993
2,229
n =1,110
ian = i,i53
247
Number of times median for concept exceeds reaction time for imagery =i.
With all of our subjects the mean for reaction time to
kinaesthetic imagery is longer than that to the concept of
purpose. Examining these results critically we find that
with some of our subjects in spite of the small number of
MEANING AND IMAGERY
203
SUBJECT LEHNER
Words
Date
Kinaesthetic
Image
T
V
Date
Concept of
Purpose
T
v
f Wiirfel
1,083
70
f Feile
997
7i
2/III
\ Lampe
I55i5
362
2/III
Uhr
1,249
323
I Pickel .
1,184
31
Stiefel
1,121
J95
Brief
1,693
540
Wiege
684
242
Leiter
i,H5
8
Eimer
1,050
124
5/ni
Fahne
1,501
348
5/ni
Kette
733
193
Flasche
1,980
827
Zwicker
815
in
Treppe
1,162
9
Finger
1,260
334
Stiefel
760
393
Brief
941
15
Uhr
804
349
Leiter
1,076
J5°
Feile
941
212
Fahne
960
34
23/VI
Wiege
826
327
Flasche
937
ii
Eimer
879
274
23/VI
Treppe
763
163
Kette
882
271
Wiirfel
738
188
Zwicker
927
226
Lampe
721
205
Finger
1,170
17
Pickel
775
I5i
16)18,452
4,264
16)14,820
2,510
1,153
266
Mean = 926
157
Median = 937
Number of times median for concept exceeds reaction time for imagery = 6.
SUBJECT MAREZOLL
Words
Date
Kinaesthetic
Image
T
V
Date
Concept
of Purpose
T
v
14/5
/Buch
\ Klavier
,157
,899
493
249
H/5
•
i Ring
Schere
962
1,215
353
100
Eimer
,148
498
Pinsel
1,942
627
I9/V
Korb
Pinsel
,745
,643
95
7
I9/V
Zirkel
Schlitten
1,966
1,404
651
89
Zirkel
,524
2,874
Schere
2,167
852
Schlitten
,573
77
Eimer
3,°97
1,782
7/VH
Schere
Ring
,071
723
579
927
30/VI
Korb
Buch
*77 f* *
650
857
665
458
Anker
,884
234
Klavier
807
508
Besen
,270
380
Burste
930
385
*9/VII
Bohrer
Burste
Zwicker
,422
,218
,557
228
432
93
7/VH
Bohrer
Besen
Anker
990
749
1,056
&
259
Rechen
,532
118
Haken
939
376
Horn
,069
58i
Meissel
,616
34
17)28,051
7,899
I5)i9,73i
7,996
1,650
464
Mean = 1,315
533
Median = 990
Number of times median for concept exceeds reaction time for imagery = o.
204
T. V. MOORE
SUBJECT MOORE
Words
Date
Kinaesthetic
Image
T
V
Date
Concept of
Purpose
T
V
Bohrer
,453
217
f Handbeil
781
36
I6/VI
Spaten
Haken
,129
,628
S4i
42
I6/VI
Sabel
Pinsel
1,162
592
345
225
Messer
,796
126
Besen
539
278
o /T 7T
Biirste
,772
102
' Nadel
918
101
i8/VI
Zange
,396
274
I8/VI
Rechen
815
2
'Handbeil
,572
98
Feile
1,046
229
9/VH
Pinsel
Pfeil
,914
,222
244
448
9/VH
Bohrer
Spaten
1,004
564
I87
253
Besen
,480
190
Haken
836
19
'Wage
,461
209
Hammer
1,215
398
Brille
,896
226
Handschuh
412
405
Feile
,O7I
401
Loffel
635
182
Kerze
,459
211
Schlitten
1,001
I84
30/VII
Ring
Zange
,234
,722
564
52
3o/vn
Leiter
Schliissel
719
669
98
I48
Bohrer
,564
106
Rechen
967
150
Haken
,888
218
Spaten
822
5
Schere
,729
59
Pinsel
1,028
211
Zirkel
2,014
344
Biirste
621
I96
20)33,400
4,672
20)16,346
3,652
1,670
234
Mean = 817
I83
Median = 818
Number of times median for concept exceeds reaction time for imagery =o.
SUBJECT STAPPEN
Words
Date
Kinsesthetic
Image
T
V
Date
Concept of
Purpose
T
V
28/11
f Kerze
Haken
Biirste
Wiirfel
Sense
4,308
f8
644
465
1,024
2,915
865
749
928
369
28/11
Auge
Sage
Sichel
Gabel
Feile
Me
Me<
1,199
682
697
8I4
362
449
68
53
388
5)6,969
1,393
5,826
5)3,754
m = 750
lian = 697
1,022
1,165
2O4
Number of times median for concept exceeds reaction time for imagery =3.
experiments we can say definitely that the concept of purpose
comes quicker than the kinaesthetic imagery. With these
subjects the median for the concept of purpose was shorter
than everyone of the reaction times to kinaesthetic imagery.
With one subject, one out of eight reactions to kinaesthetic
imagery was shorter than the median; with another two out of
MEANING AND IMAGERY
205
SUBJECT TANNHAUSER
Words
Date
Kinsesthetic
I mage
T
V
Date
Concept of
Purpose
T
V
Bohrer
1,492
231
Klavier
1,314
456
Schliissel
939
322
Burste
749
109
Trichter
,527
266
Fernglas
749
109
Bleistift
,367
106
Haken
1,069
211
Trommel
,088
173
Leiter
829
29
30/VII
.
Brille
,903
642
30/VII
Kerze
885
27
Rechen
,327
66
Pickel
1,071
213
Tasche
,255
6
Zange
1,816
958
Wage
,378
117
Ring
785
73
Hammer
,005
256
Schlitten
600
258
Handschuh
,158
103
Feile
763
95
Besen
,173
88
Spaten
554
304
Dolch
855
406
Brief
554
304
Horn
1,011
250
Anker
739
119
Auge
2,334
1,073
Eimer
1,114
256
Brunnen
Geige
2,216
694
955
567
Hut
Flasche
976
34
118
Fass
2,006
745
Glocke
717
141
Hahn
2,082
821
Facher
648
210
Beil
i>S76
315
Finger
i, 006
I48
3I/VH
Messer
831
430
31^11
Fahne
875
17
Lampe
883
378
Meissel
834
24
Koffer
797
464
Korb
952
94
Buch
1,009
252
Kette
8i7
4i
Hobel
1,007
254
Kamm
816
42
Handbeil
707
554
Schrank
637
221
Schuh
848
413
Sense
776
82
Ofen
849
412
Pfeil
700
IS8
28)35,317
10,665
Nadel
720
138
1,261
"~38i~
29)24,889
4,989
Mean = 858
172
Median = 816
Number of times median for concept exceeds reaction time for imagery =2.
twenty-eight. With one of our subjects the matter looks a
little doubtful; six out of sixteen are shorter than the median.
With another subject1 the results are too few and scattered
to give any quantitative basis for judgment.
The question is one where individual differences are likely
to play a part. Those who readily form kinsesthetic imagery
may be able to obtain such an image more quickly than they
can think of the purpose of the object. To what extent this
is true cannot be decided from the present results.2
1 It was impossible to get more experiments from this subject. He left the day
after the series above reported and did not return in the summer semester. They
are more of the nature of a preparatory series than final results.
2 When a short abstract from this paper was read last December at the meeting
206 T. V. MOORE
Taking all the results together only 12 out of 105 reactions
to kinaesthetic imagery were shorter than the median of the
various subjects' reaction-times to the concept of purpose.
(b) Introspective Data
Turning now to the introspective results, we find that
the concept of purpose and the kinsesthetic image are very
clearly differentiated. The concept of purpose differs from
the simple meaning in that it does not come with the same
necessity. It is the result of the subject's task — not of the
mere exposition of the stimulus. The same is true of the
kinsesthetic image. Both follow upon the awareness of the
simple meaning. Neither is a necessary prelude nor a se-
quence of the other. The task "image" or "concept" is the
main factor in determining which is to appear.
The following are some of the more noteworthy intro-
spective differences between the two.
(1) The concept of purpose is expressed (i) The kinasthetic image always de-
in non-sensory conceptual terms. scribes some kind of act involving a use of
the muscles.
Zwicker:" I imaged my own eye glasses Sichel: "Immediately after the word
and had clearly a consciousness of con- appeared I had an auditory-kinsesthetic
cave glasses. I was further conscious of image of it. Following this I constructed
the fact that these glasses must refract a visual and weak kinaesthetic image
the rays of light according to a definite thereof in this manner. I held a sickle in
law that the image may still fall upon the my right hand and made movements
retina — even though the lens is incapable therewith as if I were cutting grass.
of doing it this service. I then formu- Thereupon I reacted."— Kiilpe, 2/III.
lated the purpose of eye-glasses as: 'The
correction of an error of refraction.'" —
Lehner, s/III.
(2) The concept of purpose sometimes (2) This was not noted in the description
involves the consciousness of the relation of of the kinasthetic imagery.
the object to other things.
Gabel: "Immediately after the appear-
ance of the word an auditory-kinaesthetic
image thereof. Then came the know-
of the Southern Society for Philosophy and Psychology, Professor Ogden stated that
he had reported some years ago at one of the meetings a series of experiments similar
to the present in all details. He never published his results, but they were identical
with my own. In the interests of a better insight into individual differences it is to be
hoped that Professor Ogden will some day give us the advantage of his unpublished
results.
MEANING AND IMAGERY
207
ledge that the fork is an instrument for
eating, accompanied by a weak visual
image of a fork. I was also conscious that
'fork' stands in relation to 'knife.'" —
Kulpe, 2/III.
(3) The concept of purpose though often
restricted to one of various possible ends has
always a certain generality.
Kette: "I pictured to myself a toler-
ably strong chain and remembered from
the days of my youth that such chains
were used to tie animals in their stalls. I
saw the whole situation of that day rise
up before me." — Lehner, 5/III.
(4) The consciousness of purpose seldom
stops with a means but rests in a concept
conceived of as the object's end.
Uhr: "Immediately after the appear-
ance of the word I had an auditory-kinaes-
thetic image, then the thought: 'The
clock must be wound up!' and then the
further thought: 'The clock tells the
time!' Then I reacted. Weak visual
image of a clock on a wall." — Kiilpe,
2/III.
(5) The concept of purpose, even though
delayed, comes as a natural development of
thought about the object.
(3) The kinasthetic image is often per-
fectly definite and limited to an individual
act in a certain time and place.
Pickel: "I imaged a pick-ax, such as is
used for working hard ground and saw
myself in my garden in the act of lifting it
in the air. The consciousness of the
purpose of a pick-ax is a psychological
process which cannot be identified with
the act of lifting it." — Lehner, 2/V.
Wiege: "The meaning aroused the
image of a cradle. I go back in thought
to my childhood and feel how I rock my
brother. The kinaesthetic image in this
case contains a great part of the purpose."
— Griininger, 4/III.
(4) The kinasthetic image regularly
concerns an art which is a means to the
object's end.
Lampe : " I imaged the lamp that I use
in my dwelling, and saw clearly that it
did not burn brightly enough, and then
imaged the turning up of the wick. The
kinaesthetic image of the movement
cannot be identified with the conscious-
ness of purpose." — Lehner, 2/III.
Trichter: "Immediately after the
simple meaning of the word, I had the
visual image of a funnel and then the
kinaesthetic image of laying hold of it
with my right hand and placing it over
an opening. Here also the kinaesthetic
image falls short of being the fulfilment of
the purpose. For I think that the funnel
is the instrument by means of which I
pour fluid through an opening, and my
image is only the placing of the funnel
in the opening." — Griininger, 4/III.
(5) The kincssthetic image is ojten
forced and is superfluous to the under-
standing of the function of the object.
208
T. V. MOORE I
Handbeil: "I soon understood the
word, but the simple consciousness of
meaning was forced into the background
of consciousness by the task. I can ex-
press this simple consciousness of mean-
ing by the sentence: * I know well what
that is ! ' Then I asked myself under the
influence of my task: 'What purpose does
it serve?' Then there came to me the
clear concept that it is of use in cutting
wood. With this concept of purpose
were some broken, confused words. I do
not know whether they were German or
English. There was also a dark blurred
image of an island in Lake George, where
I have often cut wood in summer." —
Moore, i6/VI.
(6) The concept of purpose, though at
times more or less restricted, never mis-
carries entirely.
Fahne: It was rather difficult for me
to connect a kinsesthetic image with the
word. At first I imaged a flag as I saw
one recently waving on a little tower in
Leopold Street. But I said to myself at
once, 'This waving is not a kinaesthetic
image.' Then I imaged to myself how I
would place this flag on the little tower.
That the purpose of the flag is not cov-
ered by my motor image of it, goes
without saying." — Lehner, S/III.
(6) The kinczsthetic image is not always
pertinent to the purpose of the object.
Bohrer: "Again the meaning first and
then a visual image of the object — of a
whole situation. I attempted to screw
a drill through the wall, and instead of
that I lifted the whole wall with the
drill."— Frl. Marezoll, 9/VII.
V. MEANING AND THE WORD
(a) Quantitative Results
In the perception of the meaning of words, subjects often
spoke of the meaning being associated with an auditory-
kinsesthetic verbal image of the word itself. No attempt was
made to find out by reaction time the temporal relations of
the verbal image and meaning in the perception of printed
words. From the introspective results no definite answer
can be obtained. The two are so close together that they
appear simultaneous. One might, however, surmise that
since the word must be read, in order that it may be perceived
and understood, verbal sensations or verbal imagery are
likely to come prior to understanding.
On account of the close connection with the sensations
involved in reading and the understanding of printed words,
such material presents no little difficulty in studying the
MEANING AND IMAGERY
209
necessary relations between verbal imagery and meaning.
Pictures seemed to offer a more favorable material for study.
If meaning is the kinsesthesis of speech, then the knowledge
that a picture before me represents a tree should come when
I name the picture and not before. A series of reaction
times for the naming of pictures and perceiving the meaning
of pictures should give approximately identical results.
Three of our subjects took part in these experiments.
With all three subjects there is strong evidence that in
general it takes longer to react to the word than to the
meaning. The means for reaction to the word are, in every
case, longer than those for meaning. This excess is also
SUBJECT LEHNER
Pictures
Date
Word
T
V
Date
Simple Meaning
T
V
' Baum
845
214
Sage
492
i
Uhr
630
i
Katze
419
74
Lilie
968
337
Hahn
251
242
Sichel
638
7
9/III
Ring
923
430
9/III
<
Kafer
Hammer
626
794
.4
Eimer
Krone
500
629
I3l
Treppe
601
30
Pfau
416
77
Sense
726
95
Hobel
517
Kamel
698
67
I9/V
Spinne
579
86
, Mitra
775
144
Schlitten
572
79
f Wiirfel
840
209
Fernglas
704
211
I9/V
\ Frosch
[ Pinsel
649
689
18
58
26/V
Haue
Krug
718
536
225
43
26/V
.
Apfel
Uhr
708
493
77
138
Stuhl
Baum
380
673
H3
1 80
Wiege
780
149
Uhr
547
54
Katze
485
146
Lilie
334
159
Ring
720
89
Sichel
383
no
Haken
563
68
Hammer
558
65
Sage
594
37
Sense
349
144
Fernglas
410
221
Kamel
392
IOI
20/VH
.
Pfau
Krone
376
713
255
82
20/VII
Kafer
Pinsel
460
468
33
25
Eimer
375
256
Treppe
327
166
Hobel
598
33
Frosch
411
82
Stuhl
373
258
Wiirfel
400
93
Haue
623
8
Wiege
39i
102
Schere
560
70
Apfel
553
60
Krug
474
157
Korb
344
149
Mitra
572
79
29)18,324
3,392
30)14,798
3,350
631
117
Mean = 493
112
Median = 480
Number of times median for meaning exceeds reaction time = 6.
210
T. V. MOORE
SUBJECT KULPE
Pictures
Date
Word
T
v
Date
Simple Meaning
T
V
Horn
1,020
98
Pferd
525
108
Sofa
767
155
Biirste
510
123
Rechen
824
98
Haus
798
165
Dampsfchiff
769
153
Ente
308
325
Kette
825
97
Fahne
673
40
Zwicker
738
184
Flasche
967
334
9/IH
Finger
Stiefel
843
1,692
79
770
9/III
Trichter
Bohrer
570
659
6l
26
Dolch
837
85
Uhr
524
109
Kerze
1,718
796
Lilie
957
324
Lilie
957
35
Hahn
573
60
Sichel
834
88
Kafer
767
134
Haken
759
163
Hammer
722
89
Ring
332
590
Lowe
327
306
14)12,915
3,39i
14)8,880
2,206
922
242
Mean =633
157
Median = 616
Number of times median for meaning exceeds reaction time to word= I.
SUBJECT MAREZOLL
Pictures
Date
Word
T
v
Date
Simple Meaning
T
V
Korb
959
48
Zirkel
998
34i
Schere
825
86
Eimer
643
H
Pickel
1,862
951
Stuhl
463
194
26/V
Apfel
622
289
26/V
Sichel
1,009
352
Uhr
78i
130
Trommel
561
96
Wiege
729
182
Dampfschiff
463
194
Tasse
1,175
264
Windmuhle
526
131
Maske
873
38
Kirsche
560
97
Lyra
782
129
Zither
796
139
25/VI
Engel
•736
175
25/VI
'
Kette
739
82
Ochs
671
240
Fahne
50i
156
Mitra
913
2
Wurfel
628
29
12)10,928
2,534
12)7,887
1,825
911
211
Mean = 657
152
Median = 594
Number of times median for meaning exceeds reaction time for word = o.
greater than the mean variation. With one subject in 29
reactions to words, only 6 were shorter than the median
for meaning; with another, i in 14; with another, o in 12.
(b) Introspective Data
Turning to the introspective results we find them in
accordance with the quantitative measurements. Time and
MEANING AND IMAGERY
211
time again, whether the task were meaning or word, the
same sequence of events was perceived, viz., (i) meaning,
(2) word, (3) reaction. Often, however, when the task was
meaning, the word was reported as coming during or after
reaction.
Some special points of difference between the word and
the meaning are given below.
(1) The meaning leads to the word — the
designation of the picture.
Frosch : " The meaning was first present.
I felt a strong striving for the word, as it
were from various sides of the drawing.
The reaction followed after the entrance of
the word." — Lehner, 9/V.
(2) A meaning cannot be lacking if the
subject names the picture — no matter what
the task.
(3) The meaning is what it is by its own
right. It is never said to have a meaning.
Pferd: "Immediately after I saw the
picture I experienced a tone of familiarity
and knew what this picture represented.
At the same time, with the reaction came
the word 'Pferd.' I did not react to
the word. The tone of familiarity was
related not to the picture, but to what it
signified. The picture was a symbol of
real objects and its signification con-
sisted herein, viz. — to point to them." —
Kiilpe, 9/III.
(4) The meaning is sometimes desig-
nated by a word which is known to be in-
appropriate.
Lilie: "First I recognized in the picture
a flower, then I named it by mistake
'Tulpe.' I knew that 'Tulpe' did not fit
the picture. Then through the form of
the flower, etc., I was occasioned to say
"Glockenblume."— Lehner, 9/III.
(i) The word never leads to the meaning.
(2) The word may be lacking when the
task is meaning.
Eimer: The word did not appear at all.
Various memories were in the back-
ground of consciousness. — Frl. Marezoll,
26/V.
(3) The word may have a special mean-
ing of its own; e. g., the word has a more
general meaning than that of the picture.
Engel: "Immediately a memory image.
After this image came the word. I knew
that the meaning of the word was more
general than that of the picture." — Frl.
Marezoll, 25/VI.
(4) The word is never designated by a
meaning.
212 T. V. MOORE
VI. INFLUENCE OF THE OBSERVER'S ATTITUDE
When a short abstract of this paper was read last Decem-
ber at the meeting of the Southern Society of Philosophy and
Psychology, it was suggested that the difference in reaction
time to meaning and imagery is to be explained by a difference
in the attitude of the subject. He reacts quicker when told
to react to meaning, not because the meaning is something
different from the imagery but because he himself assumes
a different attitude.
This objection implies that there is no real difference
between meaning and imagery, but that when we call them
by different names the subject, for some obscure reason,
assumes such a different attitude that it markedly influences
his reaction time. The objector in other words does not
wish to admit a difference between meaning and imagery, and
refers the difference in reaction time to an unexplained and
perhaps inexplicable mystery.
To say the least, this explanation is highly improbable.
For supposing there is no such thing as a special 'meaning
process' and that the accruing image is identical with the
meaning, then the task of the subject in the two sets of
reactions is really identical. It is simply called by different
names. If that were the case, then the subject ought (i) to
have a real difficulty in distinguishing his two tasks. (2) He
ought to give introspective reports identifying the two pro-
cedures. (3) The reaction times ought to be identical within
the limits of the probable error.
None of these things were so, but on the contrary (i) The
tasks were readily distinguished. (2) The introspective re-
ports clearly separate the two processes. (3) The reaction
times are markedly different.
All this tends to render highly improbable, if not impos-
sible, the explanation which suggested that the difference in
the reaction times is not to be explained by a real difference
in the tasks, dependent on a difference between meaning and
imagery, but is due entirely to the difference in the attitude
of the subject. In fact, it is very hard even to imagine a
MEANING AND IMAGERY 213
mental mechanism which would produce two separate atti-
tudes with such different effects in the reaction-times, if
that to which the subject takes an attitude is in both cases
merely one and the same thing that the experimenter calls
by a different name.
Let us, however, go a step further. Our subjects reacted
to visual and kinsesthetic images. If we wish to compare the
reaction times in this case we will find them markedly dif-
ferent. Is it possible to explain that difference by a difference
in the attitude of the subject?
If we should argue visual imagery is distinct from kin-
aesthetic (i) because the subject distinguishes two different
tasks when told to react to the one or the other; (2) because
the introspective reports clearly separate the one from the
other; (3) because the reaction times to visual imagery are
much shorter than to kinsesthetic imagery, no one would
doubt the validity of the argument. When, however, the
same argument is made in regard to imagery and meaning, it
is called in question and the attempt is made to explain away
the difference by ascribing it to a difference in the attitude of
the subject. If, however, the difference in the attitude of the
subject is not the real explanation in the latter case, but a
real difference between visual and kinsesthetic imagery, then
this difference in the attitude of the subject cannot, without
any more ado, explain the shorter reaction time for meaning
as compared with imagery.
Furthermore, the difference in attitude itself must be
explained. Granted that there is a difference in attitude,
what is the most likely explanation for the fact? The first
thing that comes to mind is that in the two sets of conditions
the subject is taking an attitude to two different things. If
that is the case then, meaning and imagery must be dis-
tinguished. But how distinguished — as two different mental
processes or as two aspects of one and the same process?
In the sequence of events that follow the exposition of the
stimulus word, there may be, if you wait long enough, not
only visual but also kinsesthetic imagery. Are these aspects
of one and the same mental processes, or specifically dif-
214 T. V. MOORE
ferent items in a definite series of events? Reasons have
already been given for distinguishing them. These reasons
point to events that are qualitatively distinct, and the
distinction can scarcely be called in question. But the very
same reasons point to meaning as qualitatively distinct from
imagery. When, furthermore, one considers the fact that in
the understanding of words the meaning process is never
absent, but that visual and kinaesthetic imagery may both
be lacking, there is an added reason why meaning should not
be identified with an aspect of visual or kinsesthetic imagery.
Furthermore, a difference in the attitude of the observer
cannot be made the sole reason for the difference in the re-
action times.
(i) In the set of experiments referred to in the beginning,
the subject's task was to observe and remember a series of
words, pictures or objects. Nothing was said about attending
to meaning or imagery. He had simply to report what he
had experienced — whatever that might be. Here the ques-
tion of a difference in the "set" of the observer does not enter
at all. In these experiments, the subjects reported that in
the perception of words, meaning preceded imagery. This
suggested the problem of an objective test of the accuracy of
the introspection. The reaction time experiments followed,
and confirmed with entire satisfaction the introspections of
the earlier series.
(ii) In the reaction time experiments no matter what the
task — whether the subject is in the meaning attitude or the
image attitude, he regularly reports meaning as coming
prior to imagery. If the difference in the £ set' of the observer
were the sole reason for the difference in reaction time, we
should not expect that no matter what his 'set' he would
nevertheless observe a rather constant temporal relation
between meaning and imagery.
The introspective results and the reaction times are supple-
mentary. When taken together they leave no doubt that
we have really been investigating the temporal relation of
meaning and imagery.
MEANING AND I M ACER Y 215
VII. THE CONTEXT THEORY OF MEANING AND THE TEM-
PORAL RELATIONS OF MEANING AND IMAGERY
It may now be asked: Whom does all this concern?
Who maintains that imagery is meaning? In spite of a
certain modification of the image theory of meaning, Pro-
fessor Titchener's context theory cannot account for the
experimental facts brought out in his own and other labora-
tories. From an analysis of his theory it is apparent that he
maintains that meaning is often identical with imagery. In
fact under the conditions of our experiments the images and
words that followed upon the sensations of the stimulus words
and pictures were actually the context. Analogous conscious
states have been reported by Cornell observers as the meaning
under somewhat similar conditions. But they did not take
into consideration the temporal relations of meaning and
imagery.
A brief analysis of the context theory of meaning will show
how intimately it is concerned with the temporal relations
of meaning and imagery.
(a) Outline of the Theory
" Meaning, psychologically, is always context."1 Such is
the definition that Professor Titchener gives to a fact of
consciousness with which the modern psychology of thought
is now interested.
What is context? Context in English is a word used to
signify the setting of a sentence or a quotation — its relation
to what the author has written before and after the passage in
question. Titchener lays particular stress upon what comes
after in the definition of psychological context. "Context,
in this sense, is simply the mental process which accrues to
the given process through the situation in which the organism
finds itself." A sensation by itself has no meaning — neither
has an image. When a second mental process accrues to a
former one — this second mental process is the meaning of the
first one. It does not produce a new something called mean-
1 'A Text Book of Psychology,' New York, 1911, p. 367.
216 T.V. MOORE
ing, it is the meaning. "One mental process is the meaning
of another mental process if it is in that other's context."2
What are the mental processes that accrue to others
and thus constitute their meaning? Originally the secondary
process which constituted the meaning was a group of sensa-
tions coming from a bodily attitude of the organism. If
the animal took an attitude of defence the kinaesthetic sensa-
tions thus aroused did not exactly mean — did not signify that
something to be feared was at hand. The whole complex of
sensations involved constituted the meaning "something to
be feared."
At the present day, however, the human mind has passed
beyond the elementary stage of the primitive organism. The
essential difference between present human intelligence and
its early prototype consists in the use of imagery as well as
sensations for the constituents of meaning. "Image has
now intervened upon sensation and meaning can be carried
in imaginal terms."2 Thus spoken and written language has
become possible. A sensation arouses an image and the
image — the psychological process accruing to the sensations —
is the meaning of the sensation.
Various types of mind exist. Each has a special tendency
to form some kind of imagery in understanding sensations.
Indeed "If we were to make serious work- of a differential
psychology of meaning, we should probably find that in the
multitudinous variety of situations and contexts, any mental
process may possibly be the meaning of any other."3
It is Professor Titchener's opinion however that of all
the possible types of supplementary mental processes, two
are of special importance: kinsesthesis and verbal images.
Indeed he pushes the verbal theory so far as to say: "The
words that we read are both perception and context of per-
ception, the auditory kinsesthetic idea is the meaning of the
visual symbols."4
1 Op. cit., p. 367.
2 Op. cit., p. 367.
3 'Lectures on the Experimental Psychology of the Thought Processes/ 1909, p.
178.
4 'A Text-Book of Psychology,' p. 368.
MEANING AND IMAGERY 217
Thus far, Professor Titchener's theory is entirely psycho-
logical. But in order to meet all possible contingencies
arising from introspections that he or others may report,
where meaning shows no trace of a sensory conscious element
—a physiological factor is introduced.
Meaning is not always conscious; i. e., the imaginal
supplement to the sensation is not always to be found even
by the most careful introspection. In such cases the sensory
supplement exists — it is a physiological process in the nervous
system.
Professor Titchener thus summarizes his theory of per-
ception:
"Our account of the psychology of perception is now, in
the author's view, complete. It has embraced four principal
points :
"First, under the general laws of attention and the special
laws of sensory connection, sensations are welded together,
consolidated, incorporated into a group.
"Secondly, this group of sensations is supplemented by
images.
"Thirdly, the supplemental group has a fringe, a back-
ground, a context; and this context is the psychological
equivalent of its logical meaning.
"Fourthly, meaning may lapse from consciousness and
conscious context may be replaced by a non-conscious nerv-
ous set."1
The type of meaning in the third caption is decidedly
different from that given a few pages previous. There
meaning is context — context is the mental process that
follows upon and accrues to another mental process. The
examples given are the images spoken of in the second caption.
Here we suddenly find that meaning does not lie in the adven-
ing images — but in their fringes.
To harmonize this new idea with what has gone before
we may suppose that if meaning is conscious (in the sense of
being conscious described by Titchener) it is given by the
context which may be (a) a second group of sensations, (b)
1 'A Text Book of Psychology,' 1911, p. 371.
2i8 T. V. MOORE
an image or a group of images, (c) the fringe or background of
such images — the fringe itself being always understood as
some kind of sensory element or elements, (d) various com-
binations of (a), (b) and (c).
(b) The Evidence for the Theory
In the interests of simplicity we may leave aside the
speculations about meaning in the primitive organism and
confine ourselves to the explanation of the fact of meaning
as we experience it.
On what then, may we ask, is the statement based that
meaning is context — that it is a ' sensory complex $, following
upon sensation or image A? The points of evidence are:
1. Introspection shows that when a word or a sentence is
understood and careful search is made we always find some
kind of imagery — verbal, kinsesthetic, visual, etc.
Granted that this is so what does it prove? Nothing more
than this. In the complex of mental processes called up by
the task of understanding a word or sentence imagery is
present. It does not show that this imagery is the meaning
— which is the very point in question.
Titchener says: "The meaning of the printed page may
now consist in the auditory-kinsesthetic accompaniment of
internal speech; the word is the word's own meaning."1
He then refers in a note to introspections in the studies
of Watt and of Messer which speaks of meaning being simul-
taneous with auditory-kinaesthetic imagery. But such a
citation is not to the point. The fact that one thing accom-
panies another is certainly no evidence that the two are
identical.
2. Analysis shows no evidence of 'imageless thoughts.'
What analysis shows is the fact of meaning. Many
observers have maintained that in their consciousness of
meaning sensational elements are lacking. Professor Titch-
ener in his analysis finds also the fact of meaning and giving
to the students in his laboratory the task of reporting every
mental process that they can observe, he obtains experiences
1 'Lectures on the Experimental Psychology of the Thought Process/ p. 177.
MEANING AND IMAGERY 219
far richer in sensational elements than are elsewhere found.
Given the task, 'find imagery,' and it will certainly come.
And if the subject be told to look for imageless imagery, it
will not be found. Meaning and imagery however, have
been found both by Professor Titchener and a number of
other observers. Facts are common property. It remains
for Professor Titchener to prove that meaning is identical
with the concomitant or subsequent imagery. This he has
not done.
The context theory of imagery demands imagery, when
meaning is present. If meaning equals imagery, imagery
equals imagery. No imagery — no meaning, must be the
conclusion to be drawn from this theory. Nevertheless
Professor Titchener shrinks from admitting all that is involved
in his doctrine. Why? Because he himself has observed
that there are times when he experiences meaning and is not
conscious of imagery. He himself, therefore, in spite of the
ease with which he images things and situations, has experi-
enced the very state of mind the existence of which he denies.
"In rapid reading, the skimming of pages in quick succes-
sion; in the rendering of a musical composition, without
hesitation or reflection, in a particular key; in shifting from
one language to another as you turn to your right or left-
hand neighbor at a dinner table: in these and similar cases,
meaning has time and time again, no discoverable representa-
tion in consciousness."1 No discoverable representation in
consciousness means no sensational element — no sensational
or imaginal complex.
What is Professor Titchener's explanation of such "image-
less thoughts" that come to him as he skims over the pages
of a book? He has found "imageless thoughts," what then
is to be done with them? Explain them away and then deny
their existence. How explain them away? Refer them to
the nervous system? Meaning here is not imagery for no
imagery is present. What is it then? A physiological proc-
ess, without any conscious accompaniment. Why without
any conscious accompaniment? Because by hypothesis the
1 'A Text-Book of Psychology,' p. 369.
220 TV V. MOORE
only conscious processes that come into consideration are
sensations and these are lacking.
On the one hand, we have an hypothesis; on the other,
a fact — the imageless consciousness of meaning (imageless
thoughts) in rapid reading. The fact cannot be accounted
for by the hypothesis; therefore Professor Titchener denies
the fact. My consciousness of meaning is unconscious. I do
not think but my nervous system is thinking for me.
The reference of imageless thought to an unconscious
physiological process in the nervous system brings us to a
third point in the evidence for Professor Titchener's theory.
(3) "Our psychology is to be explanatory and our explana-
tions are to be physiological."1
Adherence to this principle and the ruling out of facts
that it cannot explain, give to Professor Titchener's theory a
certain plausibility.
What can be referred to the nervous system is explained
What cannot be referred to the nervous system is not ex-
plained. It is in fact inexplicable. There must be a mistake
in the observation. It must be explained away. The ner-
vous system with its sense organs and its centres, can appar-
ently take care of sensations and images. It gives us the
sensational elements of our conscious life and apparently
excludes anything like imageless thinking. If then we are
to explain ' imageless thought' we must analyze it in terms
of the elements given by the nervous system, or else explain
it away altogether.
Such a procedure, however, places empirical psychology
not only under the dominion of metaphysics, but subjects it
to one particular metaphysical theory. Under such condi-
tions an impartial empirical study of the mind becomes
impossible. Let us first study the facts of consciousness and
then build up our metaphysical theories.
Professor Titchener is right in demanding that the science
of psychology should be explanatory; he is wrong in main-
taining that everything must be explained in consonance
with a particular metaphysical theory.
1 'A Text-Book of Psychology,' p. 370.
MEANING AND IMAGERY 221
As a matter of fact neither Professor Titchener nor anyone
else knows the limitations nor the possibilities of the nervous
system. Nor does anyone know, for that matter, what the
nervous system may be called upon to do if it is to explain
the facts of our conscious life. We do not know all about the
facts of consciousness and until we do, explanatory psychology
must be careful. We do not know all about the nervous
system and it is not wise to distort the fact of consciousness
to fit the narrow outlines of our present horizon.
Let us first investigate the facts of consciousness without
any timidity about their ultimate explanation. Let us first
find out what we have to explain, and then explain it.
The context theory of meaning is not based entirely upon
such general considerations as we have picked out from
Professor Titchener's writings. There are a number of
experimental studies that have been put forward as tests in
confirmation of the theory.
Of these, we may analyze two, leaving a more complete
account of the literature to a full report of our experiments
which we hope to publish later.
Helen Clarke,1 in an article on ' Conscious Attitudes ' took
up the problem of the understanding of words and sentences.
She confirmed the reports of other observers that ' often the
images are adequate, irrelevant or even contradictory' (p. 241).
The inadequacy she explained by saying that 'we have no
criterion save the facts themselves, by which we can decide
how clear or complete an image must be in order to carry a
meaning' (p. 241). The contradictory character she ac-
counted for by pointing out that in every one of her cases
there was ' sufficient connection between the logical meaning
of the word, and the psychological context of the act of
understanding, for the latter to carry a general meaning'
(p. 242). The fact of irrelevancy, she said, was less easy
to explain.
Miss Clarke therefore seems to be conscious of the fact
that words have a logical meaning which cannot be identified
with the imagery that they evolve. She distinguishes be-
1 Am. J. of Psycho!., XXII., pp. 214-249.
222 T. V. MOORE
tween the word — the imagery that it evolves — and the
meaning that is carried. She finds also that imagery is often
irrelevant. Irrelevant to what, we may ask? To the mean-
ing. She therefore realizes a difference between the psycho-
logical process called an image and another something of
which she is also conscious and which may be termed the
meaning of the word. Miss Clarke1 seems to look upon
general meaning as a logical something of which no account
need be taken in psychology. If, however, the task of psychol-
ogy is to investigate all conscious processes, logical meaning
cannot be ruled out as 'outside the sphere of psychology.'2
For "logical meaning" is conscious. Its nature is therefore
a psychological problem. It is that something to which the
imagery is often inadequate, irrelevant and contradictory.
Miss Clarke has implicitly at least recognized it as a conscious
state, distinct from imagery.
Edmund Jacobson3 investigated by the Method of Intro-
spection (i) The Perception of Letters, (2) The Meaning of
Words, (3) The Understanding of Sentences. The instructions
to his subjects (three observers) were as follows:
I. Give a minute account of all the mental processes you
experience in their temporal order of sequence.
II. Put direct description of conscious processes outside
of parentheses, and statements concerning meanings, objects,
stimuli and physiological occurrences inside.
The experiments on the perception of letters showed that
under the instructions given their meaning is usually accom-
panied by the arousal of what Jacobson termed designatory
processes, viz., kinaesthetic or auditory sensations or both.
Jacobson calls attention to the fact that "The main point
to note is that the precise statement of meaning is by no
means easy." Nor does he state anything more definite as
to what the meaning of a letter is.
The experiments with the meaning of the words were made
as follows: "A written word, was laid before the observer for
a period of one minute. He was instructed to fixate the
1 Along with Geissler, Am. J. of Psychol., XXIIL, p. 194.
2 Cf. Geissler, /. c.
3 Am. J. of Psychol., 1911, XXII. , pp. 553-577.
MEANING AND IMAGERY 223
word, to utter it with quick repetition and to get at its mean-
ing. The concluding ten seconds were marked off by signals;
and the observer's task was to report what occurred in con-
sciousness during the particular interval. 9\ The observer re-
ported two kinds of imagery: (a) That which appeared as the
carrier of the meaning and (b) that which appeared as irrele-
vant. No logical or psychological test could be found to
distinguish between the relevant and irrelevant imagery.
The conclusion of Jacobson was "that the conscious
meanings brought out in these experiments are not perfect
and static logical meanings of definition. . . . Logically, the
representation of meaning is inadequate; psychologically, it
is adequate to the demands of the occasion" (pp. 568-569).
In his experiments on the meaning of sentences, Jacobson
found cases in which (i) an automatic reading was followed
by a perception of the meaning identified with images called
forth by the experiment. (2) Cases in which the mean-
ing did not come to the subject at all in spite of a wealth
of visual, organic, kinsesthetic and tactual sensationsl (3)
Cases in which the visual and auditory images and sensations
from reading were the sole processes present in consciousness
— and yet the sentence had meaning. Jacobson concludes:
(i) "Wherever there is meaning there are also processes,"
i. £., sensations and images of one kind or another. (2) "The
correlated meanings and processes are two renderings from
different points of view of the same experience."
The first conclusion seems established by the introspective
reports, but it holds only for the conditions of these experi-
ments where ample time is given for images to appear and
the task is set to report primarily mental processes, i. e.,
sensations and images; and secondarily, in parentheses, to
note meaning as it arises.
The second conclusion: (which is really the "crux" of the
whole situation) meaning is an aspect of sensation and imagery,
is simply stated and the reader is left to judge for himself
on what evidence the conclusion is based. The only evidence
in his paper for such an indentification is to be sought in the
fact that his subjects, as a rule, were not satisfied that they
224 T. V. MOORE
had anything that corresponded with their idea of meaning
till relevant imagery was present. This simply shows that
meaning in the Cornell sense is not present till such imagery
arises. From Jacobson's own data, it appears, however,
that meaning in a broader sense must have been present
when meaning in the Cornell sense was denied. When Dr.
Geissler, instructor in psychology at Cornell, for 3 seconds,
looked at the sentence, "Did you see him kill the man?,"
and then declared at the end "No meaning all the way
through," we can only conclude that "meaning" must have
been taken in a very restricted sense. When again he looked
at the sentence, "The iron cube fell heavily on the floor,"
reads it as so many meaningless words, and then on rereading
obtains the meaning, a very loud sound, the time of the
whole procedure being 4.5 seconds, the conclusion is strength-
ened that during the experiment, he was seeking for a meaning
in the sense of an imaginal representation. In this sense, and
in no other, is Jacobson's conclusion warranted. An imaginal
representation is some kind of imagery. The sweeping con-
clusion that meaning is an aspect of imagery requires the
proof of another proposition, namely that all meaning
consists in imaginal representation.
The data of this piece of introspective work is incom-
patible neither with the data nor the conclusions of the
Kiilpean school. Indeed it has confirmed the fact that the
meaning of a sentence may be present when the sole proc-
esses present in consciousness are the visual and auditory
images and sensations from reading. And if it be true that
on certain occasions, as in Geissler's case, these same processes
were present and the meaning was really absent, one should
conclude that they cannot be identical with the meaning.
In like manner, a physician refuses to admit that a definite
microorganism is the cause of a disease — if at times it is
found when the disease does not occur, and the disease
occurs when the organism is absent. Jacobson should there-
fore have admitted that there are times at least when meaning
is not a mere aspect of sensations and images.
Professor Titchener looks upon the chief value of Jacob-
MEANING AND IMAGERY 225
son's work in making the distinctions between the mere
statement that meaning is present and the analytic descrip-
tion of the psychological part of meaning. He says that
"He finds no specific ' meaning process' underlying the
statement of meaning."1
True it is that Jacobson found no special sensory or
imaginal process as the habitual carrier of meanings, but he
did not prove that meaning is not itself a conscious process.
In fact, his experiments seem rather to confirm the conclusion
that meaning is not imagery, but something else altogether.
Had the Cornell School taken cognizance of the temporal
relations of meaning and imagery, the context theory of
meaning would have been profoundly modified. Imaginal
terms may accrue to incoming sensations and constitute by
definition their context. Do they constitute their meaning?
A determination of the temporal relation that imagery bears
to meaning shows that this is impossible. What comes after
another cannot be said to cause, or constitute it, or be identical
with it. Meaning, therefore, is not context. What is it —
a mere negation? Not at all. It is a definite mental process
sui generis. What are its qualitative characters? Some of
these have been already indicated. A further development
of the concept will be given with the fuller account of these
investigations.
1 " Description vs. Statement of Meaning," Am. J. o/PsychoL, 1912, XXIIL, p. 182.
THE SHORTEST PERCEPTIBLE TIME-INTERVAL
BETWEEN TWO FLASHES OF LIGHT1
BY KNIGHT DUNLAP
The Johns Hopkins University
The determination of the minimal perceptible time-inter-
val: the shortest interval between two stimuli which allows
the stimuli to be perceived as .successive, and not simultan-
eous is important for many lines of work, including problems
of time-perception and rhythm and also problems of rate-
perception. Moreover an important theory of psychic syn-
thesis has been supported by interpretations of certain
measurements of the time-threshold for disparate stimula-
tions (i. e., stimulations of two modes of sense in succession).
My interest in these several lines of research, and also in
certain purely visual phenomena, led me to commence, in the
summer of 1912, an investigation of the time-threshold for
visual stimulation, and its relation to the ' critical frequency'
of flicker and fusion. The result of that summer's work
(done at the Johns Hopkins University, with myself as
principal observer), encouraged me to attempt further work
on the problem with better apparatus. During the next
college year (1913-1914) a graduate student was allowed to
take up the problem, and obtained some results which seemed
important.2 This student, on leaving the University, took
his unelaborated results with him, and I have not since been
able to obtain them. Last summer, having the opportunity
to work in Dr. Hyde's laboratory, I took up the problem
again with specially constructed apparatus and obtained
results which are interesting and important. I shall give in
the following paper the results of both of my experiments.
1 From the Nela Research Laboratory, National Lamp Works of the General
Electric Company.
2 This work was done with a pendulum apparatus, giving great accuracy, and having
other advantages over the rotation apparatus first used; but with some difficulties of
manipulation.
226
PERCEPTION OF TIME-INTERVALS 227
The first work on the visual time-threshold was done by
Exner,1 who worked with electric sparks, and found thresholds
of 44 cr at 280 mm. distance, and 21 a at 640 mm. Weyer,2
in Wundt's laboratory, found a much lower threshold, 12 a.
Weyer also found, using electric sparks, a flicker-threshold
from 25 a to 87 <r, according to the adaptation and other
conditions, and a threshold for separation of a series from
42 o- to 105 cr.
Shortly before my work 'was begun Bassler3 published the
results of some of his investigations, in which he found the
time-threshold (length of shortest perceptible dark interval)
to be about 40 a with two visual stimulations, and the flicker
point to be about one third as much (for serial stimulation).
None of these results are very significant, the work with
electric sparks suffering from lack of control, and Bassler's
being affected by serious defects of method.
Bassler used black discs on which were either two white
sectors or a regularly spaced series, and rotated the disc close
behind a screen in which was a hole a centimeter and a half
in diameter. This hole, in which the alternation of black
and white occurred, was observed from an unspecified dis-
tance. A student in our laboratory reproduced Bassler's
apparatus, as nearly as Bassler's description allowed, and
we found that eye movement was a very important factor in
the observation, the eye movement being induced or increased
by the motion of the black and white edges as they traveled
across the aperture.
PRELIMINARY WORK*
It is obvious that the proper attack on the problem of the
visual time-threshold involves control of the intensity and the
duration of the flashes of light, and of the adaptation and
movement of the eye, as well as of the areas stimulated.
In my first experiment I succeeded in eliminating the most
1 Exner, Pfliiger's Archiv, XI., S. 407.
2 Weyer, Philos. Studien, XV., S. 67-138.
3 Bassler, Pflugtfs Archiv, 1911, Bd. 43, 245-251.
4 This work was reported before the Natural Academy of Sciences, November 19,
1913. See abstract in Science, 1913, Vol. 38, p. 699.
228 KNIGHT DUNLAP
serious important cause of eye movement, namely the travel-
ing of the illumination across the area of stimulation, and
kept the illumination constant and the eye dark-adapted.
The first factor which I wished to investigate was the effect
of the duration of the flashes, and the second was the effect
of the intensity.
For simplicity's sake I adopted the method of rotating
sectors, measuring the time interval by determining the
speed of rotation, and computing from the angular width of
the sectors. This method has two serious disadvantages:
first, the change in speed, which is necessary to vary the
length of the interval between flashes, varies the length of
the flashes also, so that the effect of absolute flash length
cannot be easily determined, relative length only being con-
trollable. Second, the pair of flashes is necessarily repeated
rapidly again and again unless some special device is used
to cut off the exposure on all but one round of the disc; and
this repetition is a factor which adds greatly to the difficulty
of the determination.
My apparatus consisted of a Nernst glower, enclosed in a
metal box, with a lens; a disc of white plaster of paris; a
motor of controllable rate, driving, by a reducing belt, a
spindle on which discs of adjustable sectors could be rotated;
and an Ewald chronoscope for counting the rotations of the
spindle .during a given time.
The lens, 83 cm. from the Nernst glower, focused the light
into an image, of approximately the same size as the glower,
in the plane of the surface of the rotating sectors on the
spindle, with the long axis of the image in a radius of rotation.
When not interrupted by the sectors, the light fell on the
plaster surface placed 35 cm. beyond the focus, forming a
nearly rectangular spot 3.5 by 5 cm. The brightness of this
spot was not measured, but was kept constant by maintaining
a constant current through the Nernst glower, and by
frequently inserting a Lummer-Brodhun photometer in the
position of the disc, comparing the illumination of the Nernst
with that of a standardized 8 c.p. carbon lamp. The two
brightnesses used were equal to those produced on the same
PERCEPTION OF TIME-INTERVALS 229
surface by the 8 c.p. lamp at 36.5 cm. and 67.5 cm.1 respec-
tively. Since the rotating sectors moved across the beam
of light at the focus, 'traveling' of the illuminated areas was
nearly eliminated; since the focused image was narrow, the
time between the beginning of the illumination (or the dark
period) and the full illumination (or complete cut-off) was
so small as to be negligible.
The observer (myself in most cases) sat between 75 and
80 cm. from the plaster surface, the angle between his line
of sight and the axis of the Nernst beam being 45 degrees.
The plane of the plaster surface was so placed that it made
equal angles with the axis of the beam and the line of sight.
With this apparatus I made determinations both of the
time-threshold for two flashes, and of the critical frequency
for a series of interruptions. In the flicker work, a different
disc, with the appropriate sectors cut out, was used for each
of the different ratios of light to dark interval. In the work
on time-threshold, a combination of sectors was used, giving
two openings, from o° to 90° in width, separated by a 5°
sector or by a 10° sector. The lengths of flash used with the
5° interval were 5°, 10°, and by 10° steps to 90°: with the 10°
interval, flashes of 2.5°, 5° and 10° were used, also several
greater lengths, up to 180° for one flash, the other being
shorter. The flicker discs (fifteen in number) had each two
apertures, with ratios of open to closed ranging between 1/35
to 35/1-
In working on myself the method was as follows : starting
with a speed of rotation such that , distinct doubleness (or
flicker) was observable, the speed was increased by small steps
until a single flash (or fusion) was obtained. Then, by
depressing a key a circuit was completed through the Ewald,
and a circuit-breaker on the spindle, and was allowed to
continue for ten seconds: thus the number of rotations in
ten seconds was registered. After recording the speed, it
was increased somewhat (the amount of increase at this point
being purposely irregular), and then decreased by small steps
1 That is, in the first case, the lamp at 36 cm. gave a brightness clearly brighter
than that of the Nernst beam, and at 37 cm. a brightness clearly less.
23°
KNIGHT DUNLAP
until the point of doubleness (or flicker) was reached, when
the speed was again measured. Three determinations were
usually made on each setting of the sectors (or each flicker-
disc), before proceeding to the next. The longer series of
settings, or series of discs, was gone through with in this
way from one to two hours, and as one such series a day was
all an observer could endure, the progress of the experiment
was necessarily slow.
The observer was instructed to make his judgment each
time rather quickly, and then look away until the speed was
changed. Continued gazing at the lighted area was found
to cause even a pronounced doubleness or flicker to disappear.1
The observations were made with darkness adaptation.
In working on myself, I took the speed readings, and made
the record, by a very dim light, and then waited for a minute
or so for readaptation. This procedure undoubtedly had
some effect on the determinations, but this effect was probably
not large.
The series of settings in the groups were taken in different
orders on different days, the several orders being carefully
TABLE I
FLICKER AND FUSION
Observer Dunlap
Sectors in Degrees
Cycles per Second
Durations in Sigmas
Flicker
Fusion
Closed
Open
Flicker
M.V.jS
Fusion
M.V.£
Closed
Open
Closed
Open
5
175
24.12
5-86 '
28.17
5-93
i. IS
40.29
0.98
3448
10
170
28.11
4.20
31-9
4-48
i-97
33-59
1-73
29-53
15
I65
31.08
5-Si
36.00
5-25
2.68
29.49
2.31
25-45
30
150
38.56
5.16
42.59
5.62
4-32
21.60
3-91
19.56
45
135
42.59
3-5i
46.25
4-15
5-86
17.60
5-40
16.21
60
1 2O
43.84
3-55
47-94
4-69
7.60
15.20
6-95
13.90
75
105
44.68
4.00
49.46
4-38
9-32
13-05
8.42
11.79
90
90
45-32
3-07
49-47
4-34
11.03
11.03
10.10
IO.IO
105
75
45^4
4.07
50.22
4.60
12.00
9.12
11.61
8.29
120
60
44.81
3-70
49-75
4-94
14.87
7-43
13-39
6.69
135
45
44-58
4-23
49.00
4-13
16.82
5.60
15.30
5.10
150
30
41.58
4.72
46.73
3-38
20.03
4.00
17.83
3.56
I6S
15
38.72
4-57
42.92
5-65
23.66
2.15
21.35
1.94
170
10
35-66
5-39
39-8o
4-98
2548
i-SS
23.72
i-39
175
5
29.61
6.42
34-04
6.06
32.82
o.93
28.55
0.81
1 This 'flicker adaptation' is not due to brightness adaptation, as later work shows.
PERCEPTION OF TIME-INTERVALS
231
TABLE II
TIME THRESHOLD. STANDARD BRIGHTNESS
Observer Dunlap
i. A=C, £=5.
A°=C°
Double
Single
Bo
M.V.*
A . Co
Bo
M.V.*
A = Co
5
19.8
8.0
19.8
I3.8
12.3
13.8
10
14.9
10.7
29.8
II. I
9-9
22.2
20
10.5
7-9
42-3
8.2
7-9
32.8
30
7-3
14.7
43-8
6.2
9.1
37-3
40
6-3
13-5
50.9
5-2
10.7
41.7
50
54
n.8
54-6
4-5
13-3
4-59
60
4.8
IO.I
58.5
4-i
II. 2
50.0
70
44
10.7
62.3
3-7
6.3
52.9
80
4-3
9-4
69.6
3-7
74
60.4
90
4.0
8.4
72.9
3-5
9-9
63.5
C, B
10.
A°=C°
Bo
M.V.0
Ao
Bo
M.V.0
Ao
2-5
5
10
28.1
27.0
2O.6
17.8
n.6
8-7
7-o
I31
20.6
20-9
19.2
15-9
I6.5
I3.8
II.7
%
15-9
3.
^°
Bo
M.V.jJ
^<r
Bo
M.V.jf
Ao
10
30
50
70
90
16.0
n.8
8.8
6.6
5-3
"-S
5-7
8.1
7-2
7.8
33-i
71.1
88.3
92.5
96.1
II.9
94
7.0
5-6
4.6
6.2
9i
4.6
8.6
8.1
23.8
56.7
70.8
78.9
84.4
= 10, B = 10.
c°
Bo
M.V.*
Co
Bo
M.V.jf
Co
20
60
IOO
140
1 80
20.8
22.2
24-3
23-9
21.9
10.6
18.4
8.9
II-J
7-5
41.7
133-5
243.6
335-2
393-3
16.3
16.8
18.1
18.0
15.6
7-8
9.2
1:1
13-9
32.6
117.7
181.3
252.5
280.8
planned to distribute the effects of practice over the whole
series.
The results of my observations are presented in Tables I.,
II. and III. In Table I. the average flicker-points and
fusion-points for the several ratios of open to closed sectors
are given both in cycles per second (i. e., the number of
complete changes from dark to light and back to light again
in a second); and also in the duration in thousandths of a
second, of the individual light and dark periods.
232
KNIGHT DUNLAP
In Tables II. and III. the average durations are given
for 'A' (the first flash), ' B' (the dark intermediate interval)
and 'C' (the second flash) when the flashes appeared dis-
continuous ('double'), and when they appeared as one uniform
flash ('single'). Table II. gives results of work with the
higher brightness described above; Table III., with the lower
brightness.
TABLE III
TIME THRESHOLDS, Low BRIGHTNESS
Observer Dunlap
A = C, B = 5.
A° = C°
Ba
M.V.#
A = Ca
Ba
M.V.It
A = Gr
5
30
70
20.5
8.9
4-5
10.7
7.6
7-8
2O.5
53-9
63-1
12.4
7.2
4.0
942
S.6
4.8
12.4
43-3
56.8
TABLE IV
TIME THRESHOLDS: Low BRIGHTNESS
Observer G. R. Wells
i. A = C, B = 5
A°= C°
Double
Single
Ba
M. V.#
A = C<r
Ba
M.V.0
A = C<r
10
30
SO
19.9
IO.S
7-1
II.O
9-3
10.0
39-9
63.1
71.7
13-9
8.1
5-5
57
10.3
10.7
27.9
49.1
55-6
2. C = 10,
A*
Ba
M. V. 0
A,
Ba
M.VA
Aa
30
50
I3-I
II. 2
n.8
9.0
78.9
II2.4
9-9
8.0
9.2
8.6
59-5
80.7
3-
A = 10, B
= S
30
SO
18.8
19-5
10.7
6-3
113.3
195.6
14.9
15.0
i.8.6
9-7
89.6
150.1
In Tables IV. and V. the results of observations of two
other persons are given. These observations were made
after I had finished mine, and it was not deemed necessary
to use all the flash-lengths which I had observed. In these
cases I manipulated the apparatus and recorded the measure-
ments, so that the observers worked under better conditions
PERCEPTION OF TIME INTERVALS
233
TABLE V
TIME THRESHOLDS
Observer H. M. Johnson.
i. Standard Brightness
A°=C°
Double
Single
Bo
M. V.0
A-C.
^0-
M. V.j<
A = Ca
5
10
30
50
70
16.6
94
7.0
5-2
4-5
14.2
12. 1
10.0
8.0
8.6
16.6
18.8
42.2
52.4
63.1
I2.I
7.8
5.8
4-3
3.8
9-3
12.2
8.9
4.8
9.1
12.1
I5.6
34-8
434
54-5
2. Lower Brightness
5
30
13-2
7.0
7-i
I3-I
13-2
42.0
10.3
5.8
5-7
9.2
10.3
35-2
than those under which I observed, specifically as regards
adaptation.
Each of the values given in Tables L, II., III. and V.
are averages of twenty-five thresholds. The values in Table
IV. are averages of twenty thresholds.
There are two points of importance which stand out in
these data. First, the rise in rate of the 'critical frequency'
(flicker and fusion points) from the extreme inequality to
equality of open and closed sectors, in both directions (Table
II.). Second, the decrease of the time threshold with increase
in the length of the first flash (Tables II., IV. and V.). This
decrease seems to be altogether a function of the first flash;
increasing the length of the first flash with the second flash
constant (II., 3; IV., 2) has almost the effect of increasing
both flashes; while increasing the length of the second flash
(II., 4; IV., 3) alone has practically no effect. The slight
increase in the threshold in both these cases is due to the
increased difficulty of observation with the unequal length
and hence unequal appearing brightness of the flashes.
In addition to these points, it is to be noted that the time
thresholds are low, ranging (with equal flashes) from 4 to 20
sigmas. The comparison of these figures with those obtained
in other experiments is, however, not now significant, since
we have not as yet analyzed the various factors entering into
234 KNIGHT DUNLAP
determinations of this sort. In addition to the brightness,
in regard to which the a^bove data are not significant, the
factor of adaptation is probably extremely potent. These
results were obtained with fairly good darkness adaptation;
they cannot be compared with results obtained with daylight
adaptation.
Among the factors affecting the formation of judgments,
the rapid repetition of the pair of flashes was conspicuously
disturbing. The simple rotation apparatus is not suited to
determinations of this kind.
WORK ON BRIGHTNESS AND ADAPTATION
The second set of experiments I varried on, at Dr. Hyde's
invitation, in the Nela Research Laboratory during the
summer months of 1914. In carrying out these experiments
I received much help from the staff of the laboratory, and I
am especially indebted to Dr. Hyde, the director of the
laboratory; to Mr. Cady, assistant director; to Dr. Lorenz;
to Dr. Cobb; and to Dr. Johnson. The readiness of the
members of the staff to give their time to my problems, and
to release to me apparatus from their own experiments, made
possible such work as I was able to accomplish in the short
time I was there. The greatest burden of the observations
fell on Mr. Eric Martienssen, to whom I am indebted for
his careful and willing work, under conditions which were
sometimes trying.
My apparatus, which need not be described in detail,
consisted of the following units.
(a) A double rotator,1 carrying on one axis of rotation
two arbors; one on the main shaft and the other on a sleeve
on that shaft, the sleeve being geared to an auxiliary shaft
and that back to the main shaft, so that the sleeve made
one rotation for nine of the shaft. The arbor on the sleeve
carried a large metal disc with a 4<D-degree aperture. Variable
cardboard sectors were carried by the faster moving arbor.
When the axis of light is parallel to the shaft of this apparatus,
1 This piece of apparatus was made under my direction in the Physics workshop
of The Johns Hopkins University.
PERCEP TION OF TIME-IN TER FALS 235
whatever exposures are arranged through sectors on the
faster arbor are repeated every ninth revolution of the shaft,
being cut out the remainder of the time by the slow moving
disc.
The main shaft carried also a loose gear, in mesh with a
gear on the driving motor; with an electro-magnetic clutch
of my devising, so that the disc and sectors could be stopped
for adjustment without stopping the driving motor; and
could be started again without jerk by turning the current
gradually on the clutch magnet. The same shaft also carried
a cylinder of brass and hard rubber, on which rested two
brass brushes, so that the rotations could be counted by a
step-up mechanism operated by the make-and-break of the
circuit.
(b) A nitrogen-filled lamp, with the wire in a straight
compact coil; operated in these experiments at 85, 120, 200
289 and 400 watts. The image of the coiled wire was focused,
by suitable lenses, on a slit, to cut off light reflected from the
surface of the lamp bulb; and by other lenses refocused in the
plane of the sectors carried on the faster moving arbor of the
rotation apparatus described above. The axis of the beam
of light was parallel to the shaft of the rotation apparatus,
and the long axis of the image was radial to the shaft.
The lamp, the slit and the lenses were enclosed in a large
hood of black felt drawn over a wooden framework, with an
aperture just large enough for the convergent beam to emerge.
(c) A movable screen located just beyond the slow-moving
disc of the rotator and operated by a hand lever. By raising
this screen shortly before the disc made an exposure, and
lowering it shortly afterwards, a single exposure of the interval
arranged through the sectors was allowed. The manipula-
tion of this screen required no accurate timing, since the slow
disc allowed exposure every ninth rotation of the sectors only.
(d) A lens, just beyond the hand screen, decreased the
divergence of the cone of light, increasing the brightness of
the surface illuminated.
(i) A plaster disc, surfaced with magnesia, illuminated
by the cone of light. This disc was 12.5 cm. in diameter, and
236 KNIGHT DUNLAP
about half the diameter of the light cone at the point of
insertion of the disc and had a background of black velvet
upon which the light around the disc was negligible. The
plane of the disc was vertical, but was at an angle of 30° from
the plane perpendicular to the axis of the cone of light.
The observer sat so that his binocular line of sight was
perpendicular to the disc, which was about 165 cm. from his
eyes.
(/) A miniature projection lamp, with a small incan-
descent bulb entirely enclosed, established out of the range
of the observer's vision, cast on the object disc a group of
four small dots, which served excellently as a fixation mark.
(g) Eight mazda lamps, totalling 600 watts, so disposed
in the room that the walls were illuminated, especially the
wall in front of the observer — the wall behind the plaster
disc — but the lamps were screened from the observer's eyes.
These lamps were controlled by a single switch.
(h) A single mazda lamp in a long black cardboard tunnel,
arranged to throw continuous illumination, when desired, on
the disc.
(i) An Ewald chronoscope, for counting the rotations of
the sectors, as a control of the accuracy.
(/) A synchronous motor1 for driving the rotator. This
had eight poles, and working on 60 cycle A.C. current gave
15 rotations per second. This motor was geared to the main
shaft of the rotator (a), the ratio to the gear on the motor
shaft being I to 3. The main shaft therefore made five
rotations per second, so that for the sectors carried by the
arbor on the main shaft 9° equalled 5 a. The variations in
speed, due to variations in current frequency, were negligible
during the periods of work.
(k) A small D.C. motor for starting the synchronous
motor. This starting motor was belted to a one-flanged
pulley on the shaft of the synchronous motor so that the belt
could be thrown off when the synchronous motor was working
properly. A stroboscope disc mounted on the same shaft,
and illuminated by a 15-watt lamp on the A.C. current,
1 This motor was one which Dr. Lorenz had constructed for his use. The strobo-
scopic method of starting the motor was also suggested by him.
PERCEPTION OF TIME-INTERVALS 237
indicates the proper moment for turning on the synchronous
motor.
The five wattages used on the lamp gave brightnesses on
the object disc of 3, 10, 36, 82 and 168 candles per square
meter. This range of illuminations seemed adequate for the
investigation of the effects of brightness, which was the first
point I had planned to attack.
The results of the preliminary experiment, reported
above, had shown clearly that the threshold for doubleness
(measured in terms of the dark interval) depends on the
length of the flashes, especially of the first flash, although the
absolute magnitudes of the thresholds as determined in those
experiments could not be supposed to be very significant.
It would therefore be possible, theoretically, to determine
thresholds in either of two ways: first, by keeping the dark
interval constant and varying the flashes; and second, by
keeping the flashes constant and varying the dark interval.
It would seem equally useful to work out the thresholds in
flash length for several fixed dark-interval lengths, and to
work out the thresholds in dark-interval length for several
fixed flash-lengths. In either case the effect of the brightness,
and of adaptation could (it would seem) be worked out in
an adequate way.
In the manipulation of apparatus, the first procedure is
far the simpler. The sector adjustments are not so com-
plicated, and hence the progress of the experiment should be
more rapid. Realizing that the work would at best be slow,
I chose the plan which offered this important advantage.
Observations were made at first with dark adaptation
exclusively. The subject was kept in the room from ten to
twenty minutes before commencing work, according as he
had come in from outdoors, or from more or less dimly
lighted work rooms. No warning signal other than the
normal sound of the electro-magnetic clutch in taking hold,
was needed by the observer. The motor ran continuously,
and the clutch was thrown in when an observation was
desired. The rotator t picked up' full speed in less than a
second; the hand screen was lifted about two seconds after
23 8
KNIGHT DUNLAP
the clutch was thrown in. The four dots of light in the
center of the disc fixed the line of sight before the flashes
occurred. Four repetitions of the exposure were given in
succession, but the observer usually gave his judgment after
the second or third.
Observations were carried on for some time by Martienssen
and myself by this method, using the procedure of 'serial
groups/ but the results, although interesting, were of little
value for the purposes of the experiment. Variations in the
durations of the flashes produced variations in the apparent
brightness and apparent color of the disc, which were at first
extremely confusing, and on which finally the judgments came
to depend, rather than on any real appearance of ' doubleness'
or * singleness.' One set of five hundred judgments by
Martienssen, which are typical, are given in Table VI.
TABLE VI
Martienssen
Dark interval 250-. Brightness 82 c. per sq. m.
Two Flashes
Double
Single
One Flash
Double
Single
70<r
60
60
40
30
48
36
38
32
41
/>
H
12
18
9
I40a
120
100
80
60
0
0
3
II
12
SO
SO
47
39
38
In this table the first column gives the length of flash where
two were used, the second and third columns giving the
number of judgments of double and single for each flash-
duration. The fourth column gives the durations of the
single flashes, each equal to the sums of the two in the corre-
sponding pair; the fifth and sixth columns giving the number
of judgments of single and double respectively for each of
these single flash durations.
The increasing difficulty of discrimination is here shown,
not so much by the increased tendency to call the double
flashes single, as by the tendency to call the single flashes
double. Obviously, no definite threshold can be determined
PERCEPTION OF TIME-INTERVALS
239
when this tendency is present.1 This tendency, it must be
noted, is not due to mere confusion; as we shall see later, a
single flash often appears distinctly double, and with the
same sort of doubleness as is noticed in a really double flash.
In this particular set of observations, however, the judgments,
according to the observer's report, were based largely on
differences in apparent brightness and color; at least this
seemed to him to be the case in the latter part of the set.
TABLE VII
Martienssen
Dark interval 250-. Brightness, 3 c. per sq. m.
Two Flashes
Double
Single
One Flash
Double
Single
50<r
26
4
Io6<r
3
23
50
16
4
80
4
16
30
18
2
60
2
18
20
15
5
40
4
16
IO
IS
5
20
ii
9
The results of a set of observations by Martienssen with
lower brightness are given in Table VII. Results of a set
of observations by Dr. Johnson on the moderate brightness
are given in Table VIII. Other sets with different brightnesss
gave results of the same order.
TABLE VIII
Johnson
Dark interval 250-. Brightness, 82 c. per sq. m.
Two Flashes
Double
Single
One Flash
Double
Single
Soa
40
30
20
20
17
23
IO
0
3
7
10
IOO<r
80
60
40
I
7
4
5
19
11
15
In this set of observations, the observer's judgment was
influenced very largely by the apparent duration of the total
xThis condition is similar to that found in attempting to determine the 'two
point' threshold by simultaneous stimulation of the skin. No threshold can be de-
termined, since one stimulation frequently is perceived as two, and hence the least
separation of two points giving a certain percentage of perception of two has no definite
significance.
240 KNIGHT DUNLAP
exposure; the greater duration of the two flashes was noticed,
especially with the shorter flash-lengths, and this tended
more and more to become the criterion of doubleness.
The procedure by groups (£ method of serial groups'), it
was clear, could not be used in this experiment. The secon-
dary criteria — in this case the differences in brightness, color,
and duration — are made maximally conspicuous by this
method, and judgments strictly on the points under examina-
tion are made practically impossible. I therefore attempted
to use the shuffled series procedure, still clinging to the
method of constant dark interval. A few series, however,
showed that this method was not practicable, even when the
better procedure was employed, since the differences in
brightnesses and color still were very conspicuous. The
regular progression procedure ('method of minimal change')
accentuated these secondary criteria still more.
The effects of the total duration had been foreseen, and I
had expected to introduce variations in which the single flash
should be equal in length to the two flashes plus the dark
interval. This variation was found to be inapplicable because
it would have accentuated the brightness differences. For
example; the greater brightness of the 100 a flash as com-
pared with the two successive flashes of 50 a with 25 a dark
interval, would be still greater if the lengths of the two flashes
were reduced to 37.5 a each.
The next attempts were made by the method of constant
flash-length, using the ' shuffled series' procedure. With this
method the differences in apparent brightness are not so
marked as with the constant dark-interval method, and by
this procedure these differences and the differences in dura-
tion are not so disturbing as they are in the serial group
procedure. It is possible, in other words, to form judgments
on the apparent doubleness or singleness alone of the flashes,
although it required a high degree of training in order to
eliminate absolutely other criteria.
The results of these next observations by the shuffled
series procedure are given in Tables IX., X., XL and XII.
In these tables the first column gives the separation of the
PERCEPTION OF TJME-INTERVALS
241
two flashes, and the other columns give the number of
judgments of Single' and i double' for each of the five
brightnesses. The observations with all of the brightnesses
were obtained on the same days, a series being taken with
each brightness during each experimental period, the order
of brightness being altered from day to day in a regular way.
TABLE IX
Martienssen
Flash = 500-
Brightnesses, Candles per Sq. Meter
Intervals, <r
3 | 10
36
82
168
d.
s.
d.
s.
d.
s.
d.
s.
d.
s.
0
II
13
5
15
9
IS
II
9
5
19
5
8
4
5
5
9
3
5
5
7
5
IO
6
6
7
3
5
7
7
3
8
4
15
9
3
5
5
9
3
6
4
9 3
20
IO
2
4
6
ii
i
8
2
6
25
ii
I
8
2
10
2
10
0
8
4
TABLE X
Johnson
Flash = 500-
Brightnesses, Candles per Sq. Meter
Intervals, <r
3
10
168
d.
s.
d.
s.
d.
s.
0
4
16
0
20
0
20
5
S
5
O
10
2
8
10
7
3
2
8
j
7
IS
8
2
4
6
8
2
20
10
0
9
i
9
I
25
10
O
8
2
10 | O
Other series were taken with light adaptation. In this
work the room was lighted by the mazda lamps referred to
under (g), and the observer was adapted to the brightness of
the plastered wall due to this illumination. When ready to
make the observation, the lights were switched off, approxi-
mately 1.5 seconds before the exposure of the flashes (or
flash). This interval was timed by watching the exposure
on the hand-screen; and turning off the lights immediately
242
KNIGHT DUNLAP
after such exposure. Then the hand-screen was lifted, and
since the exposure occurred every 1.8 seconds (the rotation-
period of the slow moving disc) the interval between the
turning off of the mazda lamps and the exposure on the object
disc was timed sufficiently well.
TABLE XI
Martienssen
Light Adaptation. Flash = 500-
Brightnesses, Candles per Sq. Meter
Intervals, <r
3
36
168
d.
s.
d.
s.
d.
s.
O
5
25
6
24
3
27
5
2
17
12
13
6
14
10
IS
6
15
S
14
16
IS
17
2
18
2
19
i
TABLE XII
Johnson
Light Adaptation. Flash — SO<T
Brightnesses, Candles per Sq. Meter
Intervals, a
3
36
168
d.
s.
d.
s.
d.
s.
0
0
25
2
23
S
20
5
5
IS
12
13
9
II
10
13
6
10
10
18
2
15
24
i
16
3
19
I
This method of working with light adaptation seems quite
satisfactory. An interval must be allowed between the
turning off of the adaptation light and the beginning of the
stimulus light, to allow muscular recovery. The one-and-a-
half second period seemed to be about the shortest which
could be used. Of course a slight amount of adaptation
occurs within this period, but this is kept constant through-
out.
Series with darkness adaptation followed the work with
light adaptation. Results of one group of series on Mar-
tienssen are given in Table XIII. The remainder of the
PERCEPTION OF TIME-INTERVALS
243
work on this observer and on Dr. Johnson was directed to
'feeling out' methods, and does not lend itself to tabulation.
TABLE XIII
Martienssen
Dark Adaptation. Flash = 25 <r
Intervals, <r
Brightness, Candles per Sq. Meter
3
36
168
d.
s.
d.
s.
d.
s.
0
IS
20
25
7
9
13
14
7
S
i
o
3
9
13
14
II
S
i
o
I
9
8
II
13
I
3
From the results, as tabulated, little can be inferred as
to the effect of brightness. It is evident that adaptation is
an important factor. The factor of greatest consequence,
however, is the tendency to see the single flash as double.
The effects of this tendency are found in the tabulated
results, especially with the lowest brightness, and were still
more evident in the work not tabulated.1 Attempts to use
flashes longer than 50 a proved fruitless on account of this
tendency. At 75 cr, for example, there was a large increase in
the number of c double' judgments on single stimuli. There
is a limit, however, beyond which the double appearance is not
found. It may be useful, later, to determine both the upper
limit and the lower limit for the fallacious doubling, but this
is a determination of the most difficult sort.
The double appearance of the single flash may, with prac-
tice, be distinguished from the true 'doubleness.' That is,
there are times when the 'doubleness' of a single flash is
clearly different from the 'doubleness' of two successive
flashes, if the one and the two are shown with but little
pause between. This discrimination is apt to be lost at any
time, however, and the pseudo-'doubleness' taken for real
'doubleness.'
As an illustration of the discrimination, the following
observation of Dr. Johnson will serve.
1 In many cases, both with one flash and with two flashes, the appearance was
'double' on first exposure and 'single* on the succeeding exposures.
244 KNIGHT DUNLAP
1. With the brightness = 36 (c. per sq. meter), two 400-
flashes were distinguished from one 80 <r flash when the
interval was 20 <r; with 15 a interval, the one flash and the
two flashes looked equally double.
With brightness = 3, the two were distinguished from
the one with 150- interval.
With brightness = 168, discrimination was clear at 25 <r;
not at 20 0-.
2. With two 25 o- flashes, and one 50 <r flash, the differ-
ence was clear when the interval was 35 o-, with all the bright-
nesses, equal 'doubleness' at 30 a*.
3. With two 10 <7 and one 20 <r flashes, the discrimination
was clear when the interval was 55 o- for the 3 and 36 bright-
nesses, and 40 a- for the 168 brightness. Below these points
the 'doubleness' was the same.
Similar observations by Dr. Cobb gave the following
results:
1. Two 25 ff flashes and one 500- flash, with brightness
= 3, 'doubleness' clear at 400- interval. With brightness
36 and 168, 'doubleness' clear at 30 cr interval.
2. Two 50 <r flashes and one 100 a flash, with bright-
ness = 3, clear at 20 a.
With brightness = 36, clear at 25 a.
With brightness = 168, clear at 30 a.
3. Two 10 a flashes and one 20 a flash, with 3 and 36
brightnesses, not clear below 50 a (no longer interval used) :
with 168, clear at 45 <r.
4. Two 75 o- flashes and one 1500- flash, brightness = 3,
clear at 10 a interval. Brightness = 36 and 168, clear at
15 <r interval.
On the whole we cannot conclude that increasing the
brightness of the flashes increases the distinction of the
doubleness of two. This is a matter that is dependent upon
the absolute length of the flashes. In subsequent work,
carried out on the two observers listed above, and on Dr.
George R. Wells, the effect of brightness was brought out
directly by trying various intervals in succession with the
same flash-lengths. This work, while agreeing with that
PERCEPTION OF TIME-INTERVALS 245
reported above, brought out the further fact that the effect
of intensity variations on successive flashes which are hardly
distinguished at best because of the shortness of the interval,
is not the same as the effect on succession with longer inter-
vals.
These observations are not consistent with the tabulated
results, but there is no reason why we should expect them to
be so, since the conditions of observation were entirely dif-
ferent. We must always distinguish in problems of this kind
variations in the actual observable phenomena established
by the experimental conditions, and the variations in the
observations of these phenomena which may be due to the
same conditions. For example: the sensible content from
two (successive) stimulations may be different from the
sensible content due to a single stimulation, and yet on
account of the circumstances of observation, the difference
may not be noted. On the other hand, a sensible content of
a certain sort may now be judged like, now be judged different
from, a content from which it differs slightly, according as
the conditions of observation throw this difference in relief,
or minimize it.
MY OWN OBSERVATIONS
During the course of the experiments reported above, I
acquired a considerable facility in observation, since I
watched the flashes while having full knowledge of the
stimulus conditions. I did not record my observations during
the work with the other observers, since the necessity of
conducting the experiments for them, and especially the
attention to speed and accuracy in the adjustment between
exposures, was a disturbing factor.
Later I made observations (with knowledge) myself under
satisfactory conditions. In these cases I worked with the
'progressive procedure,' starting alternately with a setting
(width of dark interval) giving no doubleness, and one giving
doubleness. •
This work was done at night, and the results on different
nights did not agree absolutely. There was, however, a
246
KNIGHT DUN LAP
general uniformity, such as is indicated in Table XIV., in
which are given the results on four nights during August.
The figures given are not averages, but absolute values in the
scale of 5 a steps; the points at which (and above which) the
flashes were always seen 'double' (d.) and at which and
below which they were seen 'single' (s.) on that night under
TABLE XIV
Dunlap
Flash = 500-
i. Aug. 4
Brightness
Dark Adap.
Light Adap.
Constant Light
d.
s.
d.
s.
d.
s.
10
36
82
168
40
40-50
40-50
40
0000
2O
20
20
IO-30
5
5
5
5
2O
2O
20
20
5
5
5
5
2. Aug. 22.
Brightness
Dark Adap.
Light Adap.
d.
s.
d.
s.
3
10
36
82
168
25
25
25
25
25
IS
IS
15
IS
15
2O
15
5
5
10
IO
5
3. Aug. 23
Dark Adap.
Light Adap.
Constant Light
Brightness
d.
s.
d.
s.
d.
s.
3
20
10
20
S-io
5
2-5
10
2O
10
10
?
5
2.5
36
20
5-IO
5
?
10
S
82
20
IS
5
?
10
5
168
20
10-15
5
?
10
5
the conditions indicated. When the threshold varied during
the test, the variation is indicated. The observation lasted
from one to two hours, with periods of rest for the eyes.
In certain cases, no definite determination was made for
the 'single' point. This is indicated by a question mark.
m
PERCEPTION OF TIME-INTERVALS 247
The series with dark adaptation and light adaptation were
taken as in the work on other observers. The results in the
columns under 'constant light' were obtained while the
object disc was illuminated by the 'tunnel lamp' described
above, (h). In this case, the flashes were superimposed on a
constantly lighted surface. Except for the illumination of
the disc, the room was dark during these observations.
The observations included in Table XIV. were with 50 a
flashes only; with 25 a the results were more uniform; for all
brightnesses, with dark adaptation, the double point was at
40 or, the single, at 30 cr; with light adaptation, the points
were 20 cr and 10 cr respectively; with light adaptation and
constant light in the disc, 20 a and 5 a. With dark adapta-
tion and constant illumination, the single point was 5 a, but
the double point was variable (lOcr— 20 cr). Flashes above
50 o- (up to 75 d) gave more variable results.
The general influence of light adaptation and constant
illumination was demonstrated on a number of persons,
including the observers listed above, by a simple method.
The sectors were set so that with dark adaptation the two
flashes appeared ' single,' or the judgment was 'doubtful.'
Then the eye was light-adapted for a short time, and observa-
tion showed a striking change, it being possible with any
observer to change the judgment from 'distinctly single' to
1 distinctly double' by this means. The addition of a constant
illumination served the same purpose. With certain settings
of the sectors, and a faint constant illumination on the disc
the two flashes appeared 'single'; by increasing the constant
illumination a point was reached at which the appearance
was clearly double. This point varied with different obser-
vers, and at different times.
THE SOURCES OF DIFFICULTY
The results of the investigations of the visual time
threshold up to this point are as follows:
I. The effects of brightness of the light are variable,
depending on the other factors in such a way that no con-
clusion can be drawn as yet concerning their effects.
248 KNIGHT DUNLAP
2. The threshold is lower for the light-adapted eye than
for the dark-adapted eye. This holds, at least, for certain
light-adaptations.
3. The threshold is lower for an interval marked by
flashes added to a continuous stimulation, than flashes in a
dark field. This holds for a wide range of constant illumin-
tion, the threshold varying usually with the brightness of
the constant illumination up to the point where the additions
lose in distinctness.
4. A single flash is frequently seen as a succession of two,
and although this 'twoness' may, under proper conditions,
be discriminable from actual 'twoness/ these conditions are
not easily actualized in quantitative work.
In consequence of this (and, possibly, other factors)
quantitative work by the standard methods is not possible;
at least the results of such work are unreliable. Special
methods must be devised.
5. It is impossible to train observers on the light threshold
problem in a limited time (two or three months). Observa-
tions are of value only if made by persons having a long
training in that particular work. In this respect, the time-
threshold problem differs markedly from certain other prob-
lems, e. g., of flicker. The length of training required cannot
be specified, but possibly should extend over a period longer
than a year.
The most interesting question coming out of these observa-
tions concerns the apparent doubleness of a single flash under
certain conditions. This doubleness of appearance is un-
questionable; the flash has at times a striking £ one-two'
progression.
This fictitious doubleness is not exclusively a dark-
adaptation phenomenon, although it is less noticeable with
light-adaptation. Constant illumination, on the other hand,
even of relatively low brightnesses, completely abolishes it.
We might therefore suppose it to be due to an iris-reflex: the
stimulation beginning with dilated iris causes a strong con-
traction and immediate relaxation, so that the light-flux
entering the eye drops and rises again causing a depression
PERCEPTION OF TIME-INTERVALS 249
(' dimple') in the excitation curve of the retinal process in the
same way as in a rapid succession of two flashes.
The occurrence of the flash provokes a strong visual reflex,
noticed by every observer. One feature of this reflex is an
increase in accommodation: at the end of the stimulation,
this accommodation is for a point nearer than the object-disc,
and the relaxation necessary to re-accommodate for the disc
is easily noticed. Since accommodation and iris-contraction
go together this may be taken as indicating the iris factor
suggested above.
On the other hand, the chief factor may be retinal. The
retinal process may rise to a point higher than its 'normal'
for the intensity of stimulation, and then drop back.1 The
drop may be below normal, with an immediate second rise;
thus the 'dimple' which normally produces the appearance
of doubleness may occur independent of iris-activity.
It is possible that no 'dimple' may be required. The
two drops in the sensation, — one following the excessive rise,
and the other at the end, may be interpreted as 'twoness.'
The motor-process — adjustment of the eyes — may be
connected with the fictitious doubleness through an actual
inhibitory discharge to the retina accompanying the discharge
to the ciliary muscle. Efferent fibers to the retina are known
to exist, although their function is not known.
The motor-process is probably the cause, or connected
with, the severe effect of the observations. Both Martienssen
and myself felt the effect to a marked degree, the eyes becom-
ing very irritable, and necessitating frequent interruptions
of the work.
Instead of being towards the end or in the middle of a very
simple experiment, or small group of simple experiments, we
are now at the place where it is necessary to take up a large
number of points, not so clearly connected with each other
as they are contributory to the solution of our initial problem.
If any light is to be thrown on these problems, it can come
1 Such action of a light stimulation on the retina is called by physicists the
'over-shooting of the sensation.'
250 KNIGHT DUNLAP
only through the solution of these various problems, each of
which involves an extended investigation.
The problem, or group of problems, which stand out
above the others in importance, concerns adaptation. I am
now installing apparatus and developing methods which
may throw new light on this topic.
VOL. XXII. No. 4 July, 1915
THE PSYCHOLOGICAL REVIEW
AN EXPERIMENTAL CONTRIBUTION TO THE IN-
VESTIGATION OF THE SUBCONSCIOUS1
BY LILLIEN J. MARTIN
Leland Stanford Junior University, California
The subconscious is so often referred to and so little
attention has been given to investigating it experimentally
that it has seemed to me a condensed summary of a recent
investigation I have made, might possibly be of some interest.
In making this study the image method was employed.
That is, to state very briefly the mode of procedure: in one
half of the experiments the observer, usually with his eyes
closed or blindfolded and seated opposite the experimenter,
was instructed to sit in a relaxed position and let an image
(visual or auditory, memory or imaginative, etc., depending
upon what was desired by the experimenter), arise of itself.
The observer was not only not to arouse the image but he
was not even to know its content until he saw it before him,
and only those images were noted where such instructions
had been entirely complied with. In the other half of the
experiments, the observer was directed to arouse the image,
that is, for example he was instructed to decide on the partic-
ular thing he wished to visualize and to arouse the correspond-
ing image.
Stanford and Munich University students acted as ob-
servers.
An examination of the data regarding the content of the
images, their mode of arising, etc., shows:
1 For fuller details of the investigation as to the theory underlying it, the methods
used, the experimental data, etc., see Martin, 'Ein experimenteller Beitrag zur Er-
forschung des Unterbewussten (Earth) and Uber die Abhangigkeit visueller Vorstel-
lungsbilder vom Denken,' Zeit. fur Psych., 70, 212.
251
252 LILLIEN J. M4RTIN
1. The subconscious mental activity reveals itself through
the arising of images where the observer did not previously
know whether anything would be imaged, or if so, what it
would be ; also, in the arising of unwilled (spontaneous) images
in connection with those willed.
2. Evidently, sometimes and in some persons, the sub-
conscious thinking responds more quickly to the task set than
does the conscious. This is shown by the spontaneous images
arising more promptly than do the willed. That is, the spon-
taneous image is before the observer before he has decided
what image to arouse or it arises in place of it.
3. The images show that not only the conscious but the
subconscious mental activity differs in richness of content
in different individuals.
4. In case of all the observers — but in some of them more
than in others — some of the material stored away under the
threshold has evidently remained as originally grouped, as
for example, when the visual image of a particular man in a
particular environment arises simultaneously and at once.
On the other hand, some of the material has evidently been
more or less broken up, as for example, where an eye arises
spontaneously, when an imagination image of a face is asked
and no other features followed it until aroused by a special
act of will on the part of the observer. In case of some of the
observers the broken-up memory material, the memory
elements, have been unconsciously (as shown by the observer's
great surprise at the content of the visual images which arise)
recombined under the threshold into complicated and ap-
propriate new groups. There has been not alone a breaking
up of memory material but, to use Ribot's words, an 'un-
conscious elaboration' of it. In the observers with whom I
have experimented, the memory activity evidently predomin-
ates both below and above the threshold of consciousness.
5. The memory and imagination material under the
threshold is evidently not all on the same stratum or level as
regards consciousness, for some of it arises much more spon-
taneously and quickly and has a different content. Here too
individuality plays a great role.
INVESTIGATION OF SUBCONSCIOUS 253
6. From what has been said it will be seen, that the
image method makes it possible to obtain information regard-
ing the past life of the individual, the general character and
the personal peculiarities of the thinking going on in his
mind, not alone above but also below the threshold of con-
sciousness. The applicability of this method in the case of a
particular person will of course depend upon his ability and
habit as regards the imaging of his conscious and subconscious
thinking.
7. The introspections show that the spontaneous images
are sometimes the point of departure of the willed images,
that is, the involuntary image that arises before the observer
has decided what to will acts in the way of suggestion. This
shows how important the spontaneous images must be in our
daily life. Where the spontaneous images are in the direction
of the work in hand, they must save time in that they arise
immediately and furnish material already elaborated. On
the other hand, if they are not of such a character that they
can be used directly in the intellectual work being carried on
or as points of departure for conscious thinking along the
desired line, they must be an interruption and even a hindrance
in the continuing of such thinking. The results show also that
the spontaneous images may furnish ideals as regards action.
In this respect they may and may not be entirely helpfuL
One of the observers who took part in these experiments, has
very strong and insistent spontaneous auditory images.
So insistent are they, that she tells me that they led to her
giving up the study of music to which she had devoted several
years, and turning to a totally different field of work. She
says, that whenever she plays on the piano the spontaneous
auditory images precede what she is playing and show her
how imperfect is her execution.
8. A comparison of the content of the voluntary images
with that of those which are spontaneous, shows that in the
case of the visual images of a given observer what is above
and below the threshold of consciousness is not materially
different.
254 LILLIEN J. MARTIN
This result does not support BinetV theory regarding the
nature of the subconscious, which is, that there are two per-
sonalities running side by side, one above and the other below
the threshold of consciousness, as what is above and below
the threshold of consciousness, as was said, seems in the case
of these observers not to be materially different. It may
be otherwise in pathological persons, of course. Cases of
double personality certainly suggest this. But such special
cases do not give Binet's theory any great universality.
Nor does Meyers's2 theory, which has found support among
workers in psychical research, that the subconscious is an
expression of the infinite mind, and the conscious an individual
matter or a very limited expression of the infinite, get support,
for, as was just said, what is under the threshold does not
seem enormously richer in content than what is above. Nor
do I find anything in these results which leads me to suppose
that under the threshold a mental condition exists which
makes it necessary to suppose that communication between
different persons (telepathy) is possible and which would
more or less support Meyer's theory. The results do support
Prince's3 theory that what is under the threshold is an
expression of the observer's previous experiences.
9. The results have a farther interest from the standpoint
of general psychology.
A. They show that the differences and likenesses between
spontaneous and voluntary images ought not to be over-
looked in psychology, as has been the case in the past, since
through the study and comparison of such images we may go
below the threshold of consciousness and get information
regarding what is going on there.
B. They throw light on what is called inattention and
vacillation of attention. We see that, sometimes at least,
this grows out of the fact that the person has a flood of spon-
taneous images and ideas, which impede and even crowd out
voluntary images and ideas. They explain why the genius
is so impatient of restraint and may sometimes actually get
1 On double consciousness, etc.
2 'Human Personality,' I., 34 ff., 1904.
3 'The Subconscious,' I ff., 1914.
INVESTIGATION OF SUBCONSCIOUS 255
on faster by letting himself go, and also why the student in a
field of an exacting and foreign character as regards his natural
thinking, must take himself in hand or fail altogether in his
work. I take the following in the way of illustration from
what one of the observers gave to protocol: —
"Als sich mein Studium begann, war es mir kaum moglich,
mich in einer Vorlesung irgendwie zu konzentrieren, weil ich
bestandig durch spontan auftretende V. gestort wurde. Ich
have dann versucht, die spontanen V. zu verdrangen und
grosse Miihe darauf verwendet und habe es darin bis zu ein^r
gewissen Fertigkeit gebracht, so dass ich jetzt spontane V.
willkurlich haben oder nicht haben kann. Sobald ich mich
aber etwas gehen lasse, sind die spont. da, und ich bin ziemlich
machtlos dagegen."
C. The data obtained lead one to ask whether in future
memory investigations along quantitative lines the task of
the investigation will not be something more than a filling
in of the gap left in the work of an Ebbinghaus and a Miiller,
something more than a building upon the results already
obtained by them. May we not possibly be obliged to begin
again at the very bottom and repeat the work in order to
feel sure of its foundations. It would seem from these results
that instructions given by an experimenter favorable to
voluntary effort, or the belief on the part of the observer that
he must put forth his will in connection with the task set,
while favorable to voluntary memory may have been detri-
mental to spontaneous memory and vice versa. In short, it
does not seem entirely impossible that two persons may have
equally good memories as regards the amount that can be
reproduced, but that like instructions, as for example, that
effort (resp. no effort) is to be used in reproducing a given
material, may make it appear that one person has a much
better memory than the other or indeed that neither has a
good memory.
D. Again, these results put in question the results of
certain experiments of Rux,1 which were inspired by Ach.
Rux has attempted to measure the strength of will by using
1 'Ueber das assoziative Aequivalent der Determination,' Untersuchungen zur
Psychologic und Philosophiey Bd. II.
256 LILLIEN J. M4RTIN
the quantitative data derived from memory experiments
without apparently making any attempt to show how much
of the work done was accomplished by voluntary and how
much by spontaneous memory.
10. The results have a pedagogical interest. .
A. In that they show that it is possible to educate and
enrich the subconscious.
B. In that they lead one to ask whether we may not some-
times be placing too much emphasis on the employment of will
in connection with the intellectual work to be done. When
the student's work is of a creative nature or along the line of
discovery and his spontaneous thinking and images are in
harmony with the field in which he is working, one can think
that the director of a leading institution in America which is
devoted to scientific research, showed psychological acumen,
when he urged the investigators working under him to take
each day some time away from their work not only to give
their minds rest but to free themselves from the restraint of
thinking in one particular narrow line.
THE IMAGE METHOD VERSUS THE AUTOMATIC WRITING AND
SPEAKING METHODS OF PENETRATING BELOW THE
THRESHOLD OF CONSCIOUSNESS
Binet and others have used the automatic writing method,
in investigating the subconscious. As the image method
will naturally come in competition with the automatic writing
method in investigations along this line, I have thought it
desirable to make some experiments by this method to ascer-
tain how it compares as regards the amount of data yielded
with the visual image method in the getting of information
of what is going on under the threshold of consciousness.
The experimental results I have given in the work of which
this paper is a summary. They show (i) that while theo-
retically the subconscious experience is reproduced through
automatic writing without entering consciousness, to be cer-
tain that this actually occurred, that is, to be certain that the
experience did not enter consciousness and after such entrance
more or less influence and direct the writing, one must have
INVESTIGATION OF SUBCONSCIOUS 257
observers who have the ability and the training to introspect
very accurately. (2) That the image method has a much wider
applicability, as it can be employed with any one who has
visual and other images, while the automatic writing method,
as is shown by these experiments and by others, is very lim-
ited in its application. In these experiments only 2 out of the
19 persons were really able to respond to the task set. (3)
The image method gives more information in a given period
of time and thereby decreases the difficulty of the introspec-
tion. (4) In the image method the experience is brought above
the threshold and the observer is encouraged to give his full
attention to what occurs, and he may be directed to observe
particular things. (5) In the image method it is not necessary
to direct the movement connected with the giving .of the in-
formation into an entirely new channel by substituting the
action of lower nerve centers (centers connected with sub-
conscious thinking) for the higher (centers connected with
conscious thinking) which usually largely direct it. (6) In a
confirmatory way the writing method may be made very
useful. The great richness, for example, of what is under the
threshold of consciousness in case of M. and O. is shown by
both methods. (7) Each method also brings things to the
attention not brought out by the other method. The
tendency of the writing movements to be at the disposal of
what is in consciousness is, for example, very noticeable in case
of some observers. In case of M. and O., what is below the
threshold evidently plays also a role as regards the writing.
AUTOMATIC SPEAKING METHOD VERSUS THE IMAGE METHOD
Of some special cases of automatic speaking I have given
illustrations in my study entitled 'Die Projektionsmethode'
(p. 5, 105). From what is heard by the patient himself or
by the experimenter, an idea can be obtained of course of what
is going on under the threshold of consciousness. The words
occasionally unconsciously spoken by a normal person give
one a similar idea. It will be at once evident, however, with-
out any comparative experiments that the image method has
a very much broader field of usefulness because of the dim-
258 LILL1EN J. MARTIN
culty of getting an adequate distraction in using the automatic
speaking method.
THE IMAGE METHOD VERSUS THE PATHOLOGICAL AND THE
PSYCHOANALYTICAL METHODS OF INVESTIGATING THE
SUBCONSCIOUS
The other methods of investigating the subconscious I find
less satisfactory than the automatic writing and speaking
methods. The objection to the pathological method, where
the data regarding the subconscious is obtained for example
from cases of double personality, is the feeling of doubt and
even mistrust with which one often collects and examines such
data.
The objection to the method of psychoanalysis is that the
instruction given to the patient to speak out everything that
comes into his mind, gives a mass of data which contains not
only what is below but what is above the threshold and farther
that in applying this method no systematic effort is made, as
in the case of the image method, to separate out and classify
such data.
Taken all in all, it seems to me, the results show that the
image method offers a mode of penetrating below the threshold
of consciousness which is at least comparable if not superior
to that offered by other methods.
EMOTIONAL POETRY AND THE PREFERENCE
JUDGMENT
BY JUNE E. DOWNEY
The University of Wyoming
In a former study1 the writer reported somewhat extensive
experiments upon the imaginal reaction to poetry and the
influence of the various forms of the image upon the affective
and the aesthetic judgment. The poetic fragments utilized
in this experiment were selected largely because of their
imaginal suggestiveness. Occasional comments of reagents
upon certain fragments indicated that had highly emotional
poetry been employed instead of imaginal poetry different
results might have been obtained.
Accordingly, a second series of experiments was planned
in order to test the emotional factor in poetry. Twenty-
four fragments of poetry, somewhat longer than those of the
preceding test, chosen because of their emotional content,2
were utilized. The judgments obtained are not, however,
directly comparable with those given in the preceding series
since instead of a grouping of the emotional fragments on
the basis of pleasantness-unpleasantness, a grouping of the
fragments into eight groups according to preference was
asked for. In group I were to be placed, according to type-
written instructions placed before every reagent, the frag-
ments liked best; in group 8, those liked least; the other
fragments in the intermediate groups. After this grouping
the reagents were instructed to shade the fragments in each
group, placing first in each group the fragment most liked
and shading from that to the one liked least. Such an
1 "The Imaginal Reaction to Poetry," Univ. of Wyom., Department Psychol.,
Bulletin No. 2.
2 That the fragments utilized were actually less imaginal in content than those
employed in the previous test is shown by the fact that in proportion to the number of
fragments and for the same number of reagents they aroused only half as many images.
259
260 JUNE E. DOWNEY
arrangement was repeated five times at week-intervals. Each
reagent was instructed to record his mood before beginning
his grouping and to record it again at the close of the experi-
ment. After the first and the fifth preference arrangement
the reagent was instructed to rearrange the fragments, on
the basis of the vividness of his emotional reaction to them,
in four groups : III. Reaction vivid; II. Reaction moderately
vivid; I. Reaction slight; O. No emotional reaction. After
the number of each fragment on the second record the reagent
was instructed to write a word or phrase descriptive of the
emotional content of the fragment. In connection with the
second preference arrangement, a rearrangement into four
groups as before on the vividness with which the reagent
projected himself into the content was asked for, with a
complete account of the kind of self-projection observed in
any case. With the third preference arrangement a grouping
of the fragments with reference to the nature of the inner
speech was asked for, together with comments upon the
form of the inner speech for each fragment. A fourfold
grouping on the basis of the vividness of the concrete imagery
aroused by reading was requested in connection with the
fourth preference arrangement.
Some four weeks after the last preference arrangement a
grouping of the fragments into eight groups according to
their beauty was obtained. In group I were placed the most
beautiful fragments; in group 8 the least beautiful; in the
intermediate groups the other fragments. As before, the
fragments were shaded within the groups. In connection
with this grouping answers to the following questions were
obtained:
1. What do you mean by ' beautiful'? Answer on the
basis of your experience while arranging the fragments.
2. In your opinion is an arrangement on the basis of
beauty equivalent to an arrangement on the basis of prefer-
ence? Why?
3. Is an arrangement on the basis of beauty equivalent to
one on the basis of pleasantness?
4. Would an arrangement on the basis of pleasantness be
equivalent to one on the basis of preference?
POETRY AND THE PREFERENCE JUDGMENT 261
5. What kind of emotional appeal do you prefer in poetry?
Can you give any reason for your preference?
6. What kind of emotional appeal do you consider most
beautiful? Why?
Seven reagents took part in the experiment; all had had
considerable practise in introspective work.
A brief description of the poetic fragments employed
seems necessary. The descriptive terms used are taken from
those given by the reagents in connection with the two
groupings of the fragments made by them on the basis of
their emotional vividness. An estimate of the emotional
value of each fragment was obtained by adding the numbers
of the groups in which a particular fragment was placed by
each of the seven reagents for each of the two groupings.
The greatest sum obtainable was 42; the least, o. The sum
actually received is given for each fragment in parenthesis
after the descriptive summary.
The fragments were as follows: i, twelve lines, from
Browning's 'Saul,' expressive of the joys of living, beginning,
'Oh, our manhood's prime vigor!' (28); 2, nine lines, verse
XCIL, Canto Third, Byron's 'Childe Harold,' descriptive
of the exultation and awe aroused by a mountain storm
(33)j 3> nine lines, verse XXI. of Shelley's 'Adonais,' expres-
sive of inevitability, futility, grief (20); 4, twelve lines, a
lyric expressive of companionship (31); 5, sixteen lines, the
eleventh verse of Swinburne's 'The Garden of Proserpine,'
with an emotional toning of desire for death, annihilation
(17); 6, ten lines from Shelley's 'Prometheus Unbound,'
expressive of defiance, beginning 'Fiend I defy thee' (29);
7, fifteen lines from Browning's 'Andrea Del Sarto' beginning
'A common greyness silvers everything,' lines that voice a
twilight mood of sadness and resignation (28); 8, ten lines, a
translation by Symons of one of Mallarme's exquisite word-
pictures, expressing vague aspiration, calm (13); 9, nine lines,
fifth stanza of Swinburne's 'A Forsaken Garden,' expressive
of barrenness, weariness (19); 10, eight lines, Stevenson's
'Under the wide and starry sky,' completion (23); II, twelve
lines, fifth stanza of Browning's 'Love Among the Ruins,'
262 JUNE E. DOWNEY
descriptive of expectancy and love (25); 12, fifteen lines, from
Tennyson's 'The Princess,' 'Tears idle tears' (27); 13, eight
lines, Galsworthy's gay wind-song, 'Wind, wind — heather
gypsy' (25); 14, eight lines, Blake's 'When the voices of
children are heard on the green,' expressive of quiet happiness
(17); 15, twelve lines, Henley's famous 'Captain of my soul'
verses (34); 16, thirteen lines, first two stanzas of Poe's 'To
One in Paradise,' voicing despair (27); 17, sixteen lines, a
descriptive piece by Galsworthy, 'We'll hear the uncom-
panioned murmur of the swell,' expressive of 'God's own
quietude of things' (24); 18, seven lines, Yeats' 'Be you still,
be you still, trembling heart,' voicing mystical courage (14);
19, eleven lines from Tennyson's 'Lotus-Eaters,' beginning
'There is sweet music here that softer falls,' word-pictures
suggesting peace (31); 20, twelve lines, Marston's 'All my
roses are dead in my Garden,' expressing despoilment, hope-
lessness (27); 21, fifteen lines, the hunger for pursuit (16);
22, twelve lines, Yeats' mystical 'Outworn heart, in a time
outworn' (14); 23, nine lines, the first two and the last
stanza of Moody's 'Heart's Wild-Flower' (33); 24, fourteen
lines, Moody's 'Grey drizzling mists the moorlands drape'
(23). These fragments were typewritten on separate sheets
of paper, convenient for handling. The poet's name did
not appear on the fragment and only a few cases of recognition
occurred. The fragments, except 4 and 21, were of accepted
literary excellence, many of them being classic productions.
On the basis of the data gathered the following points
may be discussed: I. The variability and character of the
group preference judgment and its dependence upon such
factors as the emotional content, self-projection, concrete
imagery, the waxing and waning value of the separate frag-
ments; II. The variability and character of the individual
preference judgment and its dependence upon peculiarities
in the individual reactions; III. The relation of the preference
judgment to the judgment of beauty.
POETRY AND THE PREFERENCE JUDGMENT 263
I. THE GROUP PREFERENCE JUDGMENT
The average position of each fragment for each of the five
arrangements by the seven reagents was calculated with the
average M.V. for each arrangement. There is a decrease in
the average M.V. from the first to the fifth trial, although
not a constant decrease, as follows: first trial, 4.839; second
trial, 4.535; third trial, 4.783; fourth trial, 4.629; fifth trial,
4.166. The average M.V. for the first arrangement is some-
what high, although not higher than that given in certain
other reports on the subjective judgment. It is, relatively
to the number of possible positions, higher than the average
M.V. in a first arrangement of imaginal poetry on the basis
of pleasantness-unpleasantness. It is tempting to attribute
this increased M.V. to the emotional nature of the poetry and
very probably it should be so attributed. But it should
not be forgotten that increased subjectivity is not the only
possible cause of increased variability.
With repetition of the arrangements there is lowered
variability. There was at first, as has been pointed out by
other investigators of the subjective judgment, a greater
agreement on the unpreferred fragments with a shift in the
last three trials to greater agreement on the preferred frag-
ments. The average M.V. of the fragments in the first six
positions for the five different arrangements is as follows:
first trial, 4.01; second trial, 4.62; third, 4.04; fourth, 4.22;
fifth, 3.27. The greatest agreement is seen to occur on the
fifth trial. The average M.V. for the fragments in the last
six positions should also be noticed: first trial, 3.77; second,
3.94; third, 4.53; fourth, 4.52; fifth, 4.41. The difference
between the M.V.'s for the first and the last six positions is
greater for the fifth than for any other trial.
The increasing agreement of the group with repetition
of the test is shown by the extent to which every arrangement
is correlated with every other arrangement as given in Table I.
The progressive increase in coefficient values for successive
arrangements is evident, reaching a final value of .89 for the
last two trials.
264
JUNE E. DOWNEY
TABLE I
CORRELATIONS BETWEEN GROUP-ARRANGEMENTS.
Trial
I
II
III
IV
V
I..
.758
.764
•739
•773
II
.758
•75°
.699
.820
Ill
IV
.764
•739
•75°
.699
.820
.820
.863
.890
v
•773
.820
.863
.890
Av
.7158
.71:7
.709
.787
8^7
Study of the records suggests no explanation for this other
than growing objectivity of judgment with increased famili-
arity with material. With such familiarity the individual
judgment would seem to be steadied by social standards.
TABLE II
EFFECT UPON PREFERENCE OF VIVIDNESS OF EMOTION, SELF-PROJECTION, AND IMAGERY
(COMBINED RECORDS. 7 REAGENTS).
i
2
3
4
5
6
7
8
Totals
Ill
Emotion — I
Self-projection. . . .
Imagery
Emotion — 2
Total
16
25
19
13
9
9
8
7
10
6
II
7
5
3
3
i
4
2
5
5
2
2
0
2
5
i
2
I
2
4
5
7
53
52
53
43
201
II
Emotion — I
9
12
8
7
c
6
-I
•i
r-?
Self-projection. . . .
Imagery
5
3
6
c
4
A
5
7
3
c
7
•2
4
1
37
•27
Emotion — 2
IO
7
7
I
r
•2
4.
46
Total
173
I
Emotion — I
3
2
7
2
7
7
4
36
Self-projection. . . .
Imagery
Emotion — 2
Total
o
3
2
4
4
10
4
3
5
3
2
6
7
2
5
3
4
5
8
9
6
5
5
30
3i
47
\AA
o
Emotion — I
Self-projection. . . .
Imagery
0
3
•3
i
7
IO
i
6
8
2
6
7
4
c
i
7
3
7
9
9
2
24
49
4.7
Emotion — 2 . . .
2
-j
4.
^
4
c
4'
•32
Total
152
An attempt was made to determine the influence of
various factors upon the preference judgment by obtaining
as described above a four-fold grouping, twice for vividness of
POETRY AND THE PREFERENCE JUDGMENT
265
emotional toning and once each for vividness of concrete
imagery and of self-projection,1 and distributing these judg-
ments under the eight preference groups (Table II.).
From this table it is evident that the three factors are
about equally potent but that, in general, rich content,
emotional, imaginal, and self-projective, contributed to pref-
erence. The figures for the second grouping of the fragments
on the emotional basis (fifth preference arrangement) indicate
some loss of emotional vividness with repetition.
TABLE III
Five Preference-
Arrangement for
6/0
Arrangements
Beauty
c
•5,
^
«
£f
c
o
g
.
o
£
1
o
'i
Emotional Tone
. E
0
'55
>
>
|
>*
.
W
C/3
E
E
W
*
(S
ai
PH
19
I
5.20
3.20
I
1.85
.98
17
15
2O
H
Peace.
23
2
6.08
2.38
3
5-42
2.03
17
15
IS
16
Heart's wild flower.
2
3
7-45
2.86
5
6-57
4.65
18
13
19
IS
Exultation.
17
4
8.05
2.56
2
4-57
2.94
13
15
IS
ii
Quietude, companionship.
4
5
8-74
3-44
10
11.28
S.3I
18
9
10
13
"You."
ii
6
9.42
2-54
6
6.7I
3-10
10
12
13
IS
"Love among the Ruins."
7
7
9.60
5-20
4
6.57
249
12
IS
16
16
Twilight: regret.
IS
8
9.82
3.00
12
14.00
3-71
19
9
6
15
Fortitude.
i
9
9.85
4.68
7
742
I-SI
16
12
12
12
Joy of living.
24
10
12.50
4.67
14-57
4.04
13
9
IS
10
Weariness; greyness.
12
ii
12.85
2.96
8
9.71
3-75
15
ii
7
12
"Days that are no more."
22
12
12.91
3-iS
II
11.71
2.32
10
6
9
10
Mystical rebirth.
IO
13
1342
4.66
17
16.42
5-35
13
12
13
10
"Glad did I live, gladly die."
16
13-51
2-35
13
1442
3.63
14
IO
6
14
Despair.
20
IS
14.48
4-55
18
16.71
6.33
16
12
13
II
Hopelessness.
8
16
15-17
1.62
9
10.42
S46
6
6
ii
7
Aspiration.
18
17
15.40
3.20
IS
15.57
4.20
7
5
3
7 IMystical courage.
S
18
15.42
2.63
21
17-57
346
7
7
4
10
Eternal sleep.
13
19
15.54
6.00
22
18.71
3-55
12
12
12
13
Irresponsibility.
14
20
15.63
3.07
23
18.57
2.08
8
13
9
9
Content.
3
21
15.80
3-54
19
16.85
445
12
6
4
12
Futility; grief
9
22
16.48
16
16.14
4.12
12
ii
9
7
Barrenness.
21
23
17.82
342
2O
17.14
5-22
9
9
12
7
Pursuit.
6
24
18.77
3-73
24
21.00
2.OO
13
16
II
16
Defiance.
Table III. gives the position, average and M.V. for every
fragment, for five preference arrangements, and for the one
arrangement on the basis of beauty, together with numbers
representing the vividness of the emotional, self-projective,
and imaginal reactions, obtained by adding together the
1 By self-projection is meant an explicit self-reference in whatever form. Cf.
"Literary-Self Projection, PSYCHOL. REV., 19, 299-311 (1912).
266
JUNE E. DOWNEY
number of the groups in which the fragment was placed by
the seven reagents. An attempt is also made to describe
the emotional tone of each fragment. This table confirms
the conclusion that rich content contributes to preference
but suggests also that imaginal content is slightly more
potent in determining preference than are the other factors
studied. This is shown also by grouping together the six
fragments that the records show to be most emotional,
most conducive to self-projection, and most imaginal with
an indication of the position of each fragment in the preference
series (Table IV.).
TABLE IV
Most
Emotional i
Position
Most
Emotional 2
Position
Most Self-
Projective
Position
Most
Imaginal
Position
(1
8
3
5
u
24
7
2
6
/.?
24
7
4
19
2
7
I
3
7
U?
I
2
LI
i
IS
2
I
(23
1 17
2
4
u
9
15
III
6
{'i
2O
3
U4
10
The two fragments most definitely imaginal (19 and 2) rank
respectively first and third. Many fragments occur in two
or more of the groups.
Putting the matter in another way we see that of the six
fragments most preferred 19 and 23 are imaginal, favor self-
projection, and are emotionally toned; 2 is imaginal and
emotional; 17 is imaginal, and induces self-projection; 4 is
emotional; II is emotional, imaginal, and favors self-pro-
jection.
A comparison of the orders received by the different frag-
ments for the successive arrangements indicates that frag-
ments 17, n, 7, 24, 10, 8, and 22 (slightly) waxed in value;
fragments I, 16, 20, 13, 14, and 4 (slightly) waned in value;
fragments 19, 23, 3, 9, 21, 6, 2, 12, 15 remained relatively
static; fragment 18 waxed in value and then fell; 5 waned
and then waxed in value. Reference to the waxing and
waning value of the fragments will occur later in discussion
of the arrangement on the basis of beauty.
POETRY AND THE PREFERENCE JUDGMENT 267
II. THE INDIVIDUAL PREFERENCE JUDGMENT
The effect of the individual reactions upon the preference
judgment was evident and makes necessary a summary
statement of certain characteristics of the reagents. There
were seven of these reagents as stated before. The conclu-
sions, relatively to their general reactions, are based upon
extensive acquaintance with the observers in psychological
experimentation.
With reference to imaginal tendencies the observers fell
into three groups.
The first group includes reagents Rgr and Ele, subjects
in whom there is a strong preponderance of visual imagery.
Rgr's visual images are vivid and detailed. Ele shows a
strong inclination to emphasize form; she is accustomed to
changing all sounds into visual forms. Voices she pictures
in series of waves and lines at different levels; she compares
different pitches by reference to the heights at which the
translating lines are placed. She has a great liking for
mathematics.
The second group includes Jan, Hne, and Jdo. These
subjects show mixed imagery. Although they employ visual
imagery to some extent, they appear to be much more de-
pendent upon kinsesthetic and organic material. Auditory
content is, however, very potent for Hne.
The third group includes Ado and Tbu, who are strikingly
deficient in visual imagery, an incapacity which in Tbu's case
is evidently conditioned by very poor eyesight. Tbu relies
almost wholly upon inner speech and is strongly inclined to
accept the imageless thought proposition. Ado makes much
use of kinsesthetic material and in this respect might more
properly be classed with the second group.
The form which self-projection assumed was also some-
what characteristic within the same groups.
For Rgr and Ele such self-reference appeared to be highly
objective. Rgr projects herself visually within the scene
but without dramatic or kinaesthetic participation in the
scene. Ele is "there" as a spectator only. She assumes,
without visualization of self, a definite orientation toward the
268 JUNE E. DOWNEY
scene, always on the outskirts, where she is able to get a
good view of the situation.
For Hne and Ado, the self-reference is highly colored.
Hne gives a visual self-projection that is fused with kinsesthe-
tic and organic material; she is within the scene. Ado is also
within the scene, part of it, but without visualization of self.
Her participation is definitely dramatic, emotional. These
two reagents "subjectify" the poetic material.
Jan and Jdo identify themselves kinsesthetically or organ-
ically with persons or inanimate objects described. Some-
times for Jdo there is a projection of kinaesthesis into a
visualized figure, not of self. As distinguished from Hne
and Ado, these reagents appear to project or objectify the
subjective reaction.
Tbu reports little self-reference except that in inner speech
he is at once speaker and listener.
A grouping on the basis of the inner speech effects some
changes in the distribution of subjects. This inner speech
is auditory for Hne, Tbu, Jdo, Ado, and Rgr. But of these
reagents Hne is the only one who heard, to any extent,
fragments read in voices other than her own. Tbu, Ado, and
Jdo make much of inner elocution, and Tbu is almost wholly
preoccupied with this aspect of the reaction.
Ele and Jan were sceptical as to auditory content for their
inner speech. Ele reported again curious translations of the
inner speech into visual forms.
The average (with M.V.) was calculated for each frag-
ment for the five arrangements by each reagent and the
position assigned each fragment on the basis of this average.
The average M.V. for each reagent from the average of his
five arrangements was calculated and gives us an indication
of his individual variability. His average M.V. from the
average group judgment was also determined and this indi-
cates the extent to which his judgment was representative
of the group.
The variability of each reagent from his own average for
the five trials was as follows in the order of least variability:
(i) Tbu, 2.21 ; (2) Rgr, 2.28; (3) Ado, 2.71; (4) Ele, 3.11;
POETRY AND THE PREFERENCE JUDGMENT
269
(5) Jdo, 3.71 ; (6) Jan, 3.56; (7) Hne, 4.09. Increased varia-
bility seems, in general, to characterize the more emotional
reagents (determined by their fourfold grouping of the frag-
ments), while the effect of the emotional material in increasing
the variability is shown by comparison of the individual
variability in this test with that found when less emotional
poetry was utilized; it is proportionately much higher in the
present test.1
The average variability of each reagent from the average
of the seven reagents for the five arrangements gave the
following order: (i) Jdo, 2.75; (2) Ado, 2.90; (3) Rgr, 2.96;
Ele, 3.07; (5) Jan, ^3.48; (6) Hne, 3.78; (7) Tbu,5.i9. The
most interesting point in this listing of reagents is Tbu's shift
in position, which with high personal consistency indicates
a different basis of judgment from that of the other reagents,
explanation for which is to be found in his introspective
reports. On the whole, it may be noted, the variability
from the group average is no more extensive than that found
in imaginal poetry.
TABLE V
PREFERRED FRAGMENTS
Rgr
Ele
Hne
Jan
Ado
Jdo
Tbu
I
19
17
II
4
19
19
15
2
2
23
4
13
17
23
I
3
7
19
12
23
7
7
2
4
23
2
19
19
23
17
10
5
I
7
I
18
2
10
24
6
4
24
13
17
II
4
5
An arrangement in order of the six fragments which were
most preferred by each of the reagents shows at once the
effect of the individual differences in reaction (Table V.).
We note that of Rgr's preferred fragments, the first four
are exactly in the order of imaginal vividness, largely visual.
Rgr states very definitely that she prefers poetry which calls
up vivid visual images, unless, as in fragments 13, such images
are grotesque. Ele's six preferred fragments are just the six
1 A great variability from her own average in the judgments on emotional poetry
in contrast to great self-consistency in judgments on imaginal poetry was shown very
definitely by the one reagent (Jdo) who participated in both tests.
zyo
JUNE E. DOWNEY
most imaginal fragments, although not in the exact order of
the group. Ele also expresses a preference for poetry con-
veying the clearest imagery and is particularly pleased with
what she calls sound-pictures.
Hne's preferred fragments are chiefly emotional in tone;
1 1 and 4, both highly emotional, represent her first and second
choice. Jan's preferred fragment is 4, which is emotional in
its appeal, but his other choices give some indication of
dependence upon imaginal richness. He also prefers 18, a
mystic fragment of little sensuous content. Ado's and Jdo's
preferred fragments show the influence of imaginal content
as well as of emotional toning.
Tbu's preferences are distinctly individual, determined
largely by the kind of emotion expressed which Tbu prefers
to be strong in nature, expressive of a desire to act, to conquer.
Neither 19 nor 23, so generally preferred by other reagents,
occur among his first six fragments; the absence of imaginal
fragment is very evident.
TABLE VI
CORRELATIONS OF Av. REFERENCE ARRANGEMENT OF EACH REAGENT WITH THAT OF
EVERY OTHER REAGENT
Reagent
Rgr
Ele
Jdo
Jan
Hne
Ado
Tbu
Rgr
.601
•590
.46l
.232
.700
•237
Ele
.601
410
•325
.209
.702
.103
Jdo
Jan
.590
.461
410
.325
.536
•536
.230
.256
.569
.241
.200
-.272
Hne
.232
.209
.230
.256
.38l
-•059
Ado
.700
.702
.569
.241
.381
•193
Tbu
•237
.103
.200
-.272
-•059
•193
The tabulation of preferences suggested the working out
of the coefficients of correlation for the average preference
arrangement of each reagent with every other. These are
given in Table VI. It is evident from this table that re-
actions on the basis of imaginal qualities are most representa-
tive (Rgr, Ado, Jdo, Ele) and that the subject most visual in
reaction (Rgr) gives the highest average correlation. Should
such a conclusion be substantiated by a more extensive
investigation it would seem to throw light upon the kind of
POETRY AND THE PREFERENCE JUDGMENT 271
literary material that would probably have constant value
for a long period of time and the type of critic that would
best represent the average reaction in the long run.
Certain other factors influencing the individual reactions
are evident from the tables and the introspective reports.
Table III. indicates that subdued emotions are more generally
preferred by this group than are violent emotions. The
individual reports confirm this, although the effect of the
mood of the day is mentioned by several reagents as influenc-
ing their preferences.
Rgr. "In poetry the emotional appeal which I prefer
depends largely on my mood. — I like poems about nature
as they arouse emotions outside of one's self."
Ele. "An appeal to quiet, drowsy, lesiurely, reminiscen-
tial feelings suits me best. — I do not like noise, boisterousness,
confusion."
Hne. "The appeal preferred depends upon the mood.
Usually prefer something expressing longing unfulfilled, or
the joy of living."
Jan. "I do not know that I can select any one emotional
appeal; sometimes it's one sort, sometimes another. Pathos
perhaps makes the greatest appeal."
Jdo. "I prefer the emotional tone to be in harmony
with my mood which varies strongly from day to day. In
general I prefer a sad toning."
Ado. "I like an emotional appeal that is melancholy in
tone."
Tbu. "I prefer a strong emotional appeal to any of the
pleasant emotions and sometimes to those generally con-
sidered unpleasant. Usually, however, I prefer such an
appeal to emotions as are aroused by fragments I and 15,
feelings of desire to act, conquer, oppose even unconquerable
forces. The reason for this preference so far as I can judge
is that such emotions are not common in me. However
much I may consider them ideal, I do not possess them. It is
their contrary nature that appeals to me."
In order to test specifically the effect of the mood of the
day upon the reaction to strongly emotional poetry, the
272 JUNE E. DOWNEY
following tabulation was made (Table VII.). The seven
fragments most melancholy in tone were selected, 3, 5, 9,
12, 16, 20, 24; three of happy buoyant coloring were chosen,
13, 14, i; and three of strong aggressive emotion, 2, 6, 15.
Next, record was made from the introspective notes of any
cases where the reagent reported strongly depressive moods
at the time of the experiment. Eight cases of this occurred.
The effect of the mood-dominance was then determined by
subtracting the position on the day in question for the given
fragment and the given reagent from the average for that
reagent's five arrangements. A minus sign indicates in-
creased preference for the fragment for the given day; a plus
sign indicates decreased preference.
The table would seem to suggest some interesting dif-
ferences between the reagents as to the effect of mood upon
their preferences. Jan and Jdo show very evidently that a
mood of depression increased for them the preference for
melancholy poetry and in Jdo's case very considerably
lowered the liking for buoyant fragments. The effect of the
mood upon fragments 2, 6, 15 is less constant. Ado's record
indicates a general lowering of values under depression, with,
In a few cases, added appreciation of the melancholy frag-
ments. Rgr shows less effect of mood upon preference than
any other reagent, and that effect is mainly a lowering of
values. The effect of depression is somewhat variable for
Ele and Hne, both records suggest that harmony with the
mood is likely to increase preference. Under the influence of
the given mood these reagents show inclination to stress 2
and 6. The effect of mood upon the emotional reaction
is a very important one. The above discussion, however
meager, suggests a method by which the problem may be
attacked.
Jdo's comments on the effect of mood upon the reaction
for any particular day are more complete than those of the
other reagents and in certain respects instructive. There
were days of aesthetic toning and other times when it required
considerable effort to surrender to poetic suggestion. These
differences were due to general mental conditions, rather
POETRY AND THE PREFERENCE JUDGMENT 273
than to experimental conditions. At the close of the experi-
ment Jdo was usually in a more aesthetic mood than at its
beginning. The most adverse general criticism upon experi-
mental investigations of this sort she finds in the reduction
of the time needed for aesthetic absorption. Short fragments
suffer in comparison with the longer productions from which
they are taken. Fragment 12, for instance, frequently failed
4 to catch fire.' Rereading was necessary. Again, Jdo noted
that the first fragments suffered by being read before she
had assumed a poetic mood; or, at times of increased suscepti-
bility to outer suggestion, fragments I and 2 set the tone for
subsequent reactions.
The second arrangement was made under a mood of great
depression, heightened by the 'grey toning' of the weather.
Jdo recorded in her notes that fragments expressive of sadness
and futility were given a higher value than before. She was
aware also of a tendency to react against the mood by assign-
ing high value to fragments expressive of fortitude. At the
third arrangement a mood of aesthetic sadness again enhanced
the value of fragments of melancholy tone. On this occasion
the lilt and swift mocking rhythm of 9, 14, and 13 were found
very distressing. She reports, "They move too rapidly and
lightly to fit in with the tempo of my mood." At the close
of the test the pulse-rapidity was found to be 78. The
fourth arrangement was made when the subject was in a
scientific mood that contrasted strongly with the aesthetic
mood of the previous week; there was restlessness present
and distaste for taking time for the experiment. The
melancholy-toned fragments were conspicuously less pleasing
than before. Fragments 13 and 14 now fitted into the rhythm
of the day and were shifted from the eighth to the second
group. The pulse was 90. The fifth arrangement was made
under the influence of a "hurry-mood" (pulse 94); Fragment
13 was felt to express exactly the personal tempo for the day.
Besides investigating the influence of the imaginal and
emotional reactions upon preference, the experimenter made
an attempt to determine the effect upon preference of the
different forms of the inner speech. In particular, an effort
274
JUNE E. DOWNEY
was made to determine whether certain fragments encouraged
an auditory inner speech and others stressed the kinaesthetic
quality of inner speech. To this end the reagents were
requested, in connection with the third preference arrange-
ment, to place, if possible, the fragments in two groups, one
group being for fragments in which the auditory aspect was
the more pronounced in the inner speech, the other for frag-
ments that stressed the kinaesthetic factor. The attempt
TABLE VII
EFFECT OF MOOD ON PREFERENCE
Mood Reagent
Arrangement
ISt
Dis-
gusted
Jan
ISt
De-
pressed
Ado
3d
"Blue"
Ado
2d
Sad
Jdo
Dl
pressed
Jdo
D3i
satisfied
Rgr
4th
Im-
patient
Ele .
5th
Dis-
heart-
ened
Hne
No. of Frag.
3. Futility Grief.
— 6.2
—10.8
-3-8
+ 1.6
— 3-2
— 8.8
— 1.2
+6.8
±1.6
+tt
— O.2
+ 2.2
—9.2
—6.0
+ 1-4
+ 1.8
+2.2
—6.2
+5-2
-4-6
— 6.2
-7.8
— I.O
+ 2.4
+ 1.4
—i i. 6
-3-6
— 3-2
+ 2.2
~5'?
— 3-6
-3-6
— 0.6
—4.2
o.o
—44
+3-0
+ 1.4
+ 2.0
+34
+ 1.2
— 3-2
— i.o
-9-6
+ 0.6
+ 4-2
— 1.6
+ 54
— 2.8
— 4.4
— 3-o
+ 1.8
+ 0.8
+ 1.8
5. Eternal Sleep
o Barrenness
12. "Days that are no
more"
16. Despair
20. Hopelessness
24. Weariness
13. Gavetv .
— 3-2
— 4.2
+ 8.2
+0.8
—0.8
+1.8
— 1.2
+0.2
—0.2
+ 2.6
+ 3-0
+ 5-6
+ 11.6
+ 9-o
+ 6.6
—0.4
+2.8
0.0
+ 3-2
— 2.8
— 3.2
+ 6.0
+ 0.8
+ 12.2
14. Happiness
i. Joy of living .
2. Exultation
— 0.6
+ 2.6
-5.8
+ 1.6
+ 1.6
+6.4
+4-2
+ 1.6
-4.6
+ 8.6
— 2.8
— 8.0
T 2
o.o
— O.2
—0.6
— 3-2
—15.0
+ 0.2
—10.4
— 8.0
+ 2.0
6. Defiance
15. Fortitude
proved abortive, although the comments upon the variations
of the inner speech were instructive. Only two reagents
(Tbu and Jdo) were willing to attempt the grouping suggested.
Tbu's choice of fragments in which the auditory side was
most stressed was as follows: 5, 8, 9, 10, 12, 13, 14, 16, 19,
20, 21. Jdo's selection was as follows: 4, 10, 12, 13, 14, 19.
There are five fragments selected by both of these reagents.
Fragment 21 was written in dialogue form, the two
speakers being a mother and son. Nearly all the reagents
note a difference in inner speech with the transition from
one part to another. Hne hears her own voice for the
POETRY AND THE PREFERENCE JUDGMENT 275
mother's part, and a man's deep voice for the son's. She
finds the fragment pleasing. Tbu, on the other hand, who
hears the man's voice in his own, and the mother's voice in a
squeaky disagreeable voice, does not like the fragment.
Ele reported that she seemed to see the inner speech.
"Inflections are represented by different levels upon a scale."
For Jdo, there was a prominence of the visual verbal side
(with subordination of the inner speech) for fragments 4, 9,
13, and 17. These two subjects are the only ones referring
to visual factors in connection with the inner speech.
III. THE PREFERENCE JUDGMENT AND THE JUDGMENT OF
BEAUTY
The introspective reports of the reagents, except those of
Tbu and Hne, do not indicate much difference in their basis
of judgment when they shift from the category of preference
to that of beauty. Ele and Rgr assert that there is no differ-
ence introspectively between the two forms of judgment.
Rgr adds, "I would not prefer those fragments which did not
appeal to me as beautiful." Jdo makes the preferred frag-
ments largely correspondent with the beautiful ones although
she reports that there may be poetry which is merely pleasing
and preferred for that reason. Ado finds little difference
between the two categories, except that in a preference
arrangement she gives more attention to the thought and
in the beauty arrangement more attention to style and form
of expression. Hne reports considerable difference between
the two kinds of arrangement. She also emphasizes the
expression-side as important for the judgment of beauty and
the thought-side as important for preference. Tbu makes
the following distinction, "By beautiful I mean a fragment
that is pleasing in rhythm, rhyme, and meaning. Preference
is based on the emotional reaction."
Some differences are given as to the emotional appeal that
seems most beautiful in comparison with that which is
preferred. Ele, Rgr, and Ado find no difference Ele states
that an appeal to quiet drowsy emotions is at once most
preferred and most beautiful; Rgr finds an appeal to peaceful
276 JUNE E. DOWNEY
emotions most beautiful; so, too, does Jan. Jdo writes, "I
feel the appeal to renunciation, to world-sadness, to wistful-
ness, to serenity most beautiful." And Tbu, "A lack of
emotional appeal seems most beautiful to me; something
calm, quiet, not at all exciting." Throughout the group
there is evident a tendency to reduce the value of personal
emotion as a factor in the judgment of beauty.1
Correlation of the average arrangement by the seven
reagents on the basis of beauty with the average on the basis
of preference (five trials) is, however, very high, .894. A
correlation of this average judgment for beauty with the
average of the first and the fifth preference arrangement
gives as coefficients .729 and .941 respectively. This last
coefficient may indicate merely that the growing agreement
of the group shown in the correlation of successive arrange-
ments for preference (Table I.) continued in spite of the shift
in category and the lapse of four weeks; or, more probably,
it shows that with familiarity with the fragments preference
is determined more definitely by the element of beauty than
is evident in the earlier trials. A preliminary arrangement
for beauty would have been most valuable in this connection.
The average M.V. for the arrangement on the basis of
beauty is 3.61, a lower M.V. than found for any of the five
preference arrangements. This lowered M.V. suggests a
more objective standard for the judgment of beauty than
for a judgment of preference.2 Tbu, in particular, shifts
his basis of judgment. The correlation coefficient for his
arrangement on the basis of beauty with the average of the
group is high, .783.
In particular, it seems probable that a beautiful fragment
is more constant in its value than a preferred one and falls
less in value with familiarity in the repeated arrangements.
There are several ways in which one may test this assumption.
One may, for instance, inspect the order of fragments in the
first preference arrangement (group average) and note which
fragments show striking discrepancies with the order of
1 This report is quite in accord with Souriau's conclusion. " La Reverie Esthetique,"
p. 104 f.
2 Compare Miiller-Freienfels, R. "Psychologic der Kunst," II., p. 171.
POETRY AND THE PREFERENCE JUDGMENT 277
fragments in the arrangement for beauty (group average)
One might then check this list with the numbers of those
fragments that have already been cited as showing evidence
of a waxing, waning, or static value in the repeated preference
arrangements. Following this out, we find that fragments
4, 13, 14, 15, 1 6 and 20 give a much higher value (five or more
places) in the first preference arrangement than in the arrange-
ment for beauty. With one exception (15) these are just the
fragments which were found to fall in value with the repeated
arrangements. Fragments 7, 8, 9, 12, and 17 show a much
higher value in the arrangement for beauty than in arrange-
ment for preference. Three of these fragments 7, 8, and 17
were fragments that waxed in value with the repeated
arrangements. The outstanding fragments are instructive.
Fragment 15 is very definitely a preferred fragment rather
than a beautiful one; fragments 9 and 12 are judged to be
beautiful but in spite of that are not highly preferred.
Again, taking the average arrangement for beauty and
dividing it into two sections, the twelve fragments least and
the twelve fragments most beautiful, we find that of the
twelve most beautiful fragments, five were those which
waxed in value in the preference arrangements, five were
static in value, two waned in value. Of the twelve less
beautiful fragments, three fell in value in the preference
arrangements, five were static, two waxed, two (18 and 5)
fluctuated.
The conclusion seems justified that the beautiful has a
value which holds its own or waxes with familiarity.1
IV. SUMMARY AND CONCLUSIONS
As a general outcome of the experiment we conclude that
the group reaction to emotional poetry is slightly more sub-
jective than the group reaction to imaginal poetry. Famil-
iarity with the material reduces the group variability. Rich
content, emotional and imaginal, is shown to contribute to
preference, with a slight advantage in favor of imaginal
content as a determining factor. Certain fragments appear
1 Miiller-Freienfels, op. cit., II, p. 168.
278 JUNE E. DOWNEY
to fall in value with repetition of the judgment; others to
increase; others are static.
The effect of individual differences upon preference is so
evident that a grouping of the reagents on the basis of type of
reaction is instructive. The more emotional subjects appear
to be more variable in their judgments. The preferences of
four of the reagents show a great dependence upon imaginal
content; emotional vividness is potent for two subjects; and
kind of emotional content for one. The intercorrelations
indicate that preferences based on imaginal content have a
more representative value than preferences determined by
emotional content. The latter are influenced by the mood
present at the time of choice.
A group arrangement on the basis of beauty correlates
very highly with the average group preference arrangement
and would seem to be, for most of the observers, determined in
much the same way. This second judgment is, however, less
subjective than the preference judgment. Apparently the
more beautiful fragments have a more constant value than
fragments that are merely preferred. Since the former do
not wane in value as do the latter, in time the two categories
closely approximate each other.
In general, the order of merit method, in conjunction with
an analysis of individual reports, appears to afford a most
excellent means of studying the aesthetic reaction.
AN EXPERIMENT IN ASSOCIATION
BY C. G. BRADFORD
East Central State Normal, Ada, Oklahoma
THE method used in this experiment was devised by Dr.
J. E. Lough, of New York University. It is simple and easy
to execute, and at the same time it is very effective. The
experiment is dominantly mental in character.
i. AIM
This experiment was studied in the light of the influence
of (a) practise, and (b) such transient factors as distraction,
fatigue, concentration, state of health, etc.
2. METHOD AND MEANS
The experiment was conducted individually, the writer
being the subject. The means used were the test and key
sheets of the Lough Association Method. Other means were
a common lead No. 2 pencil and paper to write the associated
letters on. Time was kept by an ordinary watch.
The test sheet contained ten rows of letters; each row
contained twenty letters. These letters were in non-alpha-
betic order. The key sheet had two rows of letters on it,
one just above the other. The top row was in alphabetic
order, from A to T, inclusive. The bottom row was placed
in random order and each of its letters was directly under a
letter of the top row. For convenience a test sheet and keys
have been inserted. On this sheet seven keys are given,
the first in full, but as the top row is the same for all keys
only the bottom row of each of the others is given.
3. PROCEDURE
This experiment was taken in three divisions. The morn-
ing test (A) was taken between seven and eight o'clock, with
emphasis on speed, i. e., it was quantitative in nature. The
279
280 C. G. BRADFORD
afternoon test (J5) was taken between one and two o'clock,
and was qualitative in nature, the emphasis being put on
the quality of the work rather than speed. The evening
test (C) was taken for both speed and accuracy and came
between ten and eleven o'clock. The general course of these
three tests, except the aspects already described, was the
same in all.
TEST SHEET
KCENORAFB ILGSMPTDJHQ
MKNGOLCAEBT IFQ JPHRDS
CDATG I SKRN JMQSEHOPLF
QTCPF JNL ICDRG SAEKHBM
TPHM JNL SOFDGEQLKANB I
N IBQEHDTJRFAKSLOGMCP
AQ IKEGT S JDOHCFMBRLPN
SAPJQMGDFTRK INHLEOCB
JE IBDNGSOCMLAQFRPTKH
GNEDKB SRQHPC JLTMFAO I
KEYS USED
i
ABCDEFGH I JKLMNOPQR ST
UYTMB JCZCLYKAEGF IHND
TMH IA SCBXDE JFYWGOKUL
SXYKN JUPTP E3C OQRADEZL
MTJODPVAYC F4B SQEWGNHM
ZUFLYMBWZER5DHAVJNG IO
CPGQKTAXUNYOJFRLW IMD
FGX JRWVLDMB7J J SCY IKUZ
Making the associations of the letters was done in the
following manner: The key, until all associations were
thoroughly committed to memory, was kept just above the
line, on the test sheet, on which the subject was working.
The letters in each line of the test sheet were taken consecu-
tively as they were approached, regardless of their order.
Each one was matched with its likeness in the upper row on
the key, and under it was written the particular letter which
appeared under its likeness in the top row of the key, e. g.,
A was first in the top row on the key: suppose that under A
in the bottom row of the key was Z, then whenever A was
AN EXPERIMENT IN ASSOCIATION 281
found in a line on the test sheet Z was written under it.
For further illustration take an actual case. On the test
sheet K is the first letter in the first line; by running down
the top row of letters on the key (No. i) we find that K has
Y under it; therefore we would write Y under K on the test
sheet. All the other letters are associated in the same way
with the various letters of the alphabet.
These associations were made very slowly at first. The
subject, however, soon learned a few of the associations or
equivalents, and learned how to make short cuts from the
letters on the test sheet to their equivalents in the key,
without going to the first of the key and following it letter
by letter till the right one was reached as the tendency at
first was to do, and thus continually decreased the time for
each test. As these associations were committed to memory,
this entire process of referring to the key for the equivalent
was, of course, syncopated. When the associations became
well fixed in consciousness the sight of a certain letter on the
test sheet brought up an immediate image of its equivalent
in the key and the motor-writing impulse was discharged,
resulting in the hand movement executing the writing of the
equivalent.
In this experiment the completion of a test sheet consisting
of ten lines constituted a test, one of the ten lines constituted
a trial.
Twenty tests were made in each key, except the fifth,
which by oversight was discontinued after nineteen tests.
When one key was finished another was immediately begun.
Seven keys were finished, covering a total of one hundred
and thirty-nine days. On account of illness one test was missed
in both the morning and the evening series. No other break
occurred in the entire experiment.
4. RESULTS
The results of this series of tests are very interesting,
especially from the standpoint of practise, and the effect of
changing keys.
Curves are plotted and tables compiled from the daily
282
C. G. BRADFORD
AN EXPERIMENT IN ASSOCIATION 283
records. These curves and tables show concretely and graph-
ically the learning process or habit formation in this experi-
ment.
The tables show, for each key, the daily record in seconds
for the average of ten trials, or one test, the general average,
the maximum and minimum trials, and the mean variation
for twenty tests.
Curves A, B and C represent the daily series of experi-
ments as heretofore explained. Curve D represents the mean
variations in the various keys and for the three series. This
curve shows that in each succeeding key the mean variation
became smaller and smaller.
The subject found that to try to rush either in getting
adjusted in order to start at a given time, e. g., when the
second hand of the watch reached the thirty seconds mark,
or to finish a trial in a shorter period than the previous one
nearly always resulted in a loss rather than a gain in time.
It seemed that consciousness rebelled against any acts of
coercion. Usually when the best records were made the
process was almost unconscious, i. e., no special effort for
speed was being made.
Some associations in each key were easily made because
of their familiarity. That is, they formed initial letters of
familiar names and for that reason were easily remembered.
After these, the associations of the first and last letters with
their equivalents were the easiest to form. The remaining
associations then were retained by sheer memory.
At times it was very hard to concentrate the attention,
and as a result a very poor record was made. Many times
the subject wasted time in trying to recall a certain equivalent
instead of looking at once at the key.
In some cases the sequence of letters offered resistance to
conscious activity. For instance, in key No. I Y is the
equivalent of both B and K, and it happens that B and K
come together in the last line of the test sheet, thus causing
two Y's to come consecutively in the association. In this
case the subject had a feeling when writing the second Y
that he was either repeating the same association or making
284 C. G. BRADFORD
a false association which caused considerable distraction. In
key No. 7 the letter J comes twice consecutively; its equiva-
lents are L and M; it happens that in the ninth line of the
test sheet the letters M and L come together, thus causing a
repetition in the association as explained above.
The subject had another source of distraction closely
akin to that mentioned above, namely, making a second
association and thus writing the wrong letter. To illustrate
this process let us refer to key No. i; D in the top line has
its equivalent M in the bottom line; while M in the top line
has its equivalent A in the bottom line; when the key became
well memorized there was a tendency, upon seeing D, to run
ahead and, instead of writing its equivalent M, to write M's
equivalent A. This difficulty may have been due to a lack
of concentration of attention.
It was also found that a cold, fatigue or unaccustomed
noise, persons entering the room, knocking on the door, etc.,
caused slow work, or an increase in the time record.
The form of the curves A, B and C is very striking. Each
one of these composite curves contains curves of the seven
keys used. The broken vertical lines between these curves
are merely connections and have nothing to do with the
experiment.
It is very significant that in these curves for the various
keys the greater part of the gain comes during the first day's
practise. For instance, in curve A, key No. I, the total
gain was 32.5 seconds, while the gain of the first day was 13
seconds. In key No. 3 of the same curve the total gain was
28 seconds, while the gain for the first day was 20 seconds,
thus leaving a gain of only 8 seconds during the remaining
nineteen days.
For the most part the key curves in curves B and C make
less gains than those in curve A. This is due, however, not
so much to a failure in reaching low minima as to the fact
that the maxima were lower than those in curve A. This was
expected because of the practise in the same keys in curve A.
While the mental attitude was different and while the aim was
not identical in the three series, nevertheless, the practise
AN EXPERIMENT IN ASSOCIATION 285
had the same effect on consciousness and the habit-forming
processes; and, therefore, the second and third series had
the benefit of the practise done in the first series.
In passing from one key to another, of course there was a
considerable rise in time. It has been contended by some
authors that such rises in changing the reaction to certain
stimuli are caused by interference; others, Bair,1 for instance,
think not, unless it takes longer to do the new test than it
did to do the preceding one.
It seems to the writer that there are two possible explana-
tions: first, if the time taken in doing the new experiment is
less than that for the old one the initial rise is due simply to
the general law of habit formation and not to interference
of associations formed in the previous habit. Because if
such interference does exist, and this added difficulty plus
the normal difficulty in forming any habit be fused together,
it would undoubtedly take a longer time to perform the new
experiment the first time than it did to perform the old one
the first time; and, secondly, it is probable that there is, on
account of similarity and recency when two experiments
are related in character, a certain amount of interference
which would tend to decrease the time. It is also probable
that a certain amount of the learning in the first experiment
is transferred to the new one which would tend to decrease
the time. With one of these forces tending to increase the
time and the other tending to decrease it a resultant is
obtained which ordinarily gives as a consequence a lower
time record than occurred in the former experiment. It
seems to the writer, from his study of general experimental
findings and from a close study of his own experiments in the
learning process, that the latter hypothesis is the more
tenable.
5. CONCLUSION
In this experiment we have the typical learning process.
This process is made graphic by curves which show very
rapid progress at first, but finally the rate becomes very slow.
1 Bair, J. H., "The Habit Curve," PSY. REV., Monog. Suppl., 1903.
286
C. G. BRADFORD
TABLES FOR THE SEVEN KEYS
AVERAGE TIME IN SECONDS FOR EACH TEST OF TEN TRIALS
Key i Key 2
Date
A
B
C
Date
A
B
c
Oct. 22
" 23
45-5
33.1
48.5
3C.2
37-4
26.8
Nov. ii
" 12
38.2
26.7
30.9
3O C
33-8
26 i
" H
jj'j
26.1;
31.3
26.4
" 13
2C A
066
27 C
'" 25..
2J..2
27.4.
24.1
" 4
22 I
26 o
*/•>
22 6
" 26
21.2
24.7
21-4.
" is
2O O
23 8
21 6
" 27..
21.8
24.. c
20. 0
" 16
21.4
24 4
21 6
« 28..
i8.c
22.8
20.8
" 17
21 4
21 4
226
" 29. -
2O. I
23.7
17.0
" 18
17 0
2O 7
2O 2
" so
17.0
21. C
2O.O
" 10
18 4
2O I
2Q I
« 3!
17-7
21.7
10 I
" 20"
17 O
18 c
18 o
Nov. i
17.0
19.1
18.3
" 21
17.8
10.3
17.0
" 2 .
16.6
2O I
17 3
" 22
17 C
17 7
1C Q
" 3
15.8
18.1
17.7
" 23..
> r
i6.c
I7.O
16.2
" 4--
s
16.1
17.7
IC.O
" 24 .
i6.c
17.4
16.2
" c
16.1
18.3
I7.C
" 25 .
I4-. 0
*/•*
i6.c
16.3
« 6..
15.6
I7.O
ic.c
" 26
14.8
•v.3
18.4
1C. I
7
1C. 3
18.1
1C. 2
" 27
14.. I
18.8
*
16.0
8
I4.O
17.6
17.3
" 28
14.6
16 o
2
13.6
0. .
14-3
•/•*j
17.6
/•
14.6
" 20
13.3
IC.O
14.4
« 10
14. 7
' s
17.6
IC.O
" 30"
14. C
1C 4
IC.C
<Gen. Av
20.1
23.1
10.8
10. 1
20 8
IQ 6
Maximum .
CC.O
CC.O
42. 0
AC O
•JO O
4O O
Minimum
12.0
IC.O
13 O
II.O
11 O
12 O
M. V. .
C.i8
C.3O
3 82
4 10
•J •?•?
4 OO
Keys
Key 4
Dec I
4O 2
2C.Q
24.O
Dec. 21
37.0
33.7
2Q.O
" 2
2O I
24 4
20.2
" 22
3/-V
26.2
26.6
22.2
" 7
17 7
23 2
2O.4
" 23 .
2O.Q
23.4
22.6
« i
IQ 2
21 4
10.3
" 24
22.4
21.4
20 6
« t
17 8
***5
23 6
18.1
" 2C
10.4
21.3
^,y.«j
10 8
" 6"
ICQ
18 8
16.0
" 26
21.2
20. 3
2O 4
" 7
16 1
22 7
176
" 27
10.4
10 O
iw.ij.
18.0
" 8
14 6
IQ Q
17 3
" 28
17 4
i8.c
17 I
" 0
ICQ
18 6
17 O
" 20
1C. 2
j.u-3
18 7
i/.t
17 6
" 10
16 i
IQ 2
163
" 30
1C 7
IQ 2
16 c
" ii
***«j
13 Q
17 8
IC.C
" 31
17 C
17 7
17 2
" 12
»J'V
13 Q
17 7
HA
, J1
Jan. I
' ?
ic 6
1C Q
1C 3
" 13
13 O
'/'
16 7
14 C
2
l^j.U
JS 6
IQ 7
14 O
« 11
14 4
166
14 I
" 3
16 o
iy./
16 Q
1C 7
" 1^
I3.O
ic. 8
IC-4
" 4 ; ;
IC-4
18.4
1C- 1
" 16
12.8
*3>w
16.2
7
14.6
" s
IC.O
18.1
14.3
" 17 .
12.4
17.7
IC.O
" 6
I4.C
18.3
17.1
" 18
13. 1
•/•/
i6.c
14.1
" 7
I4.I
16.6
13.2
" I9-.
I3.O
»»».3
16.7
13.4
" 8
12.7
20. c
13.7
« 20.:....
13.2
1C. I
13.7
" 9. .
13.2
•~ 3
16.2
1C. 3
Gen. Av
l6.3
10-2
16.6
18.3
2O.2
17.8
Maximum
C7.O
3O.O
29.0
42.0
42.O
33.O
Minimum
II.O
I3.O
I2.O
II.O
I4.O
I2.O
M. V
3.36
2.7O
2. 1C
4.06
>
2.61
2.86
AN EXPERIMENT IN ASSOCIATION
287
Key -5
Key 6
Date
A
B
C
Date
,4
B
C
Jan 10
1C Q
-3 1.7
1Q I
Tan. 20
Jan. 29.
38.6
36.3
2
" II
24.. Q
23. C
21.1
6 2
28.0
22.9
" 12
21.2
22. C
•> J
21.6
" 3i"
23.2
22.1
v
10.6
" 13..
18.7
2O. Q
18.4
Feb. i
18.4
18.5
18.2
" U-.
18.0
19. 1
20. c,
" 2
16.8
2O.4
18.6
" 1C. .
16.5
18.4
-^•5
16.2
" 3
17.4
19.8
18.3
« 16..
.V.J
16.1
2O.3
17.7
" 4 -
16.6
18.4
I7.r
1 17
16.0
17.7
16.3
" I
14.7
16.6
18.1
' 18
IC..9
17.7
*V.J
16.4
« I.:
14.7
18.2
16.0
' 19. .
13-9
I7.Q
;. T
16.7
7..
14.3
17.0
17.2
' 20
14.. 0
C. "
lo o
JC Q
" 8
16 o
10. 1
ICQ
' 21
1C. I
JC. 7
168
" 9
14 8
16.0
•3***
i?. 8
' 22
13.7
17.2
14. O
" 10 "
*T-U
14 8
16.2
JC.C
' 23
JC. I
16.0
14. 1
" II
14.0
ic.q
jc.3
* 24.
JC. I
16.0
H.I
" 12
13-0
14.8
a3O
13.8
* 2C . .
I4-.7
ic.7
14.7
" 13
13-3
14.. 7
11. C
' 26..
Ii. I
16.2
I4.Q
" 4
12. C
13.8
13.8
' 27
I4.O
14.8
13.6
" il
12.3
14.0
I^.O
' 28
I'?.'?
1C. 7
11.7
" 16
12.3
13.8
I4--4.
' 2Q1. .
" 17-
12.3
XJ.W
ic.6
11.7
Gen. Av
17.1
186
17 3
l63
18 6
163
Maximum
Minimum
M. V.
38.0
II.O
3.C2
40.0
13-0
2.80
35-o
12.0
2.80
43-o
II.O
3 61
40.0
I2.O
3.74
25-0
12.0
2 ic
Key 7
Feb. 1 8
•7C.3
31 4
27 4
Mar 2
146
ic 4.
ic 2
" 10
2C.C
JA-4
23 6
22 I
" 3
14. O
1V4
16 o
14. C
20 .
20. 0
21. 1
2O O
" 4
13 8
146
14. 2
" 21
18.1
17. c
18 c
" 5
13 3
6
13 2
1 22
19. i
*/o
18.1
17. 0
« 1:
13.2
6
1-2 2
' 23. .
18.2
19.3
18.7
" 7
13.2
Hi
14. O
' M
18.1
I8.C
17.2
" 8
12.6
14 2
13 0
« 25
17.2
I8.S
18.4
" 9
13-3
14.7
12-9
4 26
16.8
17.1
17.2
tf\ T
' 27
13.6
16.8
15.6
174
177
10-7
« 28
13.8
16.6
16.0
39-°
Mar. i
15.4
15.1
14.6
Mv
7 ^0
13.0
•9 fin
3-°9
•52
Some of these curves show that the larger part of improve-
ment is done during the first test. Usually, however, there
was a constant, though slow, progress in improvement until
the end of the twenty tests. Practise showed its influence
to the end of the series, with every indication that the averages
could, with further practise, be further reduced.
Secondary factors, such as fatigue, concentration of atten-
tion, indisposition, distraction, etc., all had great influence
1One day short by oversight.
2 Periods missed on account of sickness.
288 c. G. BRADFORD
on the work. The conscious processes were very susceptive
to these influences.
In learning the associations of the different keys mne-
monics proved helpful, by taking notice that certain associa-
tion letters formed initials of familiar names, etc., while some
had to be retained by sheer force of memory.
In some cases the peculiarity in sequence of the letters
or pairs of letters associated caused retardation in association
processes.
Regarding the theory of interference upon completing
one key and taking up a different one the writer believes that
probably there is a certain degree of interference; he also
thinks there is a certain amount of learning in the former key
transferred to the new one, and that there is a resultant of
forces present, which, in general, makes possible a lower time
rate. In the present investigation there was constant lower-
ing of the beginning time rate as progress from key to key
was made; which, it would seem, shows conclusively that
there was a" transference of learning from one key to another,
and that the evil effects of interference were largely neutral-
ized.
A NOTE ON THE EFFECT OF RHYTHM ON MEMORY
BY HENRY FOSTER ADAMS
University of Michigan
In spite of the poor standing of the "class" experiment,
the writer has been so impressed by the similarity in the
results of three such tests, that he thinks the data obtained
will be of general interest to students of memory problems —
the more so, as some of the conclusions are in direct contra-
diction to the results obtained in certain other investigations.
The object of the experiment was to test the effect of some
of the various kinds of rhythm upon the memory for numbers.
The rhythms used were the trochaic, iambic, dactylic,
anapestic and amphibrachic forms, together with a non-
rhythmic series. The investigation was carried on in the
time-honored way. The material consisted of 9 and 10
digits arranged haphazard, for example, 381427695. Such
a group was read at a rate of between 90 and 100 per minute,
Eight such groups constituted one series. One 9 digit series
and one 10 digit series were used for each of the different
kinds of rhythm.
The rate of 90 to 100 per minute was used in the endeavor
to obviate subjective grouping on the part of the subjects,
and because it would be about equally advantageous for the
two-part and three-part rhythms. According to Bolton's
work upon rhythm, it was found that 75 per minute was the
most favorable rate for two-part subjective rhythms, and 130
per minute for the three-part subjective groupings. The
intermediate rate was therefore used, for this experiment con-
cerned itself with objective groupings only. The subjects
were instructed to write down as many of the numbers as
they could remember immediately after the reading of one
group. They were told to be sure to get the right number in
the right place and to put dashes for the forgotten numbers,
289
290
HENRY FOSTER ADAMS
so that those given correctly should appear in the correct
place in the combination.
The results were recorded in the usual way. A number
correctly given and in the right place in the combination was
awarded a credit of 100 per cent. A transposition of two
numbers was given half credit, or 50 per cent. What might
be called a half transposition, or a number which was shifted
one place to the right or left with an incorrect number appear-
ing in its place, was given a value of 25 per cent. The results
are consequently given in percentages of the total amount
recalled.
The series were given in such a way as to obviate as far
as possible the effects of practice for any meter. Frequent
rests were given the subjects so that fatigue might not inter-
fere with the results.
A total of 180 subjects was used, 80 men and 100 women,
all of them coming from the class in introductory psychology
at the University of Michigan. The first group, consisting
of 50 persons, 25 men and 25 women, performed the experi-
ment during the winter of 1911 and 1912. During the winter
of 1912 and 1913 about 80 persons were experimented upon,
and the papers of 25 men and of 25 women were selected at
random. During the winter just passed (1913-1914) 30 men
and 50 women performed the same experiment.
Since certain differences appear in the masculine and
feminine types of reaction, it will be well to treat them at
first separately and then note the sex differences. The fol-
lowing table shows the results registered by the men in the
9 digit series.
TABLE I
MEN. 9 DIGITS
Group
N. R.
Troc.
Iamb.
Dact.
Amph.
Anap.
Av.
I
2
3
74.1
77.0
71.0
75-o
73-2
68.2
78.9
71.2
72.0
85.0
88.5
88.7
80.4
837
89.7
83.5
88.8
80.5
78.9
78.7
Average .1 73.7
71.8
74.0
87.5
82.1
87.3
794
N. R. indicates that the numbers were read without accent,
EFFECT OF RHYTHM ON MEMORY 291
or rather with the amount of accent as nearly uniform as
possible. The other headings show the rhythm used —
trochaic, iambic, dactylic, amphibrachic and anapestic.
Until the results of the third set of papers had been
obtained, there was no intention of using them for anything
but purposes of illustration in class. The records of two sets
had therefore been destroyed before any individual char-
acteristics had been worked out. Realizing that the averages
as shown in the tables may be greatly affected by some
extreme cases rather than representative of a general tend-
ency, the writer regrets this loss. However, he has worked
out the last set of papers not only in percentage values, but
also in terms of position by The Order of Merit Method,
where the effect of extreme cases is eliminated. There were
no substantial differences in the results obtained by these
two methods, certainly no differences of kind. There were
a few differences of degree.
A study of the table brings out these facts :
I. There is a variation of 6 per cent, in the recall of the
unaccented series amounting to 0.54 of one syllable when we
compare the best group with the worst. There is no correla-
tion between the ability to recall the non-rhythmic series
and especial ability to recall any particular kind of metrical
presentation. The one thing which does appear is that the
group which had the highest recall value for the unaccented
series had the lowest for the average of the three-part meters.
This same relation will not hold for the other two groups,
however, for the group which was the worst in the non-
rhythmic series was not the best in the three-part rhythms,
unless we omit the amphibrach. If we do omit the amphi-
brach entirely, we do find the reverse correlation, that the
poorest group in the non-rhythmic series was the best in
the three-part rhythms considered as a whole, while the best
non-rhythmic group was the worst in the three-part rhythms.
Another interesting result which appears from a study of
the table is that the group of men which did best in the non-
rhythmic series did better on the falling rhythms, the trochaic
and dactylic, than they did on the rising meters, the iambic
292 HENRY FOSTER ADAMS
and anapestic; while those who were the worst in the un-
accented series recalled the rising meters better than the
falling.
2. In two out of three groups and in the average of the
three groups, the trochaic form of rhythm has not as high a
memory value as the non-rhythmic series.
3. In two out of the three cases and in the average, the
iambic form of rhythm is better for the purposes of recall
than the non-rhythmic series. We are justified in stating,
then, that the iambic form of presentation is very slightly
better than that without rhythm. Considering the average
for the two-part rhythms, we find it to be slightly worse than
the unaccented series.
But it must be remembered that the series of 9 digits read
in two-part rhythm is scarcely a fair test, for one number is
left over, making an incomplete foot at the end of the line.
When this incomplete foot is unaccented, as in the iambic
form, the group has a higher memory value, in two cases out
of three, than when the incomplete foot is accented. It is
interesting in this connection to call attention to the fact
brought out by Miss Rowland in connection with visual
rhythms, namely, that a change in the minor element of the
series is more disturbing than a change in the major element.
4. The three-part rhythms are all better than the un-
accented series. The dactylic and anapestic forms are about
equal, with the dactylic very slightly in the lead and the
amphibrachic is the worst, but better than the two-part and
non-rhythmic series. It is entirely natural that the three-
part rhythms should be the best in this part of the experi-
ment, for most of us have been trained from our earliest
days in the grade schools to group numbers by hundreds,
thousands, millions, etc.
The following table shows the results registered by the
women in the 9 digit series.
The study of this table shows the following facts :
I. There is a variation of 3.7 per cent, in the recall of the
unaccented series amounting to 0.33 per cent, of one syllable
when we compare the best group with the worst. Here we
EFFECT OF RHYTHM ON MEMORY
TABLE IA
293
Group
N. R.
Troc.
Iamb.
Dact.
Amph.
Anap.
Av.
I
2
3
76.5
76.0
72.8
72.6
7M
6S.S
78.1
70.7
70.0
91.8
944
90.8
88.1
83.7
99.1
95-3
90-3
83.6
82.6
78-7
Average .
74-5
68.1
72.2
92.0
85.0
93-7
80.9
find that the best non-rhythmic group is also the best in the
two-part and three-part rhythm series, while the worst non-
rhythmic group is also the worst for all forms of rhythmic
presentation. The best non-rhythmic group is best in the
falling rhythms, while the worst non-rhythmic group is best
in the rising two-part and the falling three-part rhythms.
2. In all three tests, the trochaic form is not as good as
the non-rhythmic.
3. The iambic form is in general below the non-rhythmic,
but is better than the trochaic. Again, as with the men, the
series having the extra unaccented syllable has a higher
memory value than the series containing the extra accented
syllable.
4. The three-part rhythms are considerably better than
either the unaccented series or the two-part meters. The
anapestic form is best, the dactylic next and the amphibrachic
worst.
Averaging the results of the men and women in the 9 digit
series, we find that the three-part rhythm is best, the non-
rhythmic next and the two-part rhythm the worst of all.
There is also a tendency for the rising meters to have a greater
memory value than the falling ones.
Considering now the differences between the masculine
and feminine types of reaction, we find:
1. The total amount recalled by women is greater than
that recalled by men, 80.9 to 79.4 or 100 to 98.1.
2. In the non-rhythmic series, the women recalled more
than the men absolutely, but relatively less. For when we
consider that the women remembered actually more than
the men did taking the series as a whole, the percentage of
the series recalled in the non-rhythmic series is less for the
294
HENRY FOSTER ADAMS
women by a ratio of 92.1 to 93.0. The non-rhythmic form
of presentation, then, is relatively worse for the women than
it is for the men in spite of the fact that they recalled more
actually.
3. The two-part rhythm is better for the men, both
absolutely and relatively, than it is for the women, the
absolute ratio being 72.9 for the men to 70.1 for the women,
and relatively 92.0 to 86.8.
4. The three-part rhythm is better for the women, both
absolutely and relatively, by the ratio of 90.2 to 85.6 abso-
lutely, or 112 to 108 relatively.
5. With the men, the dactylic form of meter had a some-
what higher memory value, while with the women the
anapestic form was best.
Turning now to the consideration of the 10 digit series,
which was given to but two sets of persons, we find a relative
rise in the memory value of the two-part rhythms, and a
relative decrease in the three-part, showing that the ir-
regularity of the last foot was a determining factor. The
total amount remembered in this series was considerably
less than in the 9 digit series, as might naturally be expected.
Table II. shows the results of the men in the 10 digit series.
TABLE II
Group
N. R.
Troc.
Iamb.
Dact.
Amph.
Anap.
Av.
2
3
71.0
62.3
697
634
72.7
69.0
77.1
75-5
71.2
76.9
73-o
73.6
69.1
Average .
644
644
70.8
76.6
71.2
74-8
704
This table shows:
1. The recall of the unaccented series shows a variation of
almost one syllable when we consider the results of the two
groups. The best group in this series was absolutely the
best in the three-part rhythms, but relatively worse than the
other group. There is no evident correlation between the
recall of the unaccented series and either the falling or rising
rhythms.
2. The trochaic rhythm is equal to the non-rhythmic
EFFECT OF RHYTHM ON MEMORY
295
series. This is true of the average, the two series differing
somewhat. In one, the trochaic is better and in the other
worse than the unaccented series.
3. The iambic series is better in both cases than the non-
rhythmic. On the average, then, the two-part rhythm with
series of 10 digits is better than the unaccented series.
4. All of the three-part are better than the two-part
kinds of rhythm, the dactylic holding the highest place, the
anapestic second and the amphibrachic coming last, as was
the case in the 9 digit series.
The following table gives the results of the women in the
10 digit series:
TABLE II A
Group
N. R.
Troc.
Iamb.
Dact.
Amph.
Anap.
AY.
2
3
63.5
61.0
68.5
64.1
69.8
654
72.9
70.4
68.S
74-3
72.2
69.8
66.9
Average .
62.0
65.5
67.0
71.2
68.5
73-o
67.9
A consideration of this table brings out the following
points :
1. There is very little variation between the two groups
of women in the recall of unaccented material, only a quarter
of one syllable. The group which was absolutely worst in
the recall of the non-rhythmic series was relatively better
in the recall of the three-part rhythms.
2. In both series, both forms of the two-part rhythm are
better than the unaccented series. The iambic form is better
than the trochaic.
3. Each of the three forms of three-part rhythm is better
than either of the two-part meters. But with the women,
the anapestic form of meter is better than the dactylic. The
amphibrachic is lowest of all.
Taking into account the sex differences, we find that:
I. The total amount recalled by the men is greater than
that recalled by the women, 70.4 to 67.9 or 100 to 96.5.
Considered absolutely, the women recall less in all the
different forms of rhythm. This is true of the average of the
296
HENRY FOSTER ADAMS
non-rhythmic, two-part and three-part rhythms. The women
did recall slightly more in the trochaic form. Considered
relatively, we find the men to be better in the non-rhythmic
series, worse in the two-part meters, and slightly better in
the three-part rhythms, when we consider all the three-part
meters together.
2. The men are best in the dactylic meter and the women
in the anapestic.
The task would be incomplete if no endeavor were made
to bring together the results of the 9 digit and the 10 digit
series. It would, of course, be unfair to average them, but
it is possible to take the average recall for each group and
reduce the amount recalled for each kind of meter to a
percentage of this average. This would do no injustice to
any group considered alone and would make comparisons
possible. Combining the previous tables for the men, and
omitting the amphibrachic form of meter — for it shows
nothing striking — we obtain this table:
TABLE III
Group
N. R.
Troc.
Iamb.
Dact.
Anap.
9 digits . .
I
92.2
97. -z
98.0
IOC.9
III.7
10 digits
2
3
2
97-4
90.2
96.7
•yj j
93-6
86.6
94-8
90.1
91.5
99.0
1 12.0
II2.8
IOC.O
105.8
II3.O
10^.2
3
90.1
91.7
99-8
109.2
105.2
Av. of 9 digits
Av. of 10 digits. . . .
92.8
9i.S
91.0
91-5
93-3
100.6
IIO.2
IO9.O
IIO.O
106.3
Difference
I.I
o.c
7.7
1.2
•5.7
The main value of this table seems to be to show the
amount of damage done by introducing irregularity into the
series. There is a difference of 1.3 in the non-rhythmic
series in favor of the shorter groups. The two-part rhythms
are better in the 10 digit series by an average of 3.9 and the
three-part rhythms are better in the 9 digit series by an
average of 2.45. On the average, the three-part rhythms
seem to be affected less by the introduction of an irregularity
than do the two-part, 3.9 to 2.45. Moreover the falling
EFFECT OF RHYTHM ON MEMORY
297
rhythms are less affected by the extra measure than are the
rising meters, 0.85 to 5.5.
TABLE \\\A
WOMEN
Group
N, R.
Troc.
Iamb.
Dact.
Anap.
0 digits
I
qi.c
86.8
03 .c
IOQ.Q
118.8
IO digits
2
3
2
92.3
92.5
Ql.O
86.6
83.2
08. i
86.0
88.9
IOO.O
1147
II5-2
I O-i.-i
"5-9
114.6
106.4.
3
91.2
95-8
98.0
I05.I
108.0
Av. of 9 digits
Av. of 10 digits ....
92.2
91.4
83.3
96.5
89.4
99.0
II3.8
105.0
116.0
107.6
Difference
0.8
13-2
9.6
8.8
8.4
This table shows that there is a slight relative difference
— 0.8 — in the recall of the non-rhythmic series, the difference
being in favor of the shorter series. The two-part rhythms
are better in the 10 digit series by an average of 11.4 and the
three-part rhythms are better in the 9 digit series by an
average of 8.6. The two-part rhythm, then, is affected a
little more by the introduction of irregularity than is the
three-part rhythm. In both the two-part and the three-part
rhythms, the falling meter is the more affected, n to 9.
There is less disturbance of the three-part rhythms than of
the two. The other points which are brought out by this
table have been considered before, so may be omitted here.
When we consider the sex differences as brought out by
a comparison of the results of the 9 digit and the 10 digit
series, we find that:
1. The feminine recall is better for the 9 digit series by a
ratio of 80.9 to 79.4 or 100 to 98.1, whereas in the longer
series the masculine recall is better by a ratio of 70.4 to 67.9
or 100 to 96.5. Considering the whole experiment, then, the
men have slightly better memories for numbers than the
women, the ratio being 100 to 99.1. This difference is very
slight and might be called negligible.
2. The irregularities introduced into the series are more
disturbing to the women than to the men, 8.16 to 2.8.
298 HENRY FOSTER ADAMS
3. The irregularity on the average affects the three part
rhythms less than it does the two-part for both sexes.
4. With the men, the dactylic form of meter is the best in
all cases; with women, the anapestic.
5. The irregularity affects the rising meters more than it
does the falling in the case of the men; whereas the falling
meters are more affected in the case of the women.
Since, as has been seen, the 9 digit series introduces a
disturbing factor into the two-part rhythms and the 10 digit
series into the three-part rhythms, it will be interesting to
consider the two-part rhythm 10 digit series together with
the three-part rhythm 9 digit series. The results obtained
from this consideration differ from the rest of the results
only in this particular, that, with one exception, any rhythm
is better than no rhythm at all. This exception occurs in
the case of the men and with trochaic rhythm, it being exactly
equal to the non-rhythmic series.
This way of regarding the results also raises the feminine
recall somewhat above the masculine, but shows that it is a
more precarious thing, very easily disturbed.
DIAGNOSTIC VALUES OF SOME PERFORMANCE
TESTS1
BY THOMAS H. HAINES, M.D., PH.D.
Bureau of Juvenile Research, Columbus, Ohio
In preliminary mental examinations of a number of recent
admissions to the Ohio Girls' Industrial School, the results
obtained from the Binet and Point Scale ratings of intelli-
gence, set apart three groups, which it seemed desirable to
investigate further. These groups follow:
1. Twenty-one high-grade morons, whose Binet ratings
average II years, with a mean variation of 3 years, and whose
Point Scale ratings, transformed into years, average n.6
years, with a mean variation of .4 year.
2. Sixteen, concerning whose intelligence defects we are
in doubt, because of the disparity between the Binet and
Point Scale findings. Binet ratings of these sixteen average
1 1. 6 years, with a mean variation of .2 year. By Point Scale
rating, these are all 12 years or more. One only is flat 12.
Four are over 15 years.
3. Twenty-six, who show no defect in intelligence by
either of these ratings, being twelve years or more by both
scales. Fifteen of them get credits for more than the 82
points for fifteen years.
The actual ages of these sixty-three girls, reckoned by
the nearest birthdays at the times of examination, range from
12 to 1 8 years. The median ages for each subgroup, in the
same order as above, are 15.8 years, 16.5 years, and 16.5 years.
The average ages for each subgroup, in the same order, are
15.2 years, 16 years, and 15.9 years. The median age of
the sixty-three girls is 16.35 years, and the average age of
the sixty-three is 15.7 years.
For the further study of these cases, recourse was had
1 Read before Section L of the American Association for the Advancement of
Science, December 31, 1914.
299
300 THOMAS H. HA1NES
to the following tests. We lack standards in all these tests.
For the more delicate diagnosis of intelligence and other
mental defects in persons of mentalities of more than ten
years, there is no method of procedure in any wise comparable
for accuracy of measurement to the Binet and Point Scale
for the earlier stages of development. These performances
and other tests have been used widely as supplementary aids
to psychologists' native intuitions. Methods of testing have
been developed, but no standards of the meaning of results.
These subjects having been rated by the two scales (Binet
and Point), and classified with respect to the 1 2-year line,
we have a means of preliminary evaluation of these supple-
mentary tests. We may ascertain which do and which do
not correlate with the scales. It is not standardization.
It is merely evaluation. Standardization is not possible
with defectives.
LIST OF THE SUPPLEMENTARY TESTS
1. Picture Form Board (farm scene, mare, colt, chicken,
sheep. Two right-angled triangular pieces to fit into
an isosceles triangle). (Healy.)
2. Construction Puzzle (A). (Healy.)
3. Construction Puzzle (B). (Fernald.)
4. Labyrinth (B). (Boston Psychopathic Hospital.)
5. Visual Verbal Memory Test. (Schmidt.)
6. Auditory Verbal Memory Test. (Thorndike.)
7. Learning Test. (10 symbols for 10 figures; 4 lines.)
8. Cross Line Test (B). (Healy.)
9. Motor Coordination. (Whipple.)
10. Opposites. (Lists I. and II. of Woodworth and Wells.)
11. Completion Test. (Boston Psychopathic Hospital.)
12. Moral Discrimination Test. (Boston Psychopathic Hos-
pital.) (For girls.)
Table I. presents, for comparison, the summaries of
results, from each of the three groups of girls for each of the
first eleven tests.
The first four tests were given with simple directions, and
the time in seconds and the moves or errors, recorded. These
are averaged for each of the three groups.
PERFORMANCE TESTS
301
E c
* 4
•5-S
.2.S
^ «
^3 S2 *^>
•0.3-, fc ^$5 *« ££3 M
fe nH *> 00 oS^'5 w
QK«^ *J^_, . .2 » J! ?^Ji
^S^eJ^cJ^WdJtjWt^OwalJ
£ t § c^ S.>: S ^^ W^ o g B.8 g-o
K . .1 5 9l tl^rSrtlSrtbSiS*3., C
*J*S^iJll^i^^*lS8l
»^ H-i i^ <-M /*» ^ **^ <»i<i\/*tH^K. S^. «^-« K «
PQ M
2 S £
o > >
t % %
5 S S
q\ vq HH
06 vd &>
32 ovo o o
.a •> v CD
O^vo ^ to
HH O* tO
: s I °$
^ «
2 5
2 ^
O HH
tO O
s a
ON VO
10 l^ 00 N T}- i O rj-
toOO 10 10 HH | to o»
$ ^
PQ ~
a S
S PH
fe jj
» >^ ^r^'2 " 2 ^
c32
b>.00 ON O HH
302 THOMAS H. RAINES
In the Visual Verbal Memory test, a sheet was handed
the subject on which the story of a fire was printed, one line
to each detail, and she was directed to read it once aloud, and
to do it carefully so she could remember as much as possible.
In the Auditory Verbal Memory test, the shipwreck story
was read to her four times. Times for recall and numbers
of details are averaged for each of the Memory Tests.
In the Learning test, times required to fill three lines, and
to learn, to the subject's satisfaction, were recorded and
averaged, as well as the numbers of errors in filling the 4th
line from memory, and the times for the memory work.
In Cross Line (B) test, the double cross was drawn on a
piece of paper and the arabic numerals, 1—9, written in the
nine spaces. The explanation of how the symbols could
stand for the numbers was accompanied by a drawing of
each symbol as the number was named, and an illustration
given of writing a symbol under a number. The subject was
asked to note carefully the arrangement. The paper was
then turned over and six numbers written in a line, and the
subject asked to make the corresponding symbol under each.
In Motor Coordination, we average numbers of dots made
in 30 seconds, and the errors, i. e., dots on lines or outside of
squares intended, or a second dot in a given square.
In Opposites, we average the time and errors (wrong
words and omissions) for 40 words. The observer wrote the
opposites and was not hurried.
In the Completion test, the same method and data were
used.
From an inspection of the table it is at once evident that
some of these tests give significantly different averages in
each of the three groups; others reveal more or less clear
differences as between the not defective (mentality of 12 years
or more), and the doubtfully defective. Others show dif-
ferences between the doubtfully and the definitely defective.
Others still show no sufficient differences between averages
of any two groups to make them of any diagnostic value in
this part of the developmental period.
I. In the first class of tests, are (a) the Picture Form
PERFORMANCE TESTS 303
Board, (b) Construction Puzzle (A\ and (c) Opposites. In
the Picture Form Board averages, there appear significant
differences in the increasing time throughout, and increased
errors in the defective. Construction Puzzle A gives marked
differences in both time and moves. It seems a valuable
means of differentiation in this region. Opposites by number
of errors gives significant differences. Defectives are set off
strikingly both by time and numbers of errors.
2. In the second class, the Labyrinth averages indicate
some diagnostic value in differentiating between the "non-
defectives" and the "Doubtfuls," both in time and number
of errors. Both the verbal memory tests on the other hand
seem to differentiate between the "high-grade defective"
and the "doubtful" in the number of details recollected.
The norms seem to be ten visual, and nine auditory details
for the defective, and twelve visual and nearly ten auditory
details for the doubtful. The Cross Line test again separates
the normals and doubtfuls strikingly. The Completion test
is much less valuable. It separates non-defective from doubt-
fuls by time, and doubtfuls from defectives by numbers of
errors. The latter difference has most weight.
3. In the third class, Construction Puzzle (£), the Learn-
ing test, and Motor Coordination seem to be of no service
for differential diagnosis in this part of the development of
mentality from ten and one half years to fifteen.
We compare the median place of location of each one of the
ten acts, in the Moral Discrimination test, for each of the
three groups, and also for a group of fourteen first-year
high-school girls of ages from 14 years to 16.5 years, with
results as in Table II.1 The figures give the median positions
assigned each act by each of the four groups.
"Flirting with a nice young man" is deemed much worse
by the normal girls than by any of the delinquents. Of the
delinquents, the high-grade defectives consider it worse than
the other subgroups.
"Taking a hair ribbon," is rated about six by all, i. e.,
1 For the data from the reactions of fourteen high-school girls, I am indebted to
Dr. R. Pintner, of Ohio State University.
3°4
THOMAS H. HAINES
TABLE II.
Normal
Delinquent
is^Yr.
H.S.
26
Not
Defec.
16
Doubt.
Intell.
Defect.
21
High-
grade
Defec-
tives
To flirt with a nice young man on the street
5-5
57
3-5
5-5
4-5
8.5
9-3
2.4
i-S
9.1
8-5
5.8
2.7
5-3
o.i
8.6
8.1
1.9
4-5
7-5
8.S
6.2
1.8
5-3
57
8.2
8.7
2.4
7-5
74
7.2
5-9
2-5
57
57
7-5
7-3
1.9
7-5
7-5
To take a hair ribbon from your employer when she
knows nothing of it ...
To spend the night in a hotel with some young man. . .
To take a box of candy from the store where you work. . .
To tell a wicked lie about some girl
To get mad and break the dishes when the woman for
whom you work finds fault with you
To spank the baby because you are out of patience
To put poison in the food of some one whom you dislike .
Not to go to Sunday school and church, and never to
read your bible
To throw scalding water on the cat
better than the middle. "Spending the night with a young
man in a hotel" is rated a better thing to do by the normals
than by any of the delinquents. Perhaps this is due to
failure to understand what is meant. Of the delinquents,
those of doubtful intelligence defect place it nearest the worst
thing to do.
"Taking a box of candy" is placed about 5.5 by all groups.
"Telling a wicked lie about a girl" is much worse with the
normal girls than the delinquents. The delinquents all place
it about the same, but those without intelligence defect
consider it least bad. "Breaking the dishes when scolded"
is worse for the high-grade defectives than any other group.
" Spanking the baby when out of patience" is least bad for
the normal girls, and judged worse by the high-grade defec-
tives than by the other delinquents. "Putting poison in the
food of one you dislike" is rated about the same by all, but
some worse by the high-grade defectives and those of no
intelligence defect.
"Not going to Sunday school and church" is the worst of
the ten offences according to the normal girls. It is moder-
ately bad for the delinquents with no intelligence defects,
and moderately good for the doubtfuls and defectives.
"Throwing scalding water on the cat" is the second best
PERFORMANCE TESTS 305
act with the normal girls, and is moderately good for all the
delinquents.
There are no significant differences between the groups of
delinquents in regard to their moral judgments. This test is
of no value for differential diagnosis as between these groups.
The startling moral judgments of the normal girls suggest
the need of psychological inquiries into ethical foundations
in the minds of girls.
By our findings then, the values of the tests for differential
diagnosis of our three groups are as follows:
1. Tests of Value for Both Distinctions
The Picture Form Board.
Construction Puzzle (A).
The Opposites.
2. Tests Good for Differentiation of the Not Defective from the
Doubtful
The Labyrinth (B).
The Cross Line (B).
3. Tests Differentiating the High-grade Defective from the
Doubtful
Visual Verbal Memory.
Auditory Verbal Memory.
4. Tests of Doubtful Diagnostic Value
Completion.
5. Tests Showing No Definite Diagnostic Value
Construction Puzzle (B).
Learning.
Motor Coordination.
Moral Discrimination.
PROCESSES REFERRED TO THE ALIMENTARY
AND URINARY TRACTS: A QUALITATIVE
ANALYSIS1
BY E. G. BORING
Cornell University
Recently the writer promised a descriptive study of certain
complex organic processes that have their origin in the ali-
mentary canal and the uro-genital system.2 The present
paper, which fulfills that promise, presents introspective
descriptions of thirst, hunger, nausea, the call to defecation,
defecation, the call to urination, and urination. Descriptive
accounts were obtained, a few in the series of the earlier
experiment, some in subsequent series, and others from written
reports prepared by the observers outside the laboratory
when the particular complex was experienced. These reports
are more reliable than the usual answers to a questionary,
because they were written down at the time of the experience
by trained observers who had consciously adopted the
attitude of psychological observation. It is from such
accounts that quotations are made in all cases where there is
no specific indication of another source.
The procedure was to provide each observer with a booklet
which contained space for the description of the various
processes mentioned above. At the beginning of the book
was the following instruction:
You are requested to give as careful an introspective description of the various
complex organic processes listed herewith as possible. Please select a time when the
particular process is intense, and arrange to retire from distractions and write a careful
report on these pages.
In the description of the complex organic processes, the observer is advised not
to avoid Kundgabe or common-sense language, but to deal freely in the meanings of
processes as well as in their pure description. It is instructive to be told that a complex
feels 'as if . . .' or 'like. . . .' In description, free use should be made of such un-
systematic descriptive words as sharp, keen, piercing, lancing, thrusting, insistent,
dull, diffuse, definite, pungent, hard, clear, bright, heavy, full, steady, stable, etc- It
1 From the Psychological Laboratory of Cornell University.
2 'The Sensations of the Alimentary Canal,' Am. Jour. Psych., 26, 1915, I.
306
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3°7
should be borne in mind, however, that the exact significance of such words is not such
that it always leads to a clear and unequivocal interpretation, for their meanings differ
from individual to individual. The sensations should therefore, whenever possible,
be compared in quality with such better known sensations as those of the cutaneous
and, possibly, of the kinesthetic senses. The observer should also be on the watch
for such characteristics as are matters of extension or of temporal course, and should,
when able to do so, localize the complex as accurately as possible within the body.
There were nine observers, — six men (A, B, C, Z), E, F)
and three women (Jf, Y, Z). Four of the observers (A, B,
X, Y) had received the doctorate in psychology; three of
these were instructors in psychology. Two others (C, Z)
were graduate students. The rest were undergraduates in
the advanced courses in the laboratory.
THIRST
Observers A, B, and X undertook to go without water or
liquid food for long periods, and to keep a running account of
the experience. Abstracts from their reports follow.
Observer A. — 6:30 P.M. Began experiment.
9 A.M. (14^ hrs.) "Tongue coated. Saliva secreted rather freely. It is
rather that I want a drink than that I am thirsty. Pressure and kinesthetic complex
in my mouth (not throat) keep reminding me at intervals of the experiment, and I
imagine how a drink would feel."
2:30 P.M. (20 hrs.) Warm day. "Have eaten a few figs. Saliva still quite
actively secreted, and mouth and throat kept wet by swallowing frequently and licking
mouth with tongue. Lips are getting dry. The thirst rather expresses itself in my
imagination than in sensation changes. While I am doing other things, I suddenly
break off automatically and start for the water faucet, without thinking of what I am
doing. Once I got the water turned on before I thought."
4:30 P.M. Ate piece of cheese.
5:30 P.M. (23 hrs.) "More unpleasantness now. A little more intense tactual
sensations in mouth. Saliva secretion seems less. Slight headache. Number of
times that I start for faucet increased; I should think I have started twenty times in the
last hour. The start is like this: I am working and come to a break in the work; auto-
matically I start to get up and move in the direction of the faucet. The tactual sen-
sations in the mouth and the kinesthetic sensations in the tongue are then in the back-
ground. Then there is usually a vague visual image of a glass for water or a visual
image of the sink or basin toward which I have made my involuntary start. Before
I have moved more than a bit with my body, before I have taken a step, I get a kines-
thetic shock or check, and remember the experiment. It then takes me some time
to get back to work. The images of the basin or faucet persist. I can not keep them
out of mind. This, is, for the most part, where the unpleasantness comes in. The
unpleasantness does not seem to attach to the mouth sensations at all. The sensations
are not strong. They seem like tactual pressure, are superficial, and spread out over
the roof of the mouth. They are not in the cheeks, on the tongue, or in the throat back
308 E. G. BORING
of the root of the tongue. The weakness and paucity of mouth and throat sensations is
the striking thing so far."
n P.M. Ate three cakes.
1 A.M. (30^ hrs.) "Weak in legs. Headache partly gone. Have taken a
bath and am less irritable. Starts toward faucet and images of drinking less frequent.
Roof of mouth very dry, an intense and thick tactual pressure with a very little livelier
sensation almost like cutaneous pain or the prickle that one gets in drinking lemonade.
In back of roof of mouth a narrow arch of deeper, duller sensation, more intense, of the
quality of muscular pressure. Some time ago there was mixed with this a deep, dull
pain, almost like a muscular cramp."
10 A.M. (39/^ hrs.) "Slight headache; tactual sensations in roof of mouth;
tendency to keep licking roof of mouth and to swallow at saliva, which is scant and
seems thick. There are frequent temptations to drink, which differ from the invol-
untary starts for the faucet in that they arise only when the mouth sensations catch
attention. Lips chapped." At this point the experiment was discontinued.
Observer B. — Midnight. Began experiment with a drink of water. 9 A.M.
(9 hrs.) "Tongue coated. Feel as if water would be good for me, but would taste
unpleasant. Not thirsty. Mouth dry and sort of puckered — drawn."
9:30 A.M. Ate cake of almond chocolate.
11:30 A.M. (ii^hrs.) "Not thirsty. Mouth a trifle dry."
1 130 P.M. (i3/^ hrs.) "Beginning to get thirsty. Thirst consists of sensations
from tongue, roof of mouth, and throat, — principally tongue. Flow of saliva copious.
I keep moving my tongue about on roof of mouth. Tongue and roof of mouth feel
sort of dead, when brought together. Mouth feels dry, a little bit drawn or puckered.
There is nothing in these complexes when I get them clear but pressure; they mean
thirst, i. e., they tend to run off into visual imagery of water or kinesthesis of going for
water. This tendency for the focus to shift makes it hard to keep the sensations, as
such, clear. The pressures are easy to refer below the surface, though they remain
on the surface too. I can feel the whole tongue affected, and, at times; this feeling does
not seem to be pure pressure, but an ache-like pressure, perhaps muscular pressure.
The ache is not explicit; I could not analyse it out. The sensation is more like the fore-
runner of an ache. I also feel a slight ache in the back of each jaw, almost a cramp,
like the ache from eating something too sour." B also fully describes the impulses to
go for water that were noted by A.
2 P.M. Ate three slices of bread and a little jam.
5 P.M. (17 hrs.) "Thirst getting insistent, a puckered, swollen, dry feeling in
tongue, cheeks, and roof of mouth. Lips very dry. Also a vague cenesthetic ache or
weakness. I am a little unsteady upon my feet, and can not articulate with certainty."
5:15 P.M. Ate piece of cheese, n P.M. Ate a few crackers and smoked.
Midnight: a few more crackers.
Midnight (24 hrs.) "Thirst insistent. Consists merely of dry mouth (pressure).
Slight ache in head and limbs. Weak all over. Nothing new." The starts for the
faucet are still reported.
10 A.M. (34 hrs.) "Feel pretty good. Mouth dry. Saliva no longer free.
General body aches diminished. Impulses to get water less frequent."
10:30 A.M. Ate crackers and cheese.
i P.M. (37 hrs.) "Very thirsty. Mouth getting actually dry. Lips have to
be moistened constantly. Impulse to get water very strong. Chief symptom is
general weakness and lassitude. Going upstairs tires me out.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 309
" I tried experiment of putting cracked ice in a thin rubber bag in my mouth. Just
as good as a drink! However, pleasant cold in mouth sets up impulse to swallow. If
I swallow it as far as my throat, the cold in the throat is even more pleasant than that
in the mouth. This surprises me because I have very few thirst sensations in my
throat and ordinarily do not notice my throat. It is perhaps barely dry and aches a
little, but I have no desire for water there until I put the bag of ice water in my mouth.
The thirst disappeared entirely as long as the bag was in my mouth. The bag even felt
wet, although it was actually dry. I could not keep it in long, however, because my
mouth soon ached from the cold. On removing it the thirst returned immediately.
As soon as the ice in the bag had melted, it ceased to be effective."
3 P.M. Ate buttered toast and ham. Rather weak. B found that two drops of
acetic acid on the tongue "tasted good" and relieved the thirst for 7 minutes. Thirst
had returned in its former intensity after 15 minutes.
4 P.M. (40 hrs.) "Headache. Mouth and face warm. Limbs ache. Mouth
very dry, a dull pressure complex, with brighter surface qualities."
7 P.M. Ate toast, ham, and cheese.
loP.M. (46 hrs.) "Tired, achy, weak. Mouth almost parched. Coated and a
bitter taste. Lips have peeled a little; feels as if cheeks might soon. Headache."
2 A.M. (50 hrs.) Same as before. The thirst is a "dry mouth (cutaneous and
subcutaneous pressure) and a taste." The rain outside reminded B of water and
prevented his sleeping, so he discontinued the experiment at this point with a quart of
water and some crackers. The mouth tended to dry quickly after the first drinks, but
five minutes later he was quite comfortable. The next morning there were no notice-
able effects of the water fast.
Observer X. — 8 A.M. Began experiment.
8 A.M. (24 hrs.) "Thirst began to show itself by dryness of the lips. I did not
clearly realize that I was thirsty, but found myself at intervals going to the water cooler.
All along I was being reminded that I had certain sensations in my mouth that in-
dicated thirst by finding the motor habit of securing a drink set off. Later the feeling
of dryness extended beyond my lips to the inside of my mouth, especially the roof,
which felt somewhat as if shrivelled. My tongue began to feel changed in shape; it
seemed to be smaller, rounder, rather swollen at the back. Later the dry, somewhat
irritated, rough feeling increased in the roof of my mouth and extended to my throat as
well. I now feel as if I had a 'dry sore throat.' There are little 'painy' sensations in
my throat and on the upper surface of my tongue. My dry lips no longer bother me.
The sensations on my tongue feel very much like those that I have after I have scalded
my tongue by drinking something too hot. What moisture there is in my mouth
seems to have got thick and sticky. The different parts of my mouth have a tendency
to stick together when I try to spread them out."
4 P.M. (32 hrs.) "Mouth cavity insistently calls attention to itself. Throat
seems hot, and inflamed, and sore; mouth somewhat dry and sticky; tongue achy and
sore. There are dull pains, with perhaps a dull pressure, localized inside the tongue
at its base. Dried-up feeling on top of tongue." Experiment discontinued at this
point.
Extracts from reports of thirst experienced under less
extreme conditions by other observers follow.
Observer C. — "Dryness expresses the complex as a whole. The qualities seem to be
a little warmth, a little pain, which is very mild but seems to be quite important as a
310 E. G. SORING
component, and a quite distinct pressure. The pressure is like the result of stretching
the skin, as if the mucosa had become shrunken and were stretched by a too voluminous
submucosa. These sensations come from the sides and back of the throat in the region
of the uvula, actually from little more than the soft palate. In the mouth there is a
sort of stickiness or dryness."
Observer D. — " Feeling of dryness in throat. Feels cottony."
Observer E. — "A taut feeling of the upper esophagus and back of throat. Tongue
feels large. Saliva thick and sticky. Sort of achy feeling in tongue and roots, extend-
ing to near-by parts of throat."
Observer F. — "Uncomfortable pressure in mouth, rather 'puckery,' localized in
middle and back of tongue and roof of mouth. Also a more vague pressure in upper
throat, not 'puckery.'"
Observer Y. — "Light pressure in mouth cavity and soft palate; faint muscular
pressure at junction of jaws; perception of wetness under tongue (saliva). Pressure
most prominent. Later pressure sensations at junction of jaws and also in cheeks
became more intense and clearer. The perception of 'warm-dryness' in mouth was
less prominent, but by no means obscure. Finally general muscular sensations took
on a weak sort of tired feeling."
Observer Z. — After eating too much candy: "Tongue puffy and furry, rather warm
and dry(?). Strong imagery of glass of water, of coolness of glass in hands, of coolness
of water in mouth and throat, and of coolness of water in stomach."
From the foregoing reports it is evident that mild thirst
consists predominantly in the going to get a drink, or in
imagery which anticipates a drink; that is to say, thirst may
not be conscious as such, but may be merely the meaning of
a complex situation. Mild thirst is accompanied by a pres-
sure-pattern on the tongue and the roof of the mouth, and
sometimes by one in the throat. This pattern, which is
seldom prominent in weak thirst, becomes marked as thirst
gets more intense. Then the pressures increase in intensity
and spread more frequently to the throat; the saliva flows
freely, and the sensations involved in swallowing it figure in
the complex. After a longer period of thirsting (twelve to
twenty-four hours) the saliva no longer flows freely, the
mouth becomes 'thick' and 'sticky' and 'dry,' the lips are
dry and have to be moistened frequently. The pressures of
the mouth are referred farther below the surface and, at least
in the case of the tongue, take on the ache-like- character of
intense muscular pressure. The oral sensations touch off
the desire to get a drink. In a still more extreme stage, the
painful qualities become more marked in the tongue, in the
roof of the mouth, and sometimes in the throat. There is
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3 1 1
general bodily lassitude and weakness. There is no evidence
of any qualitative peculiarity other than the pressure-pains
of usual organic experience. It is the situation and not the
specific quality that makes the experience one of thirst. The
oral sensations that are typical of thirst are referred always
to the mouth and sometimes to the throat. When they are
felt in the throat they are usually much less intense than
those in the mouth, and some observers, even in extreme
thirst, do not find them in this region at all. Obviously the
statement of the textbook of physiology, that cour sensations
of thirst are projected more or less accurately to the pharynx/1
needs revision.
It appears that the thirst-perception in the mouth can be
adequately neutralized, for short times at least, by a per-
ception of wetness, even though the wetness be illusory.
Cold ice-water in a rubber bag, which is dry on the outside,
afforded one observer complete relief. The bag felt wet,
like a draught of cold water.
HUNGER
The recent physiological work which has resulted in the
correlation of hunger with stomachic contraction has naturally
suggested a tentative psychological definition of hunger.
Cannon and Washburn2 have separated hunger from appetite
and characterized the former as an 'ache.'
"Appetite is related to the previous sensations of the taste and smell of food; it
has therefore . . . important psychic elements. It may exist separate from hunger
as, for example, when we eat delectable dainties merely to please the palate. Sensory
associations, delightful or disgusting, determine the appetite for any edible substance,
and either memory or present stimulation can thus arouse desire or dislike for food.
"Hunger, on the other hand, is a dull ache or gnawing sensation referred to the
lower mid-chest region and the epigastrium. It ... is likely to grow into a highly
uncomfortable pang, less definitely localized as it becomes more intense. It may exist
separate from appetite, as, for example, when hunger forces the taking of food not only
distasteful but even nauseating. Besides the dull ache, however, lassitude and drowsi-
ness may appear, or faintness, or headache, or irritability and restlessness such that
continuous effort in ordinary affairs becomes increasingly difficult. That these states
1 Howell, W. H., 'A Text-book of Physiology,' 1908, 272.
8 Cannon, W. B. and Washburn, A. L., 'An Explanation of Hunger,' Amer. Jour.
Physiol., 29, 1912, 441. See also Cannon, W. B., 'A Consideration of the Nature of
Hunger,' Pop. Sci. Mo., 81, 1912, 291.
312 E. G. BORING
differ with individuals — headache in one, faintness in another, for example — indicates
that they do not constitute the central fact of hunger, but are more or less inconstant
accompaniments, for the present negligible. The dull, pressing sensation is the con-
stant characteristic, the central fact, to be examined in detail."1
Carlson2 also distinguishes between appetite and hunger,
although he disagrees with Cannon and Washburn in regard
to the nature of appetite. Hunger is pain.
"Pure hunger, not accompanied by 'appetite,' can be experienced, if during hunger
attention is fixed on the hunger pangs themselves. . . . When this is done, hunger in its
various stages becomes different degrees of pain."3
"It seems to me that the pain experienced from contractures or 'cramps' in
skeletal muscles and in the intestines is different from hunger pangs, even though pain
is inherent in hunger. The difference may be only an apparent one, due to the fact
that the latter pains arouse the memories of previous agreeable experiences with food."
"It would . . . seem that hunger contains elements of kinesthetic sensation as well
as pain, the latter predominating in strong hunger."4
The observers in the present experiment were asked to
report upon ' hunger' only, under the following instruction:
"The observer is warned to distinguish between 'hunger' and 'apptetite.' Hunger
is more nearly 'sensational' and is said to be always experienceable in isolation when
the attention is directed toward it. Hunger usually ceases as soon as food is taken.
Appetite is more 'ideational' and persists after food is taken. It is the desire for food,
the opposite of aversion. Appetite probably constitutes the motive for eating dessert
at any meal. We are here interested in the description of hunger only."
In spite of this limitation the observers reported general
weakness and faintness, headache, visual and oculomotor
disturbances, factors which they recognized, however, as
secondary to "hunger proper." The qualitative nature of
the more immediately relevant experience may be shown by
extracts from the reports.
Observer A. — At noon after eating no breakfast: "Dull pressure of considerable
intensity in area above umbilicus. With this also pain, an achy, gnawing pain. Or
else a muscular tension, a feeling of muscular contraction in this region, gives the
meaning of 'gnawing.' I think I sometimes have mere 'emptiness,' i. e., all this com-
plex except the gnawing, achy pain, but this is only a general impression. It is common
for me to say, 'I am empty, but I am not hungry."'
1 Pp. 441 f.
2 Carlson, A. J., 'Contributions to the Physiology of the Stomach. II. The Re-
lation between the Contractions of the Empty Stomach and the Sensations of Hunger,'
Amer. Jour. Physiol., 31, 1913, 175.
3 P. 1 86.
4 P. 189. For another discussion of hunger as pain, and a resultant symptoma-
tology, see Jones, A. A., 'Hunger Pain,' Jour. Amer. Med. Assoc., 59, 1912, 1154.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3 1 3
Observer C. — "On the sensory side hunger is composed of temperature and mus-
cular sensations. The temperature is warmth, in the main, but at times there is some-
thing resembling cutaneous paradoxical cold. What I have called muscular sensation
seems to be a sort of strain or pressure, much like that from the contraction of any
skeletal muscle. The localization is in the stomach, and that organ feels as if it were
pulling itself into a knot just as the hand does when the fist is quite tightly clenched."
Observer D. — "A dull, yet insistent, ache, — very diffuse. It seems to cover an
area of about 20 cm. diameter, fairly deep, and extending upward from the point
of the sternum. Sometimes it becomes more intense for an instant at some point.
This point changes continually. At times the diffuse ache becomes weak and gives
way to a sharper pain, a little higher up. This lasts about a minute and then there is
a return to the previous conditions."
Observer E. — "Hunger begins with an unpleasant, uncomfortable feeling below
the sternum. This gradually and quickly changes to a raw painful feeling, as if of the
rubbing of the two stomach walls together. This feeling of achy, rubby pain increases
in intensity until even the esophagus seems to be uncomfortable in its lower parts."
Observer F. — "There seem to be general pressures all through the abdomen, rising
up to and above the sternum. They are most definite in the region just below the
sternum. They sometimes become a dull ache. (The whole thing is instable and
fluctuates, coming and going at almost rhythmical intervals.) They become more
definite when the attention is directed toward them. Sometimes there is a 'sharp-hot'
pain just below the sternum."
Observer X. — After 20 hours' fasting: "Hunger is such a 'total' experience that it is
difficult to pick out what should be labelled specifically 'hunger sensations.' In fact,
I doubt if there are any particular sensations that I should label specifically such. I
call them hunger sensations, I believe, because I have found regularly that, if I eat,
they disappear. Otherwise I should be inclined to label them as: a 'weakness' com-
plex, or a 'feeling of emptiness,' or 'throaty sensations,' according as one or another
aspect of the experience became prominent.
"What I feel at present is (i) a general bodily weakness, such that I am inclined
to do nothing in the way of work, not even stand, and (2) a particular kinesthetic com-
plex in the region of my diaphragm. Perhaps the latter is also part of the weakness;
respiration certainly is not so strong and deep as normally. There are also (3) sensa-
tions which I localize in my digestive tract. Some of these I localize in my stomach.
They make up the core of the empty, 'gone' feeling. They appear to be vague, dull,
quite persistent kinesthetic sensations, as if from contractions of the stomach. They
make a definite peculiar complex, but I believe that it is a complex composed of pres-
sures and kinesthesis, — dull, but definite; very strong and insistent at times, but never
sharp or bright or clean-cut. Besides the sensations in the stomach, there are others,
localized in the back of my throat, at the beginning of my esophagus, which seem like
incipient swallowing movements (esophageal kinesthesis). They draw my attention
very decidedly to that region; when I attend to them there is a veiy strong inclination
(in visual and kinesthetic terms) to put food in my mouth. The thought of food seems
to make the salivary glands more active, so that I occasionally actually swallow. The
upper part of my digestive tract seems very ready to react, and is evidently incipiently
active, — judging from the sensations. The feeling of 'wanting something' is localized
in my throat. The throat sensations are quite steady and persistent."
Observer Y. — "Strong, gnawing pressure (gnawing is elemental, a kind of pain).
This complex is unpleasant and means a need of food, — hunger. There is also a com-
3H E- G. BORING
plex localized in the pharynx, and muscular sensations in the jaws. Also a perception
of wetness, localized under the tongue, meaning much saliva. The complexes localized
in the stomach and pharynx together make up the desire for food. The pharyngeal
complex is the less prominent."
Observer Z. — This observer reports that she seldom has the intense and vivid
' hunger of childhood ' and that she has been unable to induce it for the sake of the experi-
ment. In addition to describing general faintness, however, she gives, after a short
fast, and under the caption 'Emptiness,' the following report. "Slight headache
and light feeling. No desire for food as food, but knew from experience that it would
take away the 'empty' feeling. Experience a slightly unpleasant pressure, localized
in bottom of stomach; pressure in throat from about larynx to top of back of mouth;
pressure of tongue on roof and sides of mouth (I think this is thirst); tongue felt bigger
and softer than usual. The pressure on the bottom of the stomach seemed just like
the pressure of a heavy weight on a relaxed muscle, although it was not so intense."
The writer has described elsewhere a case in which hunger
was induced in one subject by the introduction of HC1 into
the stomach.1 A few sentences will suffice to show that this
' laboratory hunger' does not differ from that occurring under
more usual conditions.
Observer B. After 5 c.c. of 5 per cent. HC1: "Hunger, a strong, intense, diffuse
ache, varying in intensity and covering an area as shown [i. e., an area extending from
the umbilicus to the sternum]. There is an especially intense and achy spot at [a point
a little above the umbilicus and to the left]. Hunger goes and then returns; then lasts
a long time, getting gradually fainter."
In another trial, a warmth was reported to "die away very slowly, fusing into a
general ache in the stomach region. This ache gets more intense and presently without
qualitative change turns into hunger."
Again, a 'stinging pain' is supplemented by a 'dull ache' below the sternum.
"The ache spreads downwards, and, as the sting disappears, it turns into hunger pains.
The hunger pains are marked, but are shot through by a little stinginess, a brighter
and more tingling pain."
The hunger-complex is a complex of pressure and pain.
Upon a background of dull pressure (A, B, C, F, X, Y),
which is sometimes recognized definitely as kinesthesis or the
equivalent muscular pressure, there is set a dull ache or gnaw-
ing pain which characterizes the hunger (A, B, D, E, F, Y\
the intense muscular pressure of C is also pain; muscular
pressure goes over into ache-like pain, which observers often
call pressure).2 Two observers (X, Z), who failed to find the
pain-quality, also had difficulty in determining just what
1 Op. cit., 48.
2 Cf. the confusion of pressure, strain, cramp, and pain in the introspections of
esophageal pressure, especially those of G, op. cit., 28 ff.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3 r 5
constituted hunger. Both pain and pressure are referred
to the region of the stomach. The pain is noted as fluctu-
ating, as rhythmical, as instable. In more intense hunger
the maxima of the fluctuations of the 'dull ache' may involve
a sharper pain-quality, which is definitely localized and limited
to a very small area (B, D, F). 'Emptiness' appears to
consist of the typical pressure pattern of hunger without the
algesic components (A, Z). Three observers (Jf, Y, Z1)
describe a complex kinesthesis in the throat and of oral
sensations arising from the free flow of saliva, a complex
which means for them desire for food, appetite, a literal
watering of the mouth. Here we have the true sensory
basis of 'appetite.' The ideation of food (mentioned speci-
fically by X) is no doubt a usual concomitant, and presumably
it often constitutes a desire for food that lacks sensory com-
ponents entirely. There can be no question that this desire
for food — appetite, if one is not disposed to limit the term too
closely — may also often be unconsciously carried, just as in
thirst the 'appetite' for water may become manifested auto-
matically in the movements of going for a drink (see pp. 307,
Our reports enable us to supplement Cannon's description by many reports of
psychologically trained observers. But we have gone farther than confirmation.
Hunger is a twofold experience. It is pressure in its weak form, pain and pressure
when intense. If one calls the intense experience 'hunger' and the weak 'emptiness,'
one has changed the phraseology but not the fact. Moreover, weak hunger appears
to be muscular pressure, and intense hunger is the ache of intense muscular pressure.
Carlson, we have seen, has also made this point: 'hunger contains elements of kines-
thetic sensation as well as pain'2; but he thinks that the pain is not 'the pain experienced
from contractures or 'cramps' in skeletal muscles.'3 However, Carlson admits that
the difference may be extrinsic rather than intrinsic, and he will doubtless welcome
evidence for the muscular quality of the entire hunger-pattern.
It is not easy to follow Carlson in his discussion of appetite.4 He objects to the
view that 'appetite requires a nervous organization capable of associative memory,'
because we have "in appetite for food conditions as primitive and essentially fixed by
1 The three woman: but it would be overhasty to discover a sex difference. They
found in general more difficulty in deciding just what hunger was — perhaps, after all,
their sex is a little less intimate with the inner man — and they gave fuller descriptions.
Unable to find a sine qua non, they described all possible concomitants. Of course
they thus noted complexes which the men overlooked.
2 Op. cit., 190.
189.
316 E. G. BORING
inheritance as in the case of the sexual desires or 'instincts'." Appetite becomes 'the
desire for food,' 'the expression of an inherited mechanism.' "The inheritance factor
in appetite, the desire to eat, is in some way caused by the hunger pains." When
appetite apparently occurs alone, it is due to a concomitant 'subconscious hunger.'
Three factors make up the food-taking impulse: hunger (pain), appetite (desire for
food; a sensation?), and 'memories of the taste and smell of foods.' There would ap-
appear to be seven possible cases: (i) hunger alone (pain), when one attends to the
hunger sensations; (2) hunger and desire for food (pain -f- appetite), when in extreme
hunger one eats disgusting food; (3) desire alone (subconscious hunger pains+appetite),
the non-rhythmical 'hungry-feeling' of Carlson's subject; (4, 5, 6) any of these states
together with 'memories' of food — the usual impulses in food-taking; (7) 'memories'
alone, 'the contemplation of a favorite dish after a full and satisfying meal.' So much
we are told. But we have no hint as to the nature of appetite. Is it a sensation? is
it a group of sensations? if so, what is it like? We do not ask for technical psychology.
If it is sensory, what makes it so fundamental that it must be reflexly aroused? And
why, in particular, must hunger, already defined as 'different degrees of pain,' be its
sole condition? And what is this conditioning hunger, conscious and 'subconscious'?
Is it sensation? or is it the nervous substrate of sensation? or is it the physiological
cause of the hunger pangs? Until these questions are answered Carlson's distinctions
will prove of little service.
If, however, classification and definition in such a simple case are wanted, the
writer would suggest the following schema:
Hunger-complex
Desire for food
Hunger = muscular pain
Emptiness = muscular pressure
Appetite = throat-mouth sensations
Imaginal desire = imagery
Unconscious desire = determining tendency
Sensory
Imaginal
Neural
It seems probable that such an account would be accepted by Meumann,1 who has
laid stress upon the variety of the digestive sensations. Meumann describes three
typical digestive experiences. In the first place there is the 'hunger sensation,' which
'is localized not only in the mouth and in parts of the throat, but also quite definitely
in the stomach,' z a complex equivalent to hunger plus appetite as given above. In the
second place, there is 'the very characteristic sensation of emptiness of the stomach/
a sensation which "has a very different character from the tension or pressure sensation
of the abdominal wall. It can become intensified in a very unpleasant way and is
sometimes connected with a vague perception of the peristalsis of the stomach."*
just such a complex we have also called 'emptiness.' Finally, after eating, there is 'a
characteristic sensation of fullness and pressure in the stomach' or, as it is called in
another place, 'satisfaction and fullness.'4 The experience is said to be not one of
mere extension, as it is partially independent of the amount of food taken, and to some
extent dependent upon the kind of food. Our introspections do not cover this point.
Hertz has described the experience and concluded that it is conditioned upon the
1 Meumann, E., 'Zur Frage der Sensibilitat der inneren Organe,' Arch. /. d. ges.
PsychoL, 9, 1907, 28ff.; 'Weiteres zur Frage der Sensibilitat der inneren Organe und
der Bedeutung der Organempfindungen,' ibid., 14, 1909, 279 ff.
z Arch., 9, 52.
3 Arch., 14, 293.
4 Arch., 9, 51 f.; 14,295. The writer does not interpret Meumann as meaning to
distinguish between 'fullness' and 'satisfaction,' although it is possible to make such a
construction.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3 1 7
tension of the muscular coat of the stomach, a tension of which the effectiveness is
independent of muscular tonus.1 Sternberg, it should be noted, distinguishes between
appetite, hunger, and satisfaction.2
NAUSEA
Nausea was experimentally induced in the laboratory by
the administration of syrup of ipecac or by a decoction of
tobacco and, in one case, by the smell of castor oil.
Numbers below in brackets indicate the time in minutes that has elapsed since the
beginning of the experiment.
Observer A. — 2 teaspoonfuls syrup of ipecac. [23] "Breathing sensations queer.
Feeling like that of respiration in abdomen, but shorter and quicker than breathing.
I feel as if I were 'ready to vomit,' which is a meaning. My stomach feels a bit fuller.
The tone of my muscles in arms, chest, and abdomen seems to have gone down; I feel
weaker." Q. "Is this nausea?" A. (after thinking): "Yes . . . Saliva is forming.
Tendency to open my mouth. Sweat comes on. Contractions of stomach, almost
painful. Breathing is irregular. I close my eyes. [He vomits five times.] Big
muscular wrench. Characteristic muscular weakness. Weeping. Throat and
stomach feel full. Achy pains across stomach. I think I feel most nauseated just be-
fore I vomit. It feels as if my stomach actually sank. To the best of my knowledge
that is like muscular pressure. There was a very, very slight dizziness in my head."
Observer D. — 4 teaspoonfuls syrup of ipecac, [i] "Begin to feel something. A
sort of sharp ache under sternum." [3] "Dull, heavy ache around stomach region."
[15] "Something runs the whole length of my esophagus, up to the back of my throat;
it means a desire to vomit. ... It is taking on a nauseous character. I feel it mostly
in the back of my throat; it seems to spread all over from the stomach up; considerable
pressure to it. [He vomits violently.] Felt just as if there was a pressure there at
bottom of esophagus up to throat. . . . Now I am getting unsettled in stomach. This
seems to be a diffused pain, localized at least two inches below sternum. With it there
is pressure, which means nausea and which gets more intense, meaning impending
vomiting." [21 : vomits again.] "Just preceding the vomiting the pressure gets very
intense. It seems as if that pressure forced the contents right out. While vomiting
I felt violent contractions in my stomach. . . . Nausea is coming again now. The
first thing is pain. It is now at the top of my stomach [indicates level of sternum]
and now lower down. There is also a band of pressure, below the sternum. Rather
suggestive of gripes." Q. "What sort of pain is this pain that comes?" A. "Diffuse,
sort of dull. It gets very intense. Not the pain of a prick at all. It seems as if the
pain and the pressure constitute nausea. The pressure alone means incipient vomiting.
... In some ways the pain is more of an ache than a pain; I suppose an ache is a dull
pain. It is quite diffuse." [40: vomits.] "Frightful unpleasantness seems to cover
up everything. There is a bodily trembling — a general feeling of weakness."
1 Hertz, A. F., 'The Sensibility of the Alimentary Canal,' 1911, 19 ff.
2W. Sternberg's classification is into (i) disgust, (2) appetite, (3) hunger, (4)
thirst, (5) feeling of satisfaction; 'Der Hunger,' Zentralbl. /. Physiol., 23, 1909, 105. For
his distinction between hunger and appetite, see in particular: 'Physiologische Psychol-
ogic des Appetits,' Zeitschr.f. SinnesphysioL, 44, 1910, 254: 'Das Appetitproblem in der
Physiologic und in der Psychologic,' Zeitschr.f. PsychoL, 59, 1911, 91.
318 E. G. BORING
Observer Y. — 5 teaspoonfuls syrup of ipecac. [15] "Dizziness. Pressure sensa-
tions in stomach, — a dull pressure, unpleasant, slightly nauseating, a sort of gnawing,
a sickish character. Pressure gives a sinking feeling. Occasionally muscular sensa-
tions, as if I were about to vomit." [20: vomits.] "I'm not sick or squeamish; it was
muscular." [22] " I am sick now. A gnawing, sinking, pressure-like quality in stomach
region, extending up a little under sternum. A trifle dizzy, but the stomach-complex
is strongest. Pressure-like quality seems to irradiate from the stomach, and I feel
generally squeamish. The feeling extends all over me." [24: vomits 5 times.]
"Nausea got very strong before vomiting. Muscular sensations are part of it. They
were fused with a 'sinking/ which got worse and, as it got worse, gnawing." [25:
vomits twice.] "I do not know whether the sinking, gnawing quality is something new
in the element-line or not, or whether it is the character of the total complex. It
belongs to the pressure family. It may be that there is a dull pain or ache, although
it is not what I usually mean by an ache. It might be an approach to a fused non-
intensive ache. Certainly its identity as such, if such it is, is so merged in pressure,
that it appears more like a coloring, a dull, gnawing, sinking affair. I am rather
inclined to think that the 'gnawiness' is in part of the achy character, but that pressure
is the clear and stronger component. It is all closely fused — a unique whole." [35]
"The gnawing is more prominent than the sinking. Yes, it is something aching; I am
quite sure now. It is dull pain, very different from pin-prick, and yet something of
the same order, — at least, a pain. . . . The achy character is more prominent than
cbefore — a gnawing."
Observer Z. — 6 teaspoonfuls syrup of ipecac. [36: vomits several times.] "Feel
awfully funny in throat; the muscles feel all tight, and yet the throat feels as if it were
bigger than usual. I think that, except when I felt the contents coming up, I did not
have any sensations at all below the bottom of my neck." Q. "Would you say you
were nauseated?" A. "No, I do not think I was. Generally, as well as I can re-
member, nausea is decidedly unpleasant in both stomach and throat." Z failed to
get any nausea within an hour and went home. There she became quite sick (i. e.y
vomited) during the night, and recorded at one time the following: "Vague moving
pressures, localized in stomach. Slight dizziness and weakness. Spasmodic contrac-
tion of muscles over whole trunk, especially in throat. Tears; perspiration." She did
not record at the time whether or not the complex was one of nausea, although in the
morning, upon being questioned, she was inclined to think that it was.
In order to obtain a nausea which would be more persistent
than that induced by the ipecac, two observers took doses of a
strong solution of tobacco juice.
Observer B. — 2 teaspoonfuls of tobacco juice. [5] "Esophageal sensations, weak,
but qualitatively like those in swallowing a hard object. Faint pressure in stomach, —
a very vague ache." [15] "Gentle achy pressure in stomach region. Also vague aches
from arms, like muscular fatigue. Whole thing makes up 'sick feeling.' Attention
to any one part seems to break it up." [19] "When I smell the tobacco juice, a wave
of achy pressure travels down the esophagus. It is nauseous." [24] "It seems as if
nausea were in this case: general bodily weakness (mostly muscular fatigue) -(-headache
(swimming sensations, eye-pressure-aches, tightness at ear, pressure wave at back of
head; eye-aches and swimming most promment-\- intense unclear pressure-aches around
sternum. It seems as if stomach sensations had to be unclear in nausea; attention
to them spoils the complex as nausea." [37] "Incipient vomiting sensations, in
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3 r9
which stomach-aches get more intense and extend up farther and aches in eyes get
intense. This u nausea. I also sometimes get stomach-pressures, which mean in-
cipient vomiting, but which are less achy." [80] "The nausea is particularly difficult
to localize; it is fleeting, evanescent, by which I think I mean merely that it (or at least
the achy complex) is intermittent. Attention always goes naturally to the head
sensations. When I voluntarily attend to the stomach-sensation, it always turns
out to be an ache; and, since it seems to remain as continuous as any process in changing
from obscure to clear, I say that achiness is its normal character. But attention
never goes voluntarily to these sensations. I say they are essential because I always
find them when I hunt for them while I am feeling nauseated. The principal part of
the nausea as regards intensity and clearness is the headache, swimming, and (just
now) jaw-aches (crampy character); but I do not think that they could be nausea
without being supplemented by the stomach sensations, i. e., I think the stomach-com-
plex gives the meaning nausea."
Observer E. — 2 teaspoonfuls of tobacco juice, 50 per cent, dilution. [7] "I begin
to feel quite sick now, — a sort of dizziness in head; also pressure in stomach. Feeling
of great discomfort all the way up from stomach to throat. More intense at times.
Also a little pain, a peculiar ache, a sort of dry achy tenseness. It is very intense every
time I smell the tobacco. Also a dull ache in head." [n] "Convulsive movements in
the stomach or esophagus." Q. "Would you say you were nauseated?" A. "Yes."
[20] "Still a discomfort in stomach. It is almost pain." [165] "Feel sick in my throat.
There are aches in my stomach."
Of all the observers, A had the greatest difficulty in char-
acterizing nausea. The nausea described above (ipecac) for
him was not intense nor was it probably typical. The strong,
characteristic experience could invariably be induced, how-
ever, by the smell of castor oil. His introspections under these
conditions follow.
Observer A. — Nausea induced by smell of castor oil. "With the smell of the oil
there was a big shiver, together with a wave of cold, all over my upper chest and
abdomen — even in my arms. There was a sensation which has something in it of the
sinking sensation that you get down here [umbilicus to sternuml when you drop in an
elevator. There is also a start of a vomiting reflex, a muscular, pressury sensation; it
seems as if I could feel a contraction. There is a bit of dizziness in head also. I feel
myself sweating a bit. . . . Both what is the start of a vomit and the sinking thing are,
I think, pressury. . . . The pressure in the stomach region is a pressure down." After
another trial: "I do believe that there is something that I haven't mentioned yet, a
sensation which forms a part of this whole situation. It is very hard to localize; cer-
tainly, however, somewhere in the trunk. I can describe it only as a sickening sensa-
tion, a kind of an awfulness and helplessness. It is not intense, if you can talk about
its intensity absolutely. The other things were definite and stood there and waited
for you; they caught your attention. This sensation is there beside, — a sickening, an
awful helplessness." At another time A assumed the attitude which stands for this
'awful helplessness:' the body is relaxed, the knees slightly flexed and the arms hanging
limp, the body bends slightly forward, the abdomen in, the head is inclined, the mouth
is open, the eyes are closed. After another experiment: "I don't know! If it is a mean-
ing, it's the meaning of the smell. And how can the meaning of the smell be down in
320 E. G. BORING
my chest and abdomen — for that is where I feel sick? This is the case: I do feel sick.
And I feel sick down here. Now there are a lot of sensations down here that I can put
my finger on and localize."
The confusion of the ache of nausea with the ache of
hunger came out in the experiments that were made upon B
with stimulation by HC1 (cf. p. 314).
Observer B. — After the introduction of 5 c.c. of 20 per cent. HC1 into the stomach.
"Ache in stomach region became definite and was recognized as nausea. It was most
intense about 3 cm. below sternum. The nausea lasted a long time (I still feel a little
bit sick.) It is a pain, very much like hunger." After 5 c.c. of 5 per cent. HC1:
"Nausea. I feel pretty sure that 'sickness' is the nausea ache below the sternum plus
muscular pressure down toward the umbilicus, the latter meaning violent contractions,
as if I were going to vomit. I do not believe that the ache is in any way different from
the ache of hunger, except that it is a little more diffuse, a little higher up, less likely
to be localized and less definitely localized when it is, less intense, and without the
rhythmical intensive fluctuations of hunger."
Observer D volunteered a general statement that is relevant: "I can not always tell
hunger from nausea. When I am nauseated [he is subject to spells of indigestion] I
generally stop eating. At such times I decide to begin with my meals again as soon as
I feel hungry, but I can not always tell when to start in, because I can not always tell
whether I am hungry or still nauseated."
The reports show that the experience of nausea is very
complex indeed. All sorts of factors are mentioned: dizziness
or swimming sensations in the head, the sensations aroused
by too free a perspiration, aches and pressure-pain complexes
in the head, in the eyes, in the jaws, in the arms, general bodily
shivers and chills, general weakness. Besides these factors,
which sometimes constitute the most prominent part of
nausea, there are the sensations which are referred to the ali-
mentary tract proper. Pressure-complexes referred to the
stomach, or pressure-waves localized in the esophagus,
indicate incipient vomiting and are often present.1 Other
alimentary pressure-sensations, however, appear to be more
nearly integral to nausea: the 'sinking feeling' and the dull
'sickishness' are described as purely pressure, the ' gnawing
pressure' and even the 'ache' are probably partly pressure.
With the exception of Z, in whom the occurrence of a true
nausea is open to doubt, and of A in his series with castor-oil,
1 It is this complex, apparently, that E. Murray describes as 'revulsion . . .
sometimes grading into a feeling of nausea' ('Organic Sensation,' Amer. Jour. PsychoL,
20, 1909, 437). To Murray belongs the credit of having obtained introspections upon
nausea under experimental conditions. Unpleasant odors were used as stimuli.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS Z21
all the observers agree that nausea involves a dull ache or
pain in the stomachic region. Two observers (B, D) declare
that this ache is indistinguishable in quality from the pain of
hunger, and one (Y) implies the similarity by describing it at
first as a 'gnawing pressure' and coming later to the con-
clusion that it was an 'ache.'
Which of these factors are essential to nausea, and which
are occasional concomitants? Dizziness, headache, bodily
weakness, shivers, perspiration, and so forth are by no means
invariably present. No one is essential, and all may be
absent. They occur more frequently as concomitants, or
perhaps indicators, of an intense nausea. The pressures of
the vomiting reflex are often absent, and were distinguished
by most of the observers as separate from nausea. Their
concurrence seems to be only casual; they have not come to
mean nausea. Frequently they occur without nausea in
vomiting.1 The pressures of the 'sinking sensation' and the
dull ache seem to be the most constant components; but the
pressures are lacking in the hunger-like nausea that B reports
for stimulation by HC1, and the pain is not found by A in the
intense nausea induced by castor oil. Apparently, then,
there is no sensory factor that is invariably present in nausea.
The facts become intelligible if we regard nausea as a meaning,
a situation. Various organic factors, alimentary or general,
may combine in nausea, or (at least after the more complex
experience has been had) a few or even one of the more usual
constituents may mean the whole situation. Nausea for
different persons, or for the same person at different times,
may thus be very different. The significance of the nauseous
situation is such that one is not likely to adopt even a casually
introspective attitude toward it; hence, even if there are varia-
tions from time to time in the same person, the qualitative diff-
erences are still not likely to be noticed. Nausea involves a
condition of the digestive tract, and undoubtedly the ali-
mentary pattern of pressure and ache is the usual result of
1 In the present experiment both A and Z vomited without nausea. In the
experiments previously reported by the writer (op. cit.) nausea was as infrequent as
vomiting was common. Observer F in that experiment was almost never nauseated;
but see the account of vomiting on pp. 6 and 12.
322 E. G. BORING
the conditions which produce nausea. In a hypoalgesic
individual, nausea may never reach the ache-stage. By
association with the pressure-ache complex, or with what
may by a given person be regarded as symptoms of nausea
(e. g., vomiting, loss of appetite, disgust, weakness, etc.),
other sensations may come to stand for the nausea or, in a
given case, to constitute the habitual form of nausea.
The foregoing hypothesis not only explains the apparent
uniqueness of the nausea experience (a factitious uniqueness,
acquired by the failure of the observers to observe sensory
quality), but also avoids the implication that a unique ex-
perience must have a unique qualitative basis. The positive
reduction of nausea to organic pressure and pain is not,
however, an easy task. The whole pattern is so complex,
the unity of the situation is so insistent, the fusion of the ele-
ments appears in consequence so intimate, that analysis is
difficult. B observed that attention to some parts seemed to
destroy the whole, and Y declared that the complex as a
whole, intimately fused, constituted nausea. A, as a matter
of fact, was not always able to make the analysis. He could
find nothing but organic sensation of pressure-like quality,
but he was not sure that there was not something else, 'a
sickening sensation, a kind of an awfulness and helplessness.'
When asked to describe, he was unable to do more than to
assume the 'helpless attitude' (see p. 319). In view of the
fact that all the other observers made the analysis, the ina-
bility of A to reduce the complex does not seem to warrant
the assumption that it involved a new element. In a more
general context, Titchener has remarked that the impossi-
bility of reduction in a single case need not imply elementari-
ness if the analysis can be made by other individuals or in
other instances;1 and it appears as if we had here chanced
upon an attitude so intimately fused that its reduction was,
under our conditions, not always attainable. It should be
remembered, too, that ^himself was not sure that the unique
residuum was not the ' meaning of the smell.'
1 Titchener, E. B., 'Experimental Psychology of the Thought Processes,' 1909, 171.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 323
THE CALL TO DEFECATION
Descriptions of the call to defecation and of the act itself
were written down by the observers at the time of occurrence.
Observer A. — "Slight strains and dull pressures of weak intensity in abdomen are
all that I can find as sensations."
Observer C. — "The call to defecation is a wave of pressure of quite general dis-
tribution and not easily localized. It is somewhere in the lower abdominal cavity and
finally reaches the rectum. When this wave has run its course there sets in a general
pressure which includes the whole abdominal contents. At high intensities the mus-
cular contractions of the abdominal wall and of the sphincter are added to the former
complex. These sensations are just those of normal contracting muscle."
Observer D. — "Insistent pressures in rectum. Faint but slightly achy pressure in
front wall of abdomen. Slight achy pressure in temples, which feels as if blood-vessels
were distended."
Observer E. — "The call to defecation is a very pleasant experience. It seems to
consist of a feeling of fullness, of distension of the bowels. Pressure, which is most
prominent, is at first rather indefinite and equal in all directions, but later becomes a
downward one. Sometimes aches and pains in the intestine accompany it."
Observer F. — "At first, vague, diffuse pressure in lower abdomen, not particularly
unpleasant. Soon, however, it becomes a dull ache and often sharply painful. There
also comes an ache about the anus, which often becomes 'hot' and 'burny '; and this is
accompanied by violent contractions of the sphincter muscle."
Observer X. — " First noted impulse to go to toilet. The observed vague sensations
localized in region of large intestine, a vague perception of movement in that part of
the intestinal tract, like a very remote dull pressure that changed its location. Also,
sensations of incipient movement in anus; a somewhat rhythmic movement of disten-
sion abruptly checked each time by movements of contraction. All this was quite
involuntary. The internal movement (intermittent) renewed itself more intensely, but
just as vaguely; the localization of it was far from definite. The relaxation-phase of
the anal movement tended to increase in duration and intensity. After a time the
checking, contracting movement ceased to occur involuntarily. Now there was a
definite sensation complex, meaning pressure against the anal opening from within and
above."
Observer Y. — "Intense deep pressure, probably muscular, located in region of
anus. Soreness (of the pain modality) fused with the pressure. The complex was
unpleasant."
Observer Z. — "Dull, diffuse, rather heavy pressure, localized rather vaguely in
lower abdomen. Neither pleasant nor unpleasant."
In experiments described elsewhere, the writer has shown
that the call to defecation may be induced in all degrees of
intensity by the inflation of a rubber bladder within the
rectum; that small amounts of warm water (50 c.c., 50—70° C.)
produced the call very intensely (although an equal amount of
cold water at o° C. did not) ; and that HC1 (10 c.c., 5 per cent.)
324 E. G. BORING
may also constitute an adequate stimulus to the call.1 The
description already printed calls attention to the specific
sensation of muscular pressure in the rectum (a muscular
'ache' in intense degrees); to the widespread general ab-
dominal response, apparently secondary and dependent upon
muscular contraction; and to the pains, which, when the call
is most intense, occur in great variety of quality and reference.
A few quotations will render the reference explicit:
Observer B. — After the inflation of a bladder in the rectum 10 cm. above the anus:
" First pressure in rectum. Then call to defecation, which differs from the first pressure
in that it is more intense, covers a larger area, and has a temporal course of varying
intensity (pulsations). Later pain was introduced. From then on intensity increased
by jumps. The increase of pain was the most noticeable. The pain was of the achy
variety, but got sharper and brighter, more definite and lively, although always diffuse,
as the intensity increased. I think the pressure also increased in intensity, although it
was largely obscured by the pain. Vague general pressures in abdomen were also
noticeable. Still later pain was very intense; it was really very sharp and tended to
run off into shoots and stings of pain. The temporal course up to this point shows, I
think, all degrees of urgency for defecation. The urgency is not only a matter of
intensity, but varies with the area affected and also with the quality of pain. . . . On
the release of the pressure there is a tremendous relief, very definite in the region of
the umbilicus and in the rectum. It is exactly like the relief of defecation without
the sensations of passage. It is kinesthetic pressure."
After the introduction of warm water into the rectum: "Violent call to defecation
includes pains and partially initiated movements at rectum. More or less confined to
rectal region. General muscular effort in resisting call, even to the circulatory warmth
of the face. Call is predominantly a pressury ache overshot with more intense thick
pains of the achy variety. Besides this there is a general muscular irradiation." In
another trial: " Intense grippy pains about umbilicus. They also shoot down intotestes
and penis. The abdominal pain has the peculiar character of 'belly ache.' There are
also pains in the rectal region, such as occur when it is hard to hold in with an urgent
call." At another time 'shivers over whole body* are noted.
After the introduction of dilute HC1: "Sets up the pressure complexes of the call
to defecation in rectum at once; very intense. There is in the call also a dull ache; but
I should not say that its unpleasantness is dependent upon the intensity of the pain."
The call to defecation is predominantly abdominal pres-
sure. The pressure may mean distension or contraction or
movement or effort; it may be weak or intense; it may be
dull, diffuse, and vaguely localized or it may be clear-cut and
accompanied by a definite visual reference. Of the nine
1 Op. cit., 5°~54- Hertz's observations, op. cit., 28 ff., have shown that the call
produced by inflation of a bladder arises in the lower rectum. His descriptions do
not indicate, however, the widespread nature of the response nor its painful character
in high degrees.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS Z25
observers five (#, Z), E, F, Y) find that the experience in-
volves aches or pains; two of these (Z?, F) also find 'sharp
pains/ Besides the general abdominal complexes, definite
rectal pressures are mentioned by Z?, Z), X, and Y. B notes
that the rectal pressures involve the ache of extreme muscular
pressure. Two observers (B, Y) mention that the call is
unpleasant, two (E, Z) that it is indifferent, one (D) that it is
pleasant.
IPs fuller introspections for the intense call indicate that
the course for increasing intensity is more or less as follows:
(i) muscular pressure in rectum; (2) rectal pressure becomes
intense and achy, general abdominal pressures develop;
(3) dull pain introduced; (4) sharp, piercing pains, of uncertain
and varying reference, appear. There is no indication of the
presence of qualities not ordinarily included in the pressure-
pain group.
DEFECATION
Reports relating to the experience of defecation itself are
as follows:
Observer A. — "Dull pressures in abdomen; increased strain in" abdominal muscles.
Pressure in rectum, localized two or three inches above anus. Sensations of cutaneous
pressure and of strain as feces pass. Slight sweat."
Observer C. — "At the act of defecation [following the call to defecation] there is no
new quality in the abdomen, unless the diminishing pressure due to decreased volume
might be considered as such. In the rectum and at the anus, however, the moving
contents may be felt as waves of pressure. . . . The only quality that I can find is just
pressure, except at times pain, which does not seem to be normal."
Observer D. — "Increase of pressure throughout abdomen and especially in rectum.
Bright stingy pressure at anus, which is pleasant and which carries the meaning of
expulsion."
Observer E. — "Defecation itself is pleasant. In it a feeling of relaxation and a
pressure, bearing down, are mixed."
Observer F. — "First the sensations of relaxation of the sphincter and abdominal
muscles; then those of the moving pressure at expulsion."
Observer X. — "Movement of issuing feces somewhat perceived somewhere in the
large intestine (as vaguely dull, indefinite, moving pressure, rather rhythmic in its
fluctuations of intensity), but chiefly precisely at the anal opening. Here there was a
definite clear-cut experience: very strong, bright, 'moving' contact and large pressure
sensations; occasional pain elements, sharp and bright. Slightly shivery sensations
ran up spine, and to some extent seemed to well out from the anal region. Felt 'goose-
fleshy.' " At another time: "Noted that preliminary internal sensations were stronger,
more definite and steadier. They included dull, diffuse, deep pressure with a vague,
weak subcurrent of dull pain. Occasionally there were sharp, knife-like streaks of
pain."
326 E. G. BORING
Observer Y. — "Defecation-complex is made up of the following factors: muscular
sensations in anus; soreness, which was more distinctly painful in character, and which
varied in intensity during defecation; and, I think, another kind of pressure, located
at the distal end of the anus and meaning contact of waste-products with that part.
There was also present a feeling of general strain."
Observer Z. — "Contraction of many muscles of abdomen with resulting strain
sensations. More intense muscular sensations localized in rectum; pressure localized
vaguely in the same place. During and after the contraction of the muscles, a pain of
a moving pressure."
As defecation follows the call, the abdominal contractions
that induce it are sensed as strain (A, Y, Z) or as pressure
(D, X). The dull rectal pressure increases in denniteness
and intensity and ordinarily becomes painful (C, -D, X, Y, Z).
This pain may be a dull ache or soreness (X, Y) or it may be
bright, stingy, sharp, and knife-like (D, X, Z). D and E
describe the experience as pleasant, D specifying that it is
the ' stingy pressure' which is pleasant. This introspection
agrees with that of the writer. In general, it may be said
that the experience of defecation is little more than a height-
ening of the call to defecation, with a consequent introduction
of algesic elements, and with perceptual additions relating
to the passage of the feces.
THE CALL TO URINATION
The descriptions of urination and of the call to urination
were obtained in the same manner as those of defecation.
Observer A. — "Weak sensations, very much like muscular pressure, spread over an
area high inside of body with base just above the pubic bone. Later these sensations
became slightly stronger in intensity and seemed more strainy in quality, although they
were still of weak absolute intensity. In the penis, especially the lower part (urethra),
there were very weak sensations like contact and very weak cutaneous pain combined.
Also a cool sensation in the glans near the opening. Still later I noticed weak strainy
sensations referred to the penis throughout its length and to the body. The sensations
were strongest at the opening of the urethra."
Observer B. — "Dull pressure-ache, like that of intense muscular pressure, referred
generally to region of penis, scrotum, and pubes. Localization not specific, for complex
appears big and round, and attention to any particular organ makes the sensation appear
to go elsewhere although it remains in the same general region. The penis is, however,
always involved; the dull ache is most intense there. Dull pressure without ache centers
in the pubic region. Besides the dull ache there is a sharper ache referred to the penis
in a region about half-way between the root and the glans. It is very much like a
'sting' that is spread out. It varies quite regularly in intensity, intense pulses being
separated from weak ones or from periods in which there is no ache at all. [I have had
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 3 27
an assistant note the time of these fluctuations for 5 min. There were 30 maxima
in this time, an average separation of 10.1 + 3.6 sees. Ten times the pulsations died
out entirely. About once a minute there is a long interval, which serves to divide the
pulsations into groups.] Ordinarily, I think, attention fluctuates, returning to the call
when this sharp ache is most intense. When the call gets strong, there are muscular
twitches — pressure sensations — prominent. Also a general restlessness."
Observer C. — "The call in its initial stages is intermittent and lapses with the
application of attention. When it really becomes insistent it is very unpleasant. The
components seem to be principally strain sensations from muscles and sensations of
warmth. The first strain seems to be from the contractions of the bladder, at least
it is in the lower abdominal cavity. It is a wave, moving from above downwards. At
once this wave is met by a wall of pressure, and the two opposing strains seem to see-
saw. The essential thing is strain, I think the resisting strain comes from the con-
traction of the sphincter muscles at the origin of the urethra. From this point lesser
waves of pressure occasionally pass outwards to the distal end of the urethra. In
addition to this pressure there is a sort of quality much like the pricking of a stiff hair
applied to a pain or pressure spot in the skin. The bladder and the sphincter strains
spread until the muscles of the abdominal wall are involved."
Observer D. — "The call to urination consists principally of bright stingy pressures
in penis, mostly near the base. There is also a slight diffuse pressure higher in ab-
domen, which feels like the pressure of a filled bladder."
Observer E. — "The call to urination is a very pleasant thing, provided it be not too
strong. It consists of a feeling of distension, of outward pressure."
Observer F. — "A vague indefinite pressure in the lower abdomen, rising as high as
the umbilicus. Intermittent pain in the glans penis. Whole thing uncomfortable."
Observer X. — "The unintense experience of normal life is as a rule scarcely con-
scious; there seems to be an automatic reaction before the sensations become at all
intense. . . . The experience, when intense, includes (i) a general bodily uneasiness,
especially in the lower portion of abdomen; (2) a definite pain-complex, localized slightly
below the middle of the abdominal cavity; it is an achy pain with something of the
strained feeling to it; it is insistent, definite, persistent; (3) a vague feeling of 'repletion'
of the abdominal cavity — a complex made up chiefly of dull pressures with perhaps a
slight pain-component; (4) intermittent sensations, localized at the opening of the
urethra. The 'general uneasiness' is centered upon this region, and the attention is
drawn strongly to very bright and lively kinesthetic sensations of incipient movement
there."
Observer Y. — "Warmth; and a sort of ticklish pressure, located in the region of
the urethra and the bladder. The 'ticklishness' belongs to the modality of pain.
This complex is set in a general muscular feeling."
Observer Z. — "Very slight warmth. Light diffuse pressure, spreading out through
a comparatively small space and very poorly localized in the lower pelvic region toward
the front of the body. Affective tone was indifferent."
All observers describe sensations of fullness or of pressure
or of strain in the region of the bladder. X and Y refer an
algesic quality to this region, and it is just possible that the
strain noted by A and C is incipiently algesic. The most
prominent part of the call for the male observers seems to
328 E. G. SORING
be, however, a pressure-pain complex, in which the pain
dominates, and which is referred to the penis (variously to
the base, the side in which the urethra lies, a point between
the base and the glans, and the glans). All the male ob-
servers (except E, whose introspection is too scanty to be
considered analytically) report pain. A finds 'weak cutan-
eous pain'; B, besides the 'dull pressure-ache,' a 'sharp,
pulsating, intermittent pain'; C, a 'pricking pain'; Z), a
'stinging pressure'; and F, an 'intermittent pain.' A^ B, C,
and D describe the pain as mingled with contact and strain,
with pressure and muscular sensations, or with pressure.
The reports of the women are quite similar. X notes an
ache in the region of the bladder, and Y a 'ticklish' pain.
X finds bright and lively kinesthesis at the urethra. For
Z the experience is quite colorless — merely a diffuse pressure
and warmth.
The affective judgments vary considerably. B finds the
pains pleasant, and E reports that the whole experience,
when weak, is very pleasant. Z records indifference, and
C and F unpleasantness — at least when the call is intense.
URINATION
The act of urination is described as follows:
Observer A. — "Voluntary relaxation of the strains referred to the urethral opening,
tactual sensations (like those from mucous membrane of mouth) and, at the very first,
very weak cutaneous pain sensations, diffused through these tactual sensations, or
spotted, peppered, around in them. These sensations are referred to the urethra,
almost for the entire length of the penis. The whole experience was pleasant."
Observer B. — "Sharp aches, like the intermittent ones in the call to urination,
become very intense just at the initiation of passage. They are referred to the same
region as before, i. e., above the glans, about one third of the way to the base. The
aches get weak as passage starts and continue so until the end. Then, as the last dribble
passes, they become momentarily intense again. After that they weaken and die out
slowly during the subsequent minute. There are dull muscular sensations in the region
of the bladder, which are not at all prominent. The relief afterwards is represented
by the persistent aches in the penis, as described, and a large diffuse ache of the same
quality in the region of the bladder. . . . All these sharp aches are very pleasant indeed.
Even when one is restraining urination with much effort, the pulsations of pain are
very pleasant. They are, I think, similar to, if not identical with, the sensations of the
sexual organs at a low degree."1
1 The completion of urination is physiologically similar to ejaculation. The last
portions of urine are expelled by rhythmical contractions of the bulbocavernosus muscle.
See Howell, op. cit., 785, 898.
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 329
Observer C. — "Strain in the muscles of expulsion but relaxation in the sphincter
muscles. In the urethra there is a warm pressure, which persists just for a moment
after the act is over. General relaxation marks the closing."
Observer D. — "At the beginning the pressure in the penis gives way to a tingling.
This changes to a complex which I have not been able to analyse, but which means
liquid flowing through the urethra. There is also general relief of pressure in the
bladder."
Observer E. — "Pleasant, but I have not been able to reduce the experience to
words."
Observer F. — "At first there is a sharp burning pain in the glans penis. Then
nothing but the pressures accompanying the flow of urine. At the end there is a
repetition of the beginning pain, only it is much stronger."
Observer X. — "Urination is accompanied by almost no sensations. There is a
lack of the strain sensations experienced in trying to hold back the reaction, — a general
bodily relaxation. There are very weak, bright, contact sensations at the opening.
The pain sensations [described in the call in the region of the bladder] do not change in
intensity or cease until some time after the operation is completed."
Observer Y. — "Muscular sensations in region of the urethra and bladder. Warm-
pressure perception (flowing of liquid) and auditory perception (liquid striking water).
This perceptual complex, set in a weak, general tension, present during urination.
Afterwards a general feeling of relaxation, slightly pleasant."
Observer Z. — "Almost sensationless. Very weak pressure sensations, moving in
scarcely perceptible waves, — just like the faintly discriminable changes in pressure
you get from floating when in swimming. With these pressures, very weak muscular
feeling of relaxed muscles all over and through the abdomen."
There has recently been reported to the writer a case of a woman who was unable
on a railroad train to tell whether she was urinating or not. The noise of the train so
obscured the usual auditory cues, she said, that she could not make the judgment.
No doubt the jar of the train prevented any faint organic sensations, which may have
been present, from being distinguished as cues.
Like the call to urination, in the men urination proper
involves principally a pressure-pain complex in the penis.
Some mention is made of muscular or strain sensations in
abdominal regions, but the characteristic experience seems
to be referred to the penis, and is caused, no doubt, by the
distension of the urethra. Distention here, as elsewhere,
may be expected to result in pain. The less experienced
observers had difficulty in making the analysis. The others
find in some cases tactual sensations, strains, or the less
definite ' pressure.' Pain is possibly universal. A finds
'cutaneous pain,' $, ' sharp aches'; F, ' sharp, burning pain.'
The 'tingling' of D implies pain, and, when one is familiar
with the usual stinging response of the penis to warm stimu-
lation, one is tempted to read an algesic quality into C's
'warm pressure.'
33° E> G. BORING
In the women the experience seems to be as indefinite and
colorless as it is striking in the men. X and Z call it 'almost
sensationless,' and they are supported by the case last cited.
Muscular sensations and weak contacts and pressures are
noted in the effort to find something to report. A true sex
difference seems to exist.
The experience is doubtless affectively indifferent for the
women. Of the men, three (A, B, E) mention the affective
aspect and all declare that it is definitely pleasant.
CONCLUSION
We may conclude that thirst, hunger, nausea, the call to
defecation, defecation, the call to urination, and urination
are all complex experiences reducible, under favorable con-
ditions, to various patterns of pressure and pain.
The experiences may be very complex and may vary from
individual to individual. Nausea is, perhaps, the extreme
example. Different processes may stand for it at different
times; and again, it may become so attitudinal as to defy
analysis. Hunger is reduced to a single pain only by isolating
it from appetite. Thirst is confined to the mouth and throat,
frequently to the mouth only, but is definitely of the per-
ceptual order. The excretory complexes are less frequently
recognized as specific, and show a correspondingly wide
variation in their many constituents.
The reduction to pressure and pain suggests the possibility
of a number of qualities within each of those modalities.
Besides occasional pressure of cutaneous quality, dull pressure
and muscular pressure were the most frequent forms reported.
As is usual, the muscular pressure in intense degrees runs off
into painful ache. There were also sharp pains which, it
may be, are of more than one kind.
Pain, although less usual than pressure, is by no means
uncommon. It may be present in the throat in thirst; hunger
is pain; the most constant constituent of nausea is the same
pain as that of hunger; the call to defecation, when intense,
involves sharp, shooting pains and dull aches; defecation
may include stinging sensations at the anus; urination and
PROCESSES REFERRED TO ALIMENTARY AND URINARY TRACTS 33 l
the call involve aches in the region of the bladder and, in
the male, bright pain in the penis. These pains may be
accompanied by any affective judgments whatever. The
intense bright pain of urination may be very pleasant, while
that in the call to defecation may be extremely unpleasant.
The present paper has been mainly analytical. The
problem that it attacks must remain unsolved until the com-
plementary synthetic operation has been performed. It is
one thing to reduce a large part of organic experience to
pressure and pain; it is another thing to say how many
pressures and how many pains there are, how they differ
from one another, and how they combine to form the typical
complex processes of organic life. This second phase of the
problem it is the writer's hope to bring into the laboratory.
VOL. XXII. No. 5 September, 1915
THE PSYCHOLOGICAL REVIEW
THE FATHER OF MODERN PSYCHOLOGY
BY PROFESSOR FOSTER WATSON
The * father' of modern psychology was, I suggest, Juan
Luis Vives. It may be objected that if we take l modern' in
a sufficiently literal sense, we must go back to Aristotle.
For that great philosopher traced the origin and development
of the pre-Socratic psychology, critically examined the
Platonic views, embodied and organized from his predecessors
whatever would serve as basis; and for the rest, supplied,
from his own researches and thought, an organic system of
psychology, which has held its own, and still claims close
study — from the fourth century B.C. to the present time —
continuously for over twenty-two centuries and a half. St.
Thomas Aquinas, fifteen hundred years after Aristotle,
re-affirmed the main points of Aristotelian psychology, but
supplemented them by a rationalized interpretation, in which
he, in many cases, anticipated modern psychological thought.1
Doubtless the line of continuity in psychological advance
is traceable, in this way, to Aristotle as the real founder of
the subject. But, our modern division of history, into
ancient, mediaeval and modern, requires us to consider new
starting-points, though, logically, in the light of the principle
of continuity, it is misleading to regard even these great
divisions as abrupt transitions. Although the Renascence
period of the fifteenth and sixteenth centuries brought a
steady concentration of attention upon psychological ques-
tions, as it did upon all humanistic problems, yet the main
1 The Rev. Prof. Michael Maher, S. J., has shown in his interesting 'Psychology —
Empirical and Rational' the parallels between St. Thomas Aquinas and modern psy-
chology.
333
334 FOSTER WATSON
setting of psychological theory had clearly been determined
by those great thinkers, Aristotle and St. Thomas Aquinas;
and any advance was most hopefully to be looked for, which
started from them as basis.
It is often stated that 'the father' of modern psychology
was Rene Descartes (1596-1650). Without attempting to
withdraw any credit from the actual accomplishments of
Descartes, in his day and generation, in the subject of psy-
chology, it would be extraordinary, on a priori grounds, if the
period of the Renascence (say from 1450 onwards) till the
time of the birth of Descartes (1596) had failed to produce a
conspicuous thinker on a subject so essentially humanistic,
as that of the Mind. Some writers, accordingly, place Francis
Bacon (1561-1626) as the pioneer of modern psychology.
Bacon was, certainly, the most influential advocate of the em-
pirical scientific method of the seventeenth century. Psy-
chology has made its great advance by the employment of
this method. Hence, it is urged, Bacon is the leader to whom
modern psychology traces its beginning. But neither Bacon
nor Descartes was the first Renascence writer to give his
attention to psychological theory, nor even to the advocacy
of the empirical inductive method. In a wide though valid
sense of the term, every man is a psychologist, and every man
employs the inductive method, and the 'fatherhood' of both
goes back to an antiquity, not merely as old as Aristotle, but
as old as man himself, at least. But the self-conscious em-
phasis on induction as a method of inquiry and discovery in
philosophical, and particularly in psychological questions,
must be taken back, even in Renascence times, beyond and
before Descartes and Bacon, at any rate, to Juan Luis Vives
(1492-1540).
Thus, Vives shows explicitly his insight into the signifi-
cance of the empirical inductive method in his account of the
origin of the arts. For the formulation of the arts, says Vives,
was due to observation joined with reasoning. "In the
beginning, first one, then another experience, through wonder
at its novelty, was noted down for use in life; from a number
of separate experiments the mind gathered a universal law,
THE FATHER OF MODERN PSYCHOLOGY 335
which, after support and confirmation by many experiments
was considered certain and established. Then this knowledge
was handed down to posterity. Others added subject-matter
which tended to the same use and end. This collection of
knowledge-material by men of great and distinguished intel-
lect, constituted the several branches of knowledge, or the
arts. . . . Whatever is in the arts was in nature first, just as
pearls are in shells, or gems in the sand."1 The insight which
Vives brought to bear on the place of the empirical, inductive
method, in the building up of the arts and sciences, was of the
first importance to him, when he came to study the subject
of psychology, theoretically.
In 1538, Vives published his book on psychology, entitled,
after the work of Aristotle: 'De Anima et Vita.' He ad-
dressed it to Francis, duke of Bejar, to whose descendant,
another duke of Bejar, Cervantes in 1605 dedicated 'El
Ingenioso Hidalgo Don Quijote de la Mancha.'
In this Preface to the 'De Anima,' Vives makes clear that
he writes on the mind, in no perfunctory or merely conven-
tional spirit, for he recognizes that the knowledge of the mind
has relation to matters 'of the greatest usefulness' To govern
oneself one must know onself, not only as so much bones and
flesh and nerves and blood (though these come into the
calculation) but each should learn to observe the nature,
quality, ability, strength, passions, of the mind. Each should
'explore' himself in his varied nooks, and even obscurities of
mental life. "On this account, it seemed good to me to
ponder deeply on some points of this great subject, all the
more because recent philosophers have brought but little
industry to the study, content with what had been left behind
by the ancients. They have added nothing at all except
problems, almost impossible of solution, and such that, even
if solved, would bring no fruitfulness. Formerly the ancients
involved themselves in great absurdities on this subject, for
they thought wrongly of the mind, which is not perceived by
1 For an account of the relation between Francis Bacon and Juan Luis Vives see
Rudolf Gunther, 'Inwieweit hat Ludwig Vives die Ideen Bacos von Verulam vor-
bereitet' (1912), and as to the influence of Vives on Descartes see, Roman Fade,
'Die Affectenlehre des Johannes Ludovicus Vives' (1893).
336 FOSTER WATSON
the bodily senses — and even their opinions on the very matters
which we do perceive by the senses were most inept."
Vives deliberately passes by the method then, — as ever — so
much in vogue of rebutting opinions deemed to be false, which
he says are more numerous in this than in any other subject of
inquiry. Such refutations would lead to thorns rather than
fruit.1 He will attempt to fit into the structure of his dis-
course, ' those expressions which have sprung from the people
and are'in common acceptance with them, and'then have become
unintelligible as they have passed into the use of the learned.'
For in dealing with the recondite problems of the mind, the
adaptation of a fitting vocabulary is a serious part of the under-
taking if a treatise is to be accommodated to learners. Vives's
•concern for his choice of language and vocabulary, it will be
remembered, was reflected, with such good consequences,
in his successors Francis Bacon, Thomas Hobbes, and Rene
Descartes, and very specially in John Locke's 'Essay on the
Human Understanding,' Book III., than which it would be
difficult to find a more masterly exposition of language as
the means of expressing ideas clearly.
In his address to the duke of Bejar, then, Vives claims
that he has parted company with the mischievous babblings
of the Stoics and their carping sophisms, and no less with the
hidden and subtle opinions of Aristotle.
In describing the contents of Vives's 'De Anima' it will be
more profitable to emphasize his independent and original
^views rather than to dwell upon the undoubtedly many
points of similarity which could be found in his book with the
.subject-matter in the works of Aristotle and St. Thomas
Aquinas.2
Tired of the atmosphere of dialectical displays which
inflated the pride of victory rather than stimulated the
search for truth, Vives was ready to try new paths. The time-
honored treatment of psychology included a discussion of the
1 Vives, however, to some extent, deals with the 'false opinions' of his predecessors
"in psychology, in his 'De Causis Corruptarum Artium' and in his 'De Veritate Fidei
Christianae.'
2 See T. G. A. Kater, 'J. L. Vives und seine Stellung zu Aristoteles;' G. Hoppe,
^Die Psychologic des J. L. Vives;' R. Fade, 'Die Affectenlehre des J. L. Vives.'
THE FATHER OF MODERN PSYCHOLOGY 337
question What is the soul? Yet Vives ventures to write a
book entitled 'De Anima' in which he says: "What the soul
is, is of no concern for us to know; what it is like, what its
manifestations are, is of very great importance."1 This calm
renunciation of the metaphysical aspect in favor of the de-
scriptive account of the activities of the mind is a natural
enough result from Vives's revolt from the older academic
disputational methods, which precluded progress, because
they could not reach further in their conclusions than they
assumed in their premises, and their premises were always
ultimately founded on a priori abstractions. Thus, disputes
on the nature of the soul either began in the unknown, and
ended there, or else sought an authoritative basis, which
was similarly substantially unknown (because not subjected to
enquiry). Vives, therefore, excluded the discussion of what
the soul is, in its essence, from his scope. Instead, he asks
for careful investigation into the manifestations of the soul
in all the activities of consciousness. "We cannot rightly
declare what the soul is in its essence, and as a bare thing
place it, as it were, before the eyes, but we can set it forth,
clothed and as if painted in a picture, in its own most apt
colors, so that it is seen in its own actions. For it has not
come under the observation of our senses, but we perceive its
works (opera} by almost all the senses, internal and external.
Surely the goodness of the Lord of Nature to us reveals itself
by many proofs, on every side, by thus aiding us, in placing
before our eyes and in such abundance, these manifestations
of the soul. For there is no sign more clear that things
are not to be converted to our use than for them to be
remote, rare, difficult to be prepared. Nor did he who bid us
know ourselves, refer to the essence, but to the actions of
our mind so that they may be ordered for moral life, and by
the expulsion of vice, we may follow virtue which will so
lead us that we may spend in full wisdom, as immortals, the
happiest eternity." It is a central doctrine of Vives that
Knowledge is of value, simply when it is 'put to use.' The
1 Anima quid sit, nihil interest nostra scire, qualis autem, et quae eius opera, per-
multum. 'Opera,' Vol. III., p. 332.
33$ FOSTER WATSON
observational method of studying the manifestations of our
minds has a value for application which cannot be gainsaid.
If indeed, the processes of cognition were to be regarded as
only of intellectual worth, even then the right study of the
passions, as they show themselves in ourselves and in the
individuals we read of in history, would be of direct ethical
significance for the purposes of examples or of warnings.
And, accordingly, Vives devotes one of the three books of the
'De Anima' to a critical and constructive study of the
passions.
We have seen that Vives places the emphasis in psycho-
logical studies upon the observation of the manifestations of the
soul, in its outward realized activities. When we thus observe
the results of mental activity in the numerous forms of cogni-
tion feeling, will, in other persons, or in ourselves, we are at the
point of view of empirical psychology. Vives may, therefore,
be claimed as a pioneer in the advocacy of this method, prior
to Francis Bacon and to Rene Descartes. It was not a long
step to take, to advance from the empirical psychology, which
is concerned in tracing the processes of mental activity in
others to that of recording, by way of psychological investi-
gation, the results of interrogation as to what has happened
in one's own mental experiences. Hence, illustrations can
be found in Vives, of the conscious employment of the intro-
spective, empirical method, the method that is especially
characteristic of later psychological students, and as these
appear to be the first instances of conscious introspective
interrogation of consciousness in psychological investigation
in modern times, i. e., from the Renascence onwards, they are
of more than ordinary interest and significance.
"As often," says Vives, "as I see a house at Brussels,
which is opposite to the Royal Palace, Idiaqueus comes into
my mind, for he is its occupier. Very often in that house,as
far as his business would allow, we used to chat over matters
pleasant to both of us. Now, as often as I revolve the idea
of Idiaqueus I do not think of the Palace, because the mem-
ory of my friend and his house is more noteworthy to me than
THE FATHER OF MODERN PSYCHOLOGY 339
the idea of the Royal Palace.1 So with sounds, tastes, smells.
When I was a boy at Valencia, I was ill of a fever. Whilst
my taste was deranged, I ate cherries. For many years
afterwards, whenever I tasted fruit I not only recalled the
fever, but also seemed to experience it again."2
These illustrations of the introspective method occur in
connection with Vives's treatment of two associated ideas
[recordatio gemina]. Of course, it is not the fact, as some
writers seem to suppose, that the doctrine of association of
ideas began with David Hartley or with the Mills nor even with
Thomas Hobbes, or John Locke. This explanation of men-
tal process, goes back, at any rate, to Aristotle. But it
may be claimed for Vives that his development of Aristotle's
theory marks the Renascence advance. No other author as
early as Vives contributed so strong and comprehensive an
exposition of it, as Vives. Indeed, Sir William Hamilton,
the most erudite of all British philosophers, in the history of
psychology, said: "Vives's observations comprise in brief,
nearly all of principal moment that has been said upon this
subject (of mental association) either before or since"*
Before Sir William Hamilton, Samuel Taylor Coleridge4
in his effort to discountenance Hobbes's claim to be the ' ori-
ginal discoverer' of the law of association, as advocated by
Sir James Mackintosh hit upon the £De Anima' of Vives, and
appears to have been the first Englishman to have drawn at-
tention to Vives's enunciation of this law of association of
ideas. This was in 1817. The words quoted by Coleridge
from Vives are: Qua simul sunt a phantasia comprehensa
si alterutrum occurrat, solet secum alterum representare.
1 Vives is explaining that our minds often travel more readily from the less to the
greater than vice-versa. Some might think that the house oposite to the Palace would
lead to the thought of the Palace. But Vives pays a compliment to his friend, as well
as expounds psychology, when he says his friend's house recalls the 'greater' i. t., the
more excellent idea of his friend and their former talks together.
2 Vives argues that, for this reason, in mnemonics, the * clues ' to excite memory
should not be themselves of such interest as to detain the attention too much from the
suggestion of what they are intended to assist in recalling.
3 'The Works of Thomas Reid' (including Hamilton's 'Dissertations'), 1872, 7th
ed., Vol. II., p. 896, column i.
4 In the 'Biographia Literaria,' Chapter V.
34° FOSTER WATSON
Coleridge proceeds to regard Vives as " subordinating all
other exciting causes of association to time. The soul pro-
ceeds ' a causa ad effectum, ab hoc ad instrumentum, a parte
ad totum'; thence to the place, from place to person, and from
this to whatever preceded or followed, all as being parts of a
total impression, each of which may recall the other." Chains
of associated ideas may have the most distant links connected
'by the same thought having been a component part of two
or more total impressions.' Vives's example quoted by
Coleridge is: "From the idea of Scipio I come to the thought
of the Turkish power, on account of his victories in that part
of Asia in which Antiochus was reigning."
Hamilton, who shows little mercy to Coleridge, tells us
that the whole of the latter's chapter on the history of the
law of association was 'conveyed,' to use the old expression,
from the German Maass, and is a ' blundering plagiarism.'
So, with regard to Coleridge's inference that Vives substan-
tially limits the law of association to that of the sequence of
connected ideas in time, i. e.7 to the phenomena of recollection,
Hamilton clearly shows that Vives does not so limit associative
ideas to time, or to place, but he points out their operation
6 in all the connections of thought and feeling*
The insight which Vives thus showed in his exposition of
the law of association of ideas prepares us for further details
in his experiential account of memory. As we have two
hands, so memory is twofold, and consists in apprehending
and retaining. The differences in memory amongst men are
implanted by nature. Some men like Hortensius remember
words more readily; others, e. g., Themistocles. Some
apprehend quickly and retain better the curious; others,
the simple; some, public affairs, others, private matters;
some, the old, others, the new; some, their own affairs, others,
those of others; some, vices, others, virtues; — each according
to the proneness of his disposition. For he attends more
willingly and therefore more closely to this or that, and at-
tention strengthens memory. A natural bodily constitution
is of the highest importance to memory, and with such were
endowed the great men whose magnitude of memory has been
THE FATHER OF MODERN PSYCHOLOGY 341
handed down, e. g., Themistocles, Cyrus, Cineas, Hortensius.
This natural endowment is capable of being strengthened by
habits of living. Memory is more tenacious in slow than in
quick minds, like an impression driven in for a long time on stone
or iron, but the swift minds more readily bring back a reminis-
cence. A deep descent into memory is made by those things
to which we give close attention and care when first perceived.
If anyone is in a state of emotion or excitement on first hear-
ing of something, and this is mingled with the memory, this
reminiscence is easier, quicker, longer, as, e. g., for what has
entered the mind when in a state of violent grief the memory
is the longest, and for that reason, adds Vives, in determining
the boundaries of property it is the custom of some races to
bring their boys there, and thrash them with severity, so
that the boundaries may be more firmly, and the longer,
retained in memory. By practice and frequent meditation
the memory gathers great strength. Unlike other gifts
of the mind, which do not deteriorate by rest and cessation,
memory grows fainter and fainter, day by day, if it is not
exercised.
The law of forgetfulness is fourfold, (i) When the im-
age painted in the memory is utterly scratched out; (2) when
it is smeared or broken in pieces; (3) when it stealthily evades
us in our search for its recall; (4) when it is covered over as
with a veil, in disease, or in the excitement of emotion.
Vives sketches the method of recall of an idea; by retracing
as it were, our steps, we may come in thought to what we are
seeking. Thus from the idea of a ring we think of a gold-
smith, from him to a queen's collar; thence to a war which
her husband waged; from the war to the leaders; from the
leaders to their ancestors or children; thence to the studies
in which the latter are being exercised. To such series of ideas
there is no end. These steps spread themselves widely
through all kinds of subjects, from cause to effect, and so on,
as we have seen in Coleridge's reference. Vives, however,
gives an example, which Coleridge omitted, but which we
may note, since we are emphasizing the appeal to experience
made by Vives. "In thinking of the name of Cicero, there
342 FOSTER WATSON
comes into my memory the name of Lactantius, for he was the
imitator of Cicero. From him, I proceed to think of the
copper-plate artist, since his book is said to have been either
the first or amongst the first printed from copper-plates."
Remembrance is of two kinds. It is either natural, or
voluntary, when we pass freely from one idea to another, or
it is ordered (jussa) when the mind makes an effort to reach
and bring back some idea. Those ideas which occur in a
series are more easily remembered, as for instance, in mathe-
matics. Verses are conducive to faithful retention on account
of the rhythm of composition, which keeps the mind from
straying outside its limits. The art of mnemonics is based
on the order of what is committed to memory. It is at this
point that Vives introduces his definition of association.
For with regard to ideas 'which have been included in the
imagination [phantasia] at the same time if one of the two
should occur, it is wont to bring back with it, the other.'
According to his custom, Vives illustrates this definition.
"From the sight of a place there comes into the mind what
we know once happened in that place, or what was situated
there. When something joyful happened along with a voice
or sound, we are delighted when we hear that same voice or
sound again. If it was a sad event, we are saddened. This
is also to be noted in the lower animals, who if they receive
something pleasant, after they have been called by a sound,
they will run towards it readily and gladly, on hearing the same
call again. But if they were beaten, after being called, they are
frightened by the memory of the blows, on hearing the same call
again. Returning to memory, another factor in clear memory,
Vives points out, is the time-element. A distinction of times
is necessary in reminiscence, otherwise images are confused,
as if in a picture, other pictures should get painted on the
top, after an interval. Those images which we have received
with a quiet, leisurely mind imprint their trace for a longer
time, and more permanently, if we gave our attention to
them. It is for this reason that the things we have seen and
heard in early life are recollected by us so clearly. For at
that age the mind is free from cares and thoughts. Every-
THE FATHER OF MODERN PSYCHOLOGY 343
thing is new. We therefore watched closely these things
which won our admiration and they sank deeply into the
mind. Older men are preoccupied, and in an * internal
agitation of thoughts' so that they cannot so restfully admit
new ideas, or find out readily the old amid their experiences.
In connection with associations by similarity, Vives is
impressed by the opportunities for error in memory and in
thought, as we pass from like to like. We mistake Georgius
for Gregorius; problema for enthymene; Pindarus for Pan-
darus. Similarity may appear in the beginning, middle, or
end, of words. Or the similarity may be in things, from the
manner in which we concentrate our attention on them. In
philosophy we may think of Xenocrates for Aristotle; in
speaking of the Carthaginian Wars we may confuse Scipio
with Q. Fabius. In a question of poverty, we may say Irus
instead of Codrus. In eloquence, we may ascribe something
to Demosthenes instead of to Cicero. In a matter of beauty,
we may speak of Narcissus for Adonis. Or in a question of
smell, mistakes may be made of garlic for onion. So, in
matters of times, actions, qualities. Similarity, therefore,
disturbs the memory, as it does the eyes of the body, so that
judgment is confused. Vives sees that errors can creep in at
the * first attention' when an idea is received or at the 'second
consideration' when reminiscence wrongly draws forth those
ideas which were received as wholes. "Yesterday," illus-
trates Vives, "in the market-place, Peter of Toledo saluted
me. I did not notice the fact sufficiently, nor remember it
accurately. If anyone asks me: * Who saluted you yesterday
in the market-place? ' and adds nothing further, I shall answer
more readily than if he were to say: ' Was it John Manricus or
Ludovicus Abylensis?' The labor thus becomes two-fold,
first to reject what does not fit, secondly to replace it by what
is right."
We can thus see that Vives is permeated with the idea of
an empirical and introspective method of the modern type
in his 'De Anima,' and that the application of this method
has taken him far along the path of development of the
doctrine of association of ideas, and in the exposition of ob-
344 FOSTER WATSON
servational aspects of memory. There are many other views
of Vives which have special interest. Thus his emphasis on
the necessity of observing and distinguishing the great
variety of differences in men's minds became the basis of the
treatise of his fellow countryman, Juan Huarte, written in
1557, and translated into English1 by Richard Carew in 1594,
under the title of the 'Examination of Men's Wits.' This
book following on explicit suggestions in Vives's £De Anima,'
demands that children's impulses and tendencies, both in
play and at work, should be studied so as to afford a psycho-
logical basis for their studies and after-occupations. Vives
distinguishes carefully between the ratio speculativa, whose
1 end' is the truth and the ratio practica whose end is the
'good.' The Spanish biographer of Vives, Senor Professor
D. Adolfo Bonilla y San Martin, points out the parallel
between Vives's view and that afterwards developed by Kant.
Vives introduces a priori subjective forms of reason which
he terms anticipationes sen informationes naturales, and as
Bonilla remarks, anticipationes naturales is also the term used
by Francis Bacon in 'Novum Organum,' Book I. Nor will
theologians and philosophers, if they turn to Vives's 'De
Anima,' pass by without notice his treatment of the problems
of free-will, and of immortality, for they present an interesting
individual statement of these problems — of this Renascence
period.
Educationists ought to realize that Vives writes in the
'De Anima,' a section which he terms fde Discendi Ratione.'
He attempts the well-trodden path of an evaluation of the
order of intellectual resources and discipline afforded by the
several senses. He remarks: "The course of learning is
from the senses to the imagination, and from that to the mind
of which it is the life and nature, and so progress is made
from singulars to combinations, from singulars to the univer-
sal; which is to be noted in boys. . . . And so the senses are
our first teachers2 in whose home the mind is enclosed" It is
in this section of the 'De Anima' that Vives refers to the ad-
1 Indirectly, from the Italian of Camillo Camilli.
2 Cf. Rousseau's often-quoted "Our first teachers of philosophy are our feet, our
hands, and our eyes." 'Emile' (Payne's ed.), p. 90.
THE FATHER OF MODERN PSYCHOLOGY 345
miration, almost bordering on incredulity, with which he
heard that a deaf-mute had become taught, and he points
out that the method of learning necessarily implied a large
measure of self-teaching.
Whilst the psychological views of Vives, sketched above,
are of this most significant modern cast, yet it must be stated
clearly that the general setting from which they are taken is
that of the old Aristotelian psychology, and if there are modi-
fications in statement, these are usually clearly based on scho-
lastic writers, for as a Spaniard, Vives was thoroughly unwilling
consciously to borrow or adopt from Moorish sources. He
treats of 'souls' as distinguished from ' torpid things' (or the
inorganic world), the distinction being founded on the power
of self-movement. He is thus provided by the Aristotelian-
Scholastic psychologists with (i) the anima alens of the plants,
(2) the anima sentiens of zoophytes, (3) the anima cogitans
of birds and four-footed beasts and (4) the anima rationalis
of man. The soul of man is the 'form' of which the body is
the 'matter.' The human body therefore is a potentiality,
which is actualized by the union with it of the soul. But this
principle of the form giving actuality to the potentiality of
suitable matter is characteristic of all life, e. g., in plants,
zoophytes, birds, beasts — up to and including man. The
lowest kinds of life, e. g., plants, have the motions of nutrition,
growth and decay. Animals, in addition, have sensation,
and developments from sensations. Man combines all these
stages, including the appetency which springs from sensation,
cognition and reasoning. The soul of man is therefore an
epitome of all lower life and also possesses psychical ingredients
of its own. Hence, an investigation is necessary into the vege-
tative 'soul,' and the 'animal' soul, as well as the human soul.
Thus the physical phenomena of nutrition, growth, decay, gen-
eration, sensation (in general) and the special senses (and the
hierarchy amongst them), are passed in review, then 'interior
cognition' including imagination, phantasia (into which angels,
good and bad, insinuate themselves), the sensus communis, the
cognitive judgment and reason. Vives then offers his defini-
346 FOSTER WATSON
tion of the soul1 and makes the usual inquiry as to the seat
of the soul, deciding, with Aristotle, that it ' informs' the
whole body; though certain functions are localized, e. g., the
front part of the brain is the seat of phantasia and in the back
part of the brain memory is localized, and so on.
The above topics comprise the contents of Book I. of
Vives's 'De Anima.' Book II. is devoted to the rational
soul and its faculties. Man has been created for eternal
felicity and provided with the means for accomplishing it.
This implicitly demands the intelligence to know the good, the
memory to retain that knowledge, and the will to act it out
in life. We thus have the trinity of the soul. Vives then
describes in detail the functions of the simplex intelligentia
(simple apprehension), memory and reminiscence, composite
ideas, reason, judgment, mental ability and its individual
varieties, speech, the method of learning (in which Vives
inquires why there are so few learned people), knowledge,
contemplation, will (in which occurs the discussion on its
freedom),2 on the mind in general, on sleep and dreaming,
'habit,' old age, length of life, death, and the immortality of
the human soul.
In the third book of the 'De Anima,' the subject of which
is the emotions or passions (affectus), Vives undoubtedly
followed3 St. Thomas Aquinas. We have seen that Vives
regarded the intellect as supplying knowledge as to the good,
which the will was to carry into effect. We are, therefore,
prepared to find that 'passions' are looked upon in their
relation to the supreme end to be achieved in volition. They
are defined by Vives as 'the natural faculties of our soul by
which we are carried towards the good and endeavor to avoid
1 This follows substantially the teaching of Aristotle. Animam esse agens praeci-
pium, habitans in corpore apto ad vitam.
2 It is worth noting that Vives strongly protests against the doctrine current long
after his age, that the will is controlled or influenced by the motions of the stars. The
significance of the protest is only realized when we remember that Tycho Brahe,
Kepler and even Galileo 'cast nativities.*
8 Aristotle is out the question. As Mr. Hicks says Aristotle 'exalted the cognitive
element, while his treatment of the emotions and the will is wholly inadequate even if
the ethics and the rhetoric be called in to redress the balance.' Aristotle de Anima,
p. LXXII. On the parallels between Vives and St. Thomas Aquinas see Roman Fade,
'Die Affectenlehre des J. L. Vives,' Miinster i. W., 1893.
THE FATHER OF MODERN PSYCHOLOGY 347
the evil.' The close relation, therefore, between psychology
and ethics is evident. Vives's treatment of the passions,
though based on St. Thomas Aquinas, is yet largely supple-
mented by his own introspection and observation.
Whilst thus he tends to emphasize the interest in the
description and analysis of the separate emotions, and to give
the result of his wide intercourse with men of very varied
kinds, he has not been so thoroughgoing in his psychology of
emotion in general. He does not offer an elaborated theory
of the passions like Descartes and Spinoza, leading in the one
case to a discussion of emotion in the abstract, and in the
other to an a priori mathematically-based theory, but he is
quite as comprehensive and at points shows depth of empirical
interest, which place him in the direct line of continuity be-
tween the Scholastics and the modern descriptive school of
psychology, in the treatment of the passions. Vives reduces
all the passions to two, love and hate. All that stirs and
stimulates towards the good, comes from the incitement of
love, and all that stirs to evil, from the passion, in some
form, of hate. Yet Vives describes fully and separately the
passions : good-will, respect, sympathy, gaiety, hope, laughter,
annoyance, scorn, anger, hatred, envy, jealousy, indignation,
vengeance, cruelty, sadness, mourning, fear, shame, pride.
Vives is strongly attracted to the Platonic treatises, especially
the 'Phsedrus' and 'Symposium.' The value of the third
book, as with the other two, is in its empirical method, for
it also contains numerous observations, personal illustrations
based on Vives's own introspective experiences. The dis-
tinguished psychologist Harold HofFding1 in the section on
the psychology of the feelings, whilst discussing the topic of
laughter, quotes Vives, who described himself as unable to
refrain from laughter as he first tastes food after a long fast
(*De Anima,' Bk. IIP). And, again, HofFding recalls the
observation of Vives, that what man expresses by laughter
may be expressed by animals in other ways3 (e. g., by wagging
the tail).
1 'Outlines of Psychology' translated by Mary E. Lowndes, 1891, pp. 291-2.
2 'Opera,' III., p. 469.
3 'Opera, 'III., p. 470.
348 FOSTER WATSON
In his professedly psychological work — throughout the
three books of the fDe Anima' — Vives discloses himself as
the pioneer of the modern empirical method in psychology.
But we only realize the full sense of conviction, which ani-
mated Vives in his use of the method of introspection and
observation, when we further note in his other writings, his
constant application of the same method in the ordinary affairs
of life. He applied psychological principles to professional
practice, to individual conduct, and particularly to the func-
tion of teaching. In other words, in practical affairs, he
sought to introduce psychological precepts and methods, to
create a habit of introspection which might be turned to use —
to create an atmosphere of psychology, to thinkpsychologically.
No writer of the Renascence period was so distinguished
by his application of psychology to education, as Vives. In
his 'Transmission of Knowledge' ('De Tradendis Disciplinis,'
1531) he is permeated with the desire to bring education to a
psychological basis. We have seen that his account of the
memory was an outstanding feature in his de Anima. But in
dealing with the subject from an educational point of view,
he is in accordance with the most modern of writers in pointing
out that both quick comprehension and faithful retention
in memory are helped by a right arrangement of facts. This,
he adds, is just that art of memory ' which beasts are said
to lack.' His rules for the cultivation of memory have not
lost their suggestiveness. For instance, he says: "What we
want to remember must be impressed on our memory while
others are silent. We need not be silent ourselves, for those
things which we have read aloud are often more deeply
retained. ... It is a useful practice to write down what we
want to remember, for it is not less impressed on the mind
than on the paper, by the pen. The attention is kept fixed
longer by the fact that we are writing it down." " Great is
the help to memory if reasons are associated with the matter
taught." We have seen that Vives recognizes interest as a
strong stimulus to the attainment of knowledge. He goes
far in the direction of Herbart in the recognition of interest
not merely as a means in the acquisition of knowledge but also
THE FATHER OF MODERN PSYCHOLOGY 349
in the advocacy of a wider scope of interests as the outcome of
our studies. Vives in his story of Charles Virulus the school-
master of Louvain is as modern in significance as Herbart
himself. When a pupil's parent came to visit, especially to
dine with him, he made a point of finding out what his visitor's
work and interests were and prepared himself carefully to con-
verse on matters familiar to his guest, and lead him to speak
freely on what was best known to him. "He would thus hear
in the briefest time details which he could scarcely have
gleaned from the study of many years."
Education with Vives is not the preparation for a career,
but the increase in practical wisdom of life and the preparation
for moral excellence. In each school masters should meet
four times a year and discuss the 'nature' of each boy and
then apply the boy to those studies for which he seems most
fit. The fruit of studies is not honor or money, but the
culture of the mind — 'a thing of exceeding great and in-
comparable value — that the youth may become more learned
and more virtuous through sound teaching.' Boys should
only at first be taken on trial in the school. The teachers are to
determine who are fit and who unfit for learning. As Ascham
maintained, Vives previously urged that the slow wit is usually
the surest. The wonderful variety of dispositions in boys
requires the closest attention of teachers, in 'choosing'
scholars. Yet there is scarcely anyone who will not profit
by being taught, if the right sort of teaching is given him.
Probably no Renascence writer has taken so much notice of
the problem of the feeble-minded, the deaf and dumb, the
blind, as Vives, though naturally he has not been able to sug-
gest the most effective lines of training in each case. But his
firm grip of the principle of suiting instruction to the individual
capacity puts him in the direction of perceiving the problem
involved. Vives, again, sees clearly that the problem of
education is essentially that of self-activity. He requires the
pupil to keep paper notebooks, in which he gathers for himself
the main materials of his own instruction. His notebooks
must have divisions and heads, and be provided by himself
with indexes. In these he must enter, under proper heads,
350 FOSTER WATSON
all he learns from teachers and books. In other words, he
must largely make his own text-books. He presents a well
thought out psychology of school punishment, and indeed his
psychology of examinations, if we may so call it, is in advance
of present-day methods, for, instead of pitting boy against
boy (when he has emphasized the great variety of original
mental capacity) he logically requires the comparison of the
boy with himelf at an earlier stage. "Let scholars keep what
they have written in earlier months, in order to compare it
with that written at a later month, so that they may perceive
the progress made, and persevere in the way of improvement."
Thus, whilst Vives sees the overwhelming importance of
building up the art of pedagogy upon a sound psychological
basis, he by no means limits the value of the application of
psychology to the work of the schoolmaster. The knowledge
of psychology is essential to all who have to deal with spiritual
affairs. "The study of man's soul exercises a most helpful
influence on all kinds of knowledge, because our knowledge is
determined by the intelligence and grasp of our minds, not by
the things themselves." The text-books for psychology recom-
mended by Vives are the sacred writers of the Bible, and the three
books of Aristotle's ' De Anima ' (especially Books II. and III.).
Other writers to be read are Alexander, Themistius, Timseus
of Locris, and Plato's ' Timseus/ Proclus, Chalcidius, and of
the Renascence writer, Marsilius Ficinus, who will act as guide
for Plotinus. The physician, to Vives, is on the very border-
land of nature-study and soul-study, and must be at home in
both. But the psychological aspect must be present not only
in his studies, but in his professional habits. He must
himself "not be in infirm health, not pallid in countenance,
so raising the suggestion put in the sacred Gospel: ' Physician
heal thyself.' Further let the doctor be clothed neatly
rather than sumptuously. At the first sight of his patient
he will immediately take in his appearance and constitution,
age and vitality. All necessary information he will gather
in an urbane and affable fashion. He will listen with patience.
. . . He will neither exchange views nor discuss with other
physicians in the presence of patients, or of lay-people, who
THE FATHER OF MODERN PSYCHOLOGY 35 *
know not which side to take. To do so, easily raises a ' certain
despair' in them and a hatred against knowledge, 'which
comes to be regarded as a matter of uncertainty.' '
In the case of the historian, wars and battles are to be
regarded as 'cases of theft.' The historian should study
peaceful affairs, trace the glory and wisdom of virtuous acts
and note the disgrace of evil-doers. The wisdom of great
statesmen, and those who have excelled in 'good arts,'
philosophers, saints of the faith, and all that has been said
in practical affairs should be studied with the full force of
weighty intellect and judgment. For "it is unworthy to
hand over to our memories historical actions due to our pas-
sions, and not also to study what took place as the outcome
of the rational judgment." History discloses the essential
nature of human beings, and discloses the manifestations of
the affections and judgment of the human mind, in short, the
subject has a distinctly psychological basis.
Again, the politician and economist have to study the dis-
positions and minds of the people. Herein is the predominant
value of experience. " Sometimes old men converse with one
another in an experienced way and allow youth to listen."
This is a privileged method of study to youth, if they kept free
from the company of cavillers and obstinate dialecticians.
"For a man to be more anxious about achieving a dialectic
victory than of discovering truth leads to the ruin of practical
wisdom, as indeed Cicero has said." Vives points out the
mental characteristics desirable in the administrator, and for
the study of political philosophy recommends not only Plato
and Aristotle and other classical writers, but also the 'Utopia'
of Sir Thomas More and Erasmus's 'Christiani Principis
Institutio.' Equally clearly the lawyer must be a psychol-
ogist. He must understand "the common nature of man-
kind, the views and customs of many kinds of people, espe-
cially of his own country. This is brought about by wide
experience in seeing, hearing, observing things; through read-
ing of the deeds of ancestors and varieties of changes which
have befallen the state. Such men need alert minds and keen
judgments, so as to observe and to estimate circumstances, one
by one."
352 FOSTER WATSON
In all these instances Vives's introduction of psychological
observation bears a modern aspect, and affords illustration
of the attraction which he felt towards the empirical stand-
point and self-exercised thought on the environment rather
than the older type of abstract, metaphysical explanation or
discussion of the more ultimate foundation of psychological
phenomena. The most marked characteristic of the early
Renascence writers is the backward-looking concentration on
the golden age of Roman and Greek culture and knowledge.
In a notable passage1 however we find Vives exclaiming:
"The student should not be ashamed to enter into shops and
factories, and to ask questions from craftsmen, and to get to
know about the details of their work. Formerly, learned
men disdained to inquire into those things which it is of such
great import to life to know and remember. This ignorance
grew in succeeding centuries up to the present ... so that
we know far more of the age of Cicero or of Pliny than of that
of our grandfathers." His keen interest in the experiential
side of psychology, therefore, infused itself into his whole
outlook, in wishing to get an intellectual grip of the problems
of the human mind, and its manifestations, in its relations to
its own growth and development, and also in its actions and
reactions, in connection with its environment. No doubt he
believed that the glorious achievements and experiences of
the past threw light on those questions, perhaps to a degree
which has been lost in modern times, but Vives's psycho-
logical attitude towards life, in its present environment is
clearly modern, rather than ancient or mediaeval.
In England, Vives's name has fallen into undeserved
oblivion. For the Spaniard of Valencia made his home in
this country for portions of the year from 1523 to 1528, the
year in which he had to withdraw from England, on account of
his known adhesion to the cause of Queen Catharine of Aragon,
who had such a belief in his abilities that she desired him
to act as her advocate in the court, so adroitly consti-
tuted by Henry VIII to try her case. During his visits
in England, Vives lectured on rhetoric at Oxford, where
1 'De Tradendis Disciplinis,' Book IV., Chapter 6. 'Opera,' VI., p. 374.
THE FATHER OF MODERN PSYCHOLOGY 353
he was associated with Corpus Christi College. He was one
of the friends of Cardinal Wolsey, and of Sir Thomas More.
Yet, curiously, as we have seen, it is to a representative of
Scotland, Sir William Hamilton, that Vives particularly
owes his acknowledgment, in the last century, though
another Scot, Dugald Stewart, perhaps even more fittingly
brings out the modern aspect of Vives, in the striking passage:
"Of all the writers of the sixteenth century, Ludovicus
Vives seems to have had the liveliest and the most assured
foresight of the new career on which the human mind was
about to enter. The following passage from one of his works1
would have done no discredit to Francis Bacon's 'Novum
Organum': 'The similitude which many have fancied between
the superiority of the moderns to the ancients, and the ele-
vation of a dwarf on the back of a giant is altogether false
and puerile. Neither were they giants, nor are we dwarfs,
but all of us men of the same standard, — and we the taller of
the two, by adding their height to our own : Provided always,
that we do not yield to them in study, attention, vigilance and
love of truth; for, if these qualities be wanting, so far from
mounting on the giants' shoulders, we throw away the ad-
vantages of our own just stature, by remaining prostrate on
the ground."1
1 'De Causis Corruptarum Artium,' Bk. I., Chap. 5. 'Opera,' VI., p. 39.
AN INVESTIGATION OF THE LAW OF EYE-
MOVEMENTS
BY MILDRED WEST LORING1
The first investigation of eye-movements was made in
1826 by Johannes Miiller (i). He stated that the eyes in
their movements do not rotate about their long, i. e., sagittal
axes. He said: "I have convinced myself, while observing
various points on the white of the moving eye, which were
marked beforehand with ink, that the eye, through action of
the oblique muscles, does not rotate about its long axis."
In 1838, however, Hueck (2) demonstrated the compensatory
rotation of the eyes around the line of sight, by observing
that a given horizontal blood-vessel on the conjunctiva re-
mained horizontal even when the head was inclined to the
right or left. Burow (3), 1841, reached the same conclusion
using his own paralyzed iris for the demonstration. But
Ruete (5), in 1846, like Miiller, got negative results, using
after-images for a criterion. Donders, (7) 1848, showed his
work to have been careless, and demonstrated the well-
known principle, that the after-image of a vertical strip
remains parallel to itself, with vertical and horizontal move-
ments of the eyes, but with oblique movements, becomes itself
oblique. He however formulated no law.
It remained for Listing to put the law concisely in the
form now known as Listing's Law, although he did not prove
it, or publish it himself. It first appeared in this form in an
article by Ruete (n) in 1855, in which he says: "The principle
of the mechanism of the eye can be expressed according to
Listing in the following simple way; 'From the above-men-
tioned normal position of the eye which may be called the
primary, the eye will be moved into any other secondary
position by the cooperation of the six muscles, in such a way
1 The results of this paper were obtained in the psychological laboratory of the
University of Washington, 1912-13, under the direction of Dr. H. C. Stevens.
354
LAW OF EYE-MOVEMENTS 355
that this displacement can be represented as the result of a
rotation about a definite axis different from the above three,
which always passes through the center of the eye, and is
perpendicular to the primary and secondary position of the
optic axis, so that each secondary position of the eye stands
in such a relation to the primary, that the rotation projected
on the optic axis will equal zero."
Meissner (9), in 1854, was the first to use the method of
double images, which he observed are not parallel for a given
object with near or far fixation. The inclination of the two
images he took as representing the torsion of the two eyes
about the line of sight. Both Meissner (13), (1860) and Pick
(12), (1858), investigated the subject by means of the change
of orientation of the blind spot with eye-movements. Pick's
results are irregular, and he says that the movement of the
eyes is not a simple geometrical one, about a definite axis, as
Meissner held, but that it is a physiological change of position,
whose beginning and end only, we know. Meissner found
his results to be similar to those of Listing.
The next important work on eye-movements was done by
Wundt (14) in 1862. He used the method of after-images,
and presented the theory that the eye rotates to such a degree
as to allow it to take the desired position of the line of sight
with the least muscular effort. He found a slight torsion
inward for movements of the eyes vertically above, and torsion
outwards for movements vertically below. Helmholtz (15),
in 1863, denied the reliability of Wundt's hypothesis, owing
to individual differences in muscle strength, and the general
unreliability of the muscle sense. He preferred to state the
facts in the form of Bonder's law, which says: When the
position of the line of sight in relation to the head is given, we
have a definite and unchangeable amount of rotation.
Volkmann (16), in 1864, attacked the problem somewhat
after the fashion of Meissner, using two rotating discs, one
for each eye, on each of which was marked a diameter. These
were rotated till judged parallel, for any given convergence,
and the real torsion thus determined. The torsion as ex-
pressed by its effect on the two diameters placed vertical,
was as follows:
356 MILDRED WRING
Primary position 2°.2i
30° above to the right 2°.74
30° above to the left 2°.92
30° below to the left i°.3 1
30° below to the right i°.4O
This small amount of variation accounts, he says, for the
difficulty of using after-images.
Helmholtz (17), 1866, using after-images, got the following
results:
A. Turning the eyes to the right above or left below
1. After-image of horizontal line is turned to the
left.
2. After-image of vertical line is turned to the
right.
B. Turning the eyes to the left above or right below
1. After-image of horizontal line is turned to the
right.
2. After-image of vertical line is turned to the left.
In 1868, Hering (18) published his results on binocular
vision. | In the main he agrees with Helmholtz. He
worked with (a) after-images, using parallel lines of sight,
observing the customary phenomena of Listing's figure, (b)
double images, using convergent lines of sight. His general
conclusion was that Listing's Law holds for parallel lines of
sight, but not for convergence lines of sight. According to
him, then, Donder's law is invalid, inasmuch as it maintains
a constant torsional value for a fixed position of the line of
regard in relation to the head.
Le Conte (20), in 1881, published a severe criticism of
Helmholtz's version of Listing's Law, and framed it in terms
exactly opposite to those of Helmholtz. The whole occasion
of the difference of conclusion rests, as Le Conte shows, on the
fact that he has used the normal spherical surface for pro-
jection of the after-images, while Helmholtz stated the law for
projection on a plane surface. Le Conte denies all real rotation
about the sagittal axis. It is only apparent, he says, and
attributes it to a rotation about some axis perpendicular to
the sagittal axis, which, because we observe all the movements
of a spherical surface like the eyeball from one point of view.
LAW OF EYE-MOVEMENTS 357
we interpret as rotations about the line of sight. So Le Conte
concluded there was no torsion with parallel lines of sight;
but for convergence he finds not only an apparent but a real
torsion about the line of sight.
One of the most recent studies on eye-movements has been
made by Bernice Barnes (23) in 1905, at the University of
Michigan. Her apparatus which she calls a torsio-meter,
consisted of 'an iron arc of 180°, one meter in diameter,
mounted on a standard, so that its center may be directly in
front of the eye of the subject, who is seated before it.' The
arc could be swung into horizontal, vertical and oblique posi-
tions. It was fitted with a telescope which was directed at
the observer's eye. The latter seated himself in front of the
arc, with head firm; a thread stretched between the extremities
of the arc served to determine the center of the arc, at which
point the observer kept the pupil of his eye. The cross-hair
in the telescope was set in coincidence with a line on the iris,
which reading was taken as zero. Then the telescope was
changed in position, and the eyes were again directed at the
telescope. The cross-hair in the telescope was set again on
the same line on the iris. In this way any real torsion of the
eye about its sagittal axis was determined. Miss Barnes
found real torsion for every position different from the primary
position, even with parallel lines of sight, a conclusion very
different from Listing and his successors'. Her chief con-
clusions are these:
1. There is a contradiction between Listing's and Bonder's
Laws for torsion in eye-movement.
2. Experiments by the after-image method seem to con-
firm Listing's Law. But there are two sources of error —
inaccuracies in measurement, and false torsion, which it is
difficult to make allowances for.
3. More accurate direct measurements show that there
is always torsion with rotation, and the amount of torsion is
proportional to the amount of rotation.
4. Donder's law holds.
Miss Barnes's work is interesting for the reason that
she found, even with parallel lines of sight, an invariable real
358 MILDRED WRING
torsion with movements away from the primary position in
any direction whatever. Yet a closer analysis of her results
shows certain ambiguities.
1. There is nothing in her article to show which eye of
the observer was used in the experiment, or whether both
were used indifferently.1 This is an important point, for it is
possible that the two eyes undergo torsion symmetrically,
but in opposite senses, when the lines of sight are either
parallel or converged. An adequate interpretation of her
results is impossible without a knowledge of this point.
2. Her results show that there are some regions, for in-
stance between o° and 30° to the right above, where the sense
of the torsion changes, for at o° it is to the right and at 30°
it is to the left. This must indicate one of two things;
(a) that somewhere between o° and 30° there is no torsion at
all, this position being at the point where the direction of the
line of sight changes from right to left. This point, mathe-
matically speaking, could be called the point of inflection.
(b) Otherwise the phenomenon must indicate a region in the
field of vision such that, if one directs the line of sight there,
the sense of the torsion depends on chance, that is, it is as
likely to be to the right as to the left and cannot be predicted.
Since Miss Barnes denies any direction in which there is no
torsion, the first alternative is for her out of the question.
She would of necessity have to accept the second. But this,
however, violates Donder's Law, which Miss Barnes upholds,
inasmuch as the law mentioned demands a constant torsional
value for every given position of the line of regard.
The only conclusion that can be drawn from Miss Barnes's
work is that further investigation is necessary in the above-
mentioned regions of the fields of vision, using each eye, and
comparing the results for the two eyes.
Wichodzew (24), 1912, has recently contributed to the
subject. He has used the size of the field of vision (both
monocular and binocular) to determine the influence on eye-
movements of the inclinations of the head to the shoulder.
His main conclusions are:
1 1 have since learned that only one eye of the subject was used, but which, is not
recalled by the author.
LAW OF EYE-MOVEMENTS 359
1. There is a compensatory torsion (Raddrehung) of the
eyes about the sagittal axis, which causes a change in the
mutual relation of the fixation points of the eye muscles and
so hinders the eye movements.
2. The capacity of the eyes for symmetrical torsion about
a sagittal axis increases with inclination of the head to the
right or left shoulder. This increase might be explained as a
stimulation by the compensatory torsion to the innervation
of the muscles, which turn the eye about the sagittal axis.
The work of Wichodzew, like that of his predecessors,
still leaves the question of the sense of the torsion for each
individual eye still undecided. Besides there is no reason
for accepting a priori that the movements of the eyes are the
same when the fixation point is kept constant and the head
rotated, as when the head is kept constant and the fixation
shifted.
REPETITION OF PREVIOUS EXPERIMENTS BY HERING AND
HELMHOLTZ
Hering. — When two parallel strips, at interocular distance,
are looked at with the eyes in the primary position, the two
strips are seen as three, all parallel. But as soon as the
fixation point approaches nearer than infinity, the three
break into four, the inner one making two images not parallel
but converging at the top. This is due to torsion of the
eyeball, which increases as the fixation point is brought nearer
the observer. Likewise, when the eyes are shifted horizontally
to right or left, the convergence of the two middle strips
increases, the greater the shift of the eyes. The purpose of
this experiment was to measure the amount of torsion and
to measure individual differences in the amount of torsion.
Essentially the same apparatus and procedure was used
as that described by Hering (18). Likewise the results
obtained agree with Hering.
I. The amount of torsion of the eyes, as measured by the
convergence of double images of two straight lines, increases
directly as the amount of convergence of the eyes increases,
from an apparent minimum of zero to the maximum with
maximum convergence.
MILDRED WRING
2. The torsion of the eye increases as the point of fixation
is directed to a greater distance horizontally either left or
right of the primary position.
3. Keeping the fixation point constant, a bending of the
head forward increases the amount of torsion; raising the
head decreases the amount of torsion.
Helmholtz. — (a) It is an established fact that we do not see
all actual straight lines as straight. In fact, if we look at a
line which is really straight, thus with parallel
lines of sight and a card held in front of the nose in the median
plane of the body so that each eye sees but half the line, it
will not seem straight but broken, the two arms converging
thus (allowing for much exaggeration):
If now by some device we compensate for this apparent
brokenness, the line must actually be made thus:
Now if the eyes are converged to the middle point of the line,
where the card meets the line, the line which had been made
straight apparently, seems broken again thus:
And to compensate for this additional apparent slant of the
arms of the line, the two arms of the actual line must be bent
toward one another downward, still more, thus :
The apparent convergence of the two arms of the line there-
fore increases as we pass from an infinite fixation point to
one very near the eyes. This is due to actual torsion of the
eyeball about its sagittal axis, the right eye turning to the
left and the left eye turning to the right, from the point of
view of the observer.
LAW OF EYE-MOVEMENTS 361
The apparatus, based on Helmholtz's idea (17), consisted
of a strip of cardboard 10 X 1.5 cm., fastened horizontally
at one end by a screw to a vertical screen of cardboard.
Along the length of the strip, which was at the level of the
eyes, was drawn a black line which was extended through the
point of rotation to an equal horizontal distance on the screen.
This gave a horizontal line whose apparent brokenness when
fixated under certain conditions, with a vertical screen per-
pendicular to the face in the median plane, could be com-
pensated for by rotating the strip, and so give a measure of
the rotation of the eyes.
The results indicate that with any given fixation point,
there is a torsion of the eyes, the left eye to the right, and
the right eye to the left, from the point of view of the observer
himself. This result agrees with that obtained in Hering's
double image experiment above, and with the results of our
original experiment below.
Helmholtz. — (b) It has been shown that if the head is
held in the primary position, and a colored cross, with vertical
and horizontal arms (Listing's figure), be located on a vertical
screen with the intersection point at the middle point of the
line joining the points of intersection of the lines of sight of
the eyes with the screen, the shifting of the eyes to any
oblique position when an after-image has been developed
with head kept fixed, results in the following distortions of
the arms of the cross (17):
1. Turning the eyes to the right above or left below
After-image of horizontal arm is turned to the left.
After-image of vertical arm is turned to the right.
2. Turning the eyes to left above or right below
After-image of horizontal arm is turned to the
right.
After-image of vertical arm is turned to the left.
According to Listing and corroborated by Helmholtz, if
the eyes are turned straight above or below, or directly to
right or left, the after-images of horizontal and vertical arms
remain horizontal and vertical.
A much modified form of Helmholtz's apparatus was
362 MILDRED WRING
employed. It consisted essentially of a black strip rotating
at its middle point against a white screen. The after-image
of the black strip could be directed to any angle of any quad-
rant of the field of vision at a fixed distance from the point of
rotation. The torsion of the eye as determined by a torsion
of the after-image was measured by making the strip parallel
to the after-image, and reading the angle of torsion directly.
The results show that there is a torsion for all angles of
oblique fixation, for the right eye to the left and the left eye
to the right, from the point of view of the observer. This is
contrary to Listing, Helmholtz and Sanford in that they deny
torsion for the horizontal and vertical, and Sanford also for
45°. The results show too that the method of after-images
is capable of showing the vertical and horizontal torsion,
which Miss Barnes denied.
EXPERIMENT
Problem. — It was the purpose of this investigation to see
whether or not the long-debated question of torsion of the
eyeball around the sagittal axis could be settled by means of
a very carefully made apparatus which would detect any
rotation ordinarily unobserved by other methods of investi-
gation. To that end, recourse was had to the principle of
the method of Johannes Miiller, namely the observation of a
fixed line upon the eye when that eye is placed in different
positions in the field of regard. The fixed line of the eye was
one of the radial striae of the iris. A description of the ap-
paratus, by which the observation was made, follows.
Apparatus. — The apparatus consists of a large, vertical,
semicirculcar iron arc, one meter in radius, attached solidly
to a heavy brick chimney in the dark room of the psycho-
logical laboratory. By means of a galvanometer, it was
determined that the vibration of the chimney is so slight as
not to overcome the inertia of the apparatus. The iron arc,
which was turned in a lathe to insure smooth surfaces, is
5 cm. wide and 2.5 cm. thick, and is graduated into arcs of
10°. The whole arc rotates in a horizontal direction about
its extremities, but its points of rotation are shifted 7.5 cm.
LAW OF EYE-MOVEMENTS 363
to the right by means of iron plates, to accommodate for the
eccentric position of the telescope which is attached to the
arc. In this way, the vertical cross-hair of the telescope
coincides with the axis of the arc. The upper center of rota-
tion has on it a horizontal, circular protractor, 7.5 cm. in
radius, graduated in degrees, to measure the amount of rota-
tion of the arc. Attached to the right side of the arc is a
transit telescope, with a magnifying power of 34 times, at the
distance from the telescope to the middle point of the arc.
The magnifying power was determined as follows:
At a distance of one meter from the telescope a meter
stick was placed in a vertical position. A rider, made of a
narrowband of paper, 1.5 cm. wide, was freely movable along
the length of the meter stick. The experimenter then looked
at the rider through the telescope with his left eye, and
directly at it with his right eye, and adjusted the rider, so
that the upper edge of it, magnified in the telescope, coincided
with the upper edge of it, unmagnified, as seen by the right
eye. With the images of the left and right eyes thus super-
imposed, the experimenter observed the length of the magni-
fied rider on the unmagnified scale of the meter stick. The
left eye image of the magnified meter stick and right eye
image of the rider, were inhibited. The quotient then of the
magnified length of the rider and its unmagnified length
represents the magnifying power of the telescope.
Magnified length of the rider 51.5 cm.
Unmagnified length of rider 1.5 cm.
Magnifying power 51.5/1.5 = 34.3 times
This method was checked by the method of focal lengths.
The magnifying power = /"//, where F is the focal length of
the objective and / the focal length of the eye-piece. Meas-
urement of the focal lengths of the telescope showed:
Focal length objective, F 26.5 cm.
Focal length eye-piece, / 8 cm.
Magnifying power 26.5/.S =33.1 times
The telescope of the apparatus is detachable and can be
slid along the arc to any desired position and fastened there.
It is directed inward, perpendicular to the arc.
364 MILDRED WRING
Exactly at the middle point, between the two points of
rotation of the arc, suspended by means of an iron support,
extending out from the brick chimney, is a small brass ring,
3 cm. in diameter, permanent in position, at which the obser-
ver places his eye. Also from this same support is attached
a head rest. It is made of an iron band 27 cm. in diameter
surrounding the observer's head at the temples, and supported
by four vertical iron strips adjusted in length vertically, which
at a distance of approximately 12 cm. above their attachment
to the iron band, bend inward to a common point, which is
the center of support. Screws, with flat discs at their ex-
tremities, point inward through the iron band, which are
adjustable to accommodate differences in size of head of
observers. The chin is supported by a U-shaped piece of
iron, swinging from the iron band so that it hangs about in
the plane of the brass ring. It too is adjustable vertically by
means of screws.
The head rest can be set for either right or left eye position.
The observer places his head in the iron band, with the eye
exactly in the center of the ring; the screws are tightened to
fit the head, the band is lowered or raised to accommodate
his height, the chin rest is adjusted likewise and then the head
is so firmly fixed that there is no error from head movements.
The telescope is attached at about its middle point to the
right side of the arc, and is adjustable in three planes, ver-
tically, horizontally, and circularly about its long axis. It is
equipped with vertical and horizontal cross hairs to obtain
accurate settings. A vernier, which reads to o'.5 is attached
to read the circular rotation. It is to be noticed that since it
is necessary practically to have the telescope fastened to one
side of the arc, the points of rotation of the arc have been
shifted an equal amount in the same direction. By dropping
a plumb line from the upper point of rotation, the center of
rotation of the arc was obtained, at which point the center
of the brass ring was placed and marked, by two perpendicu-
larly intersecting threads, the vertical one being determined
by the plumb line. The telescope was then set in a horizontal
position at the zero point on the arc, which is half way between
LAW OF EYE-MOVEMENTS 365
its extremities, and the intersection of the cross-hairs of the
telescope coincides with the intersection of the threads on the
brass ring. In this way the points of rotation of the arc, the
center of the brass ring, and the intersection point of the
cross-hairs of the telescope all lie in a plane, parallel to the
plane of the arc. With a new position of the telescope on
the arc, or a rotation of the arc, the horizontal and vertical
cross-hairs of the telescope still coincided with the horizontal
and cross-threads passing through the center of the brass
ring. This statement must be modified. Since the ap-
paratus is to measure rotation in a vertical plane, perpendicu-
lar to the plane of the arc, the aim was so to construct it that,
with the arc and telescope in a given position, a change in
position of the arc, or of the telescope on the arc, the vertical
and horizontal cross-hairs would still be horizontal and
vertical. By making the axis of rotation of the arc absolutely
vertical, placing the center of the brass ring precisely at the
center of the diameter of this axis, and since the arc was
constructed as true as possible, it was found that the vertical
and horizontal cross-hairs remained so with various positions
of the apparatus. There was but one variation in the ab-
solute trueness of the instrument. When, for instance, the
arc and the telescope were both placed in the zero position
and the vertical cross-hairs set coincident with the vertical
cross-thread of the brass ring, and then the arc was rotated
to right, the thread and cross-hair, though both remained
vertical, were no longer coincident but were laterally dis-
placed. Since they were parallel, by adjusting the screw
controlling the horizontal shift of the telescope, they could
be brought into concidence again. In this operation, the
cross-hair was not rotated from its vertical position, so that
whenever a rotation of the cross-hair was necessary in the
actual experimentation, to cause coincidence, it indicated, not
imperfection in the apparatus, but a rotation of the line which
the cross-hair of the telescope was being set upon. This was
the intended function of the apparatus. This parallel shift
was due to the size of the apparatus and the high magnifying
power of the telescope, but did not amount at a maximum to
366 MILDRED WRING
more than two or three turns of the small screw, controlling
the horizontal adjustment. The same displacement occurred
in a vertical direction but to a much less degree. Two electric
lights served for illumination. One light was attached to the
arc by a jointed arm and could be slid along the length of the
arc. It was used to illuminate the vernier; the other was
suspended from a horizontal arm pivoted to the top of the
head rest, and illuminated the eye of the observer. A frosted
globe was used in this lamp, to lessen the fatigue of the eye,
and to give a more diffuse illumination.
Procedure. — When the observer was seated with his head
in the head rest as indicated above, the arc and telescope
were placed in the primary or zero position; that is, the arc
was swung to a zero position on the dial, and the telescope
set horizontally on the arc at its middle or zero point. The
observer sat with his left eye at the brass ring and the plane
of his face perpendicular to the plane of the arc, so that looking
straight ahead of him, his eye was fixed on the telescope. A
small spot of white paint was placed at the middle point of
the objective to facilitate fixation. Next, the operator ex-
amined the eye through the telescope, and selected on the
iris a line or set of two points upon which to set a cross-hair
of the telescope. The same line was not used through the
whole experiment, inasmuch as it might have an unfavorable
location for observation with certain oblique positions of the
eye. This initial reading on the vernier was taken as the
zero reading. Whenever a new line was taken on the iris,
and with each sitting of the observer, a new zero was taken.
With the telescope at zero position still, the arc was swung
15° to the right. The head of the observer was not moved;
in fact, its position was kept constant throughout a sitting.
The eye, however, was turned horizontally in its orbit to the
new position of the telescope and fixed again on the white
spot. The operator reexamined the eye, set the cross-hair
on the experimental line of the iris, and read the vernier.
The difference between the first and second readings repre-
sented the angle of torsion of the eyeball. Similarly angles of
torsion were obtained for 15° to the left and 30° to both right
LAW OF EYE-MOVEMENTS 367
and left. The arc was then returned to the zero position of
the dial, and the telescope was then moved along the arc to a
position 15° above the zero, or middle point of the arc. The
reading was taken here, and again the arc was swung 15° to
the left and to the right and also 30° to the left and to the right,
and readings taken. The difference of the readings in any
position from the reading of the zero, represented the torsion
for that position. Observations were made upon both right
and left eyes. As summary of the positions of the eye in
which readings were taken is as follows:
Telescope. Arc.
I. 50° above 50° left— if left— zero— if right— 30° right
II. 75° above 50° left — 75° left — zero — 75° right — 30° right
III. Primary position 30° left — 75° left — zero — 75° right — 30° right
IV. 75° below 30° left — 75° left — zero — 75° right — 30° right
V. 30° below 30° left— if left— zero— if right— 30° right
In spite of the accuracy of the instrument, certain un-
avoidable errors arose in taking data. These were:
1. Certain reflex pupillary expansions and contractions
even with a constant intensity of illumination. This changed
the size of the iris and therefore the direction of the line set
upon, in spite of every care to have the line a true radius of the
iris.
2. Nystagmoid movements of the eyeball, entirely involun-
tary, due to reflex contractions of the eye-muscles, as a result
of continued fixation.
3. A marked change in the pattern of the iris, owing to
torsion. Sometimes it was difficult, in a new position of the
telescope, to recognize the line to be set upon, owing to certain
twistings in the pattern, which, while not very extended, yet
interfered markedly with an accurate setting of the cross-hair.
Results. — I. Each number in Table I. (a) and (b) repre-
sents the average of five observations in that position.
2. In Table I., the column headed "Position," the sign +
before a parallel indicates a position of the telescope above
the zero parallel on the arc, the sign — , a position below the
zero parallel on the arc; the sign + before an angle in the
subgroup ( + 30°, +15°, o, —15°, —30°) underneath a parallel,
indicates a position of the arc to the right of the zero on the
M t^oo ^^o ^ *•; "f* *> "t" 9^9 *? 9s
°o °o °o °o °o °o °o °o °o °o °o°o °o°o
9 *
N M 00 VO
HI M M M
°o *b °o °o °o
M C* HI HI C*
°o °o °o °o
oo ON >J-> O VO
Oo0o0o0o°o
vo O
* *
b* TJ-OO O O
1-1 LD rj- -^-
TJ- "*• ON qvqoNoo -^vq q "f -*
N ••$• O ON OO M N "t^^O t/i tr> vi Q
LO Q HI 10 ur> N
C» Tj-VO N-«1- NCOMMHI OW OO r*^Tj-O»-<'«!j- ONLTiN
1 ++ 1 1 +++ II + 1 ++ 1 + 1 1 1 +++++
OO OO VO -^- O c^O^VOO OOVO OO
"J-00 O 00 00
++ 1 i i + 1 i i i ii ii i i i i i i i i +
OOOOON NOO qo
voTj-
ooooo
N ,prp
O
N HI O O VO O
I I ++ M
1 1 1 + +
vb *lr)Vi*bN*b
Oo0o0o°o0o Oo0oo0o°o °o°o °0°o Oo0o0o0o°o Oo0o0o0o°o
°0 °0 °0 °0 °0 °0 °0 °0 °0 °0 °0 0 00 OOOOO 0 °0 0 °0 °0
4^vMOOMi/-> vot^-^voco NOOr^t^
HITt-l-HCOLO MHlOc^M l_,Mr<N
^ 9
CO *to
++ 1 i i i i i i
c^NN^l-Tj- ONOOOOLo
I I I I I ++ I I I
HHTj-
ooooo
i-< -
HHC^Tj-ThU
ooooo
b oo
MHH WHH
ooOoo
oooo
00000** «/i
lOTj-N-^-hH
ooooo
++++ 1 iiii ii ii i i + i++
oo oo q *o N q ONVO oo •«*• oo N vq ^ -<j-oq q q q
*O *ir>*O ^C^OO *O "l^*ON*'''f*Hi OO *O %J-">t> *|O"N "V»VO ^Q OO "HI u->VO f>
O
O O VJ o O
ooooo
ooooo
i i M +++ M I+M IIM+ iii
t-iOOOOO t-iOooOO l-iOOOOO £oOOOO S-iOOO
O^rOHi i-c «<-) o, c<-j HI HI r^> CXi m HH HicoOnCOHi Mc^D<r<^Hi HH
I I ° ++ I I <k++ II
LAW OF EYE-MOVEMENTS
369
dial, the sign — before an angle indicates a position to the
left.
parallel.
indicates a position where the
For instance __ o
telescope has been placed 15° above the zero or middle point
of the arc, and the arc rotated 30° to the left of the zero on
the dial. These signs are taken from the point of view of
the experimenter, who looks through the telescope.
3. In the column headed "Torsion," the sign + in-
dicates torsion of the eye to the right from the point of view
of the experimenter; the sign — indicates torsion to the left.
Conclusion. — I. The results are to some extent irregular,
both as to the amount of torsion and direction.
(a) As the eye travels from the primary position to scu-
cessively more oblique positions, there seems to be a tendency
for the angle of torsion to increase, but some exceptions occur.
Likewise a movement of one eye to a given position does not
give the same amount of torsion as the movement of the other
eye through the same angle. Similarly, the sense of the torsion
does not remain constant for one eye in all cases, and the
torsions for the two eyes for movements through similar
angles is not always opposite or always the same.
(b) There is a predominance of torsion to the right of the
right eye, and to the left of the left eye, from the standpoint
of the experimenter. Out of 24 positions for each eye away
from the primary position, for each observer we find the
following:
Torsion of Right Eye
Torsion of Left Eye
To the Left
To the Right
To the Left
To the Right
I
2
3
9 cases
20 cases
10 cases
15 cases
4 cases
14 cases
19 cases
17 cases
20 cases
5 cases
7 cases
4 cases
It will be seen that for the left eye, the great predominance of
torsion is to the left for all three observers, and that for the
right eye, the predominance is to the right for two of the three
observers.
II. The most definite conclusions consist then mainly in
the two following observations:
37° MILDRED WRING
(a) A tendency for the torsion to increase as the eye
passes from the primary position to successively more oblique
positions, including horizontal, vertical and 45° meridian.
(b) A predominance of torsion of the left eye to the left,
and of the right eye to the right, from the point of view of the
experimenter.
LITERATURE
1. JOHANNES MULLER. Zur vergleichenden Physiologic des Gesichtssinns, 1826,
p. 254.
2. ALEXANDER HUECK. Die Achsendrehung des Auges, 1838.
3. A. BUROW. Beitrage zur Physiologic des Auges, 1841, p. 8.
4. A. W. VOLKMANN. Wagner's Handworterbuch der Physiologic, 1846, Band 3,
Abt. I., p. 273.
5. THEODOR RUETE. Lehrbuch der Ophthalmologie, 1846, p. 14.
6. G. VALENTIN. Lehrbuch der Physiologic, II. (2), 1846, p. 32.
7. F. C. BONDERS. Hollandische Beitrage zu dem anat. und phys. Wissenschaften,
I., 1848, pp. 105-145, 384-386.
8. A. FICK. Ztschr. fur rat. Medizin, IV., 1854, p. 101.
9. G. MEISSNER. Beitrage zur Physiologic des Sehorgans, 1854.
10. G. MEISSNER. Graefe's Arch.f. Ophthalmologie, II. (i), 1855, pp. 1-123.
11. THEODOR RUETE. Lehrbuch der Ophthalmologie, I., 1855, p. 37.
12. A. FICK. Moleschott's Untersuchungen z. Naturlehre der Menschen, V., 1858,
P- 193-
13. G. MEISSNER. Ztschr. fur rat. Med. (3), VIII., 1860, p. I.
14. W. WUNDT. Archiv fur Ophthalmologie, VIII. (2), 1862, pp. 16-17.
15. H. HELMHOLTZ. Archiv fur Ophthalmologie, IX., 1863, pp. 153-214.
16. A. W. VOLKMANN. Physiologische Unters. im Gebiete der Optik (2), 1864, pp.
199-240.
17. H. HELMHOLTZ. Handbuch der Physiologischen Optik, Ed. 2, 1866, pp. 613-669.
1 8. EWALD HERING. Die Lehre vom Binocularem Sehen, 1868, pp. 83-92.
19. H. AUBERT. Graefe-Saemische Handbudi der gesammten Augenheilkunde, II.
(2), 1876.
20. JOSEPH LE CONTE. Sight, 1881, pp. 185-212.
21. A. MEINONG. Ztschr. fur Psych, u. Physiol, XVII., 1898, p. 161.
22. E. SANFORD. Experimental Psychology, Appendix, 1898.
23. BERNICE BARNES. American J. Psych., Vol. 16, 1905, p. 199.
24. A. WICHODZEW. Ztschr. f. Sinnesphysiologie, II., 1912, p. 394.
VARIABILITY^ IN PERFORMANCE DURING BRIEF
PERIODS OF WORK
BY A. T. POFFENBERGER, JR., and GLADYS G. TALLMAN
Columbia University
Numerous researches have been carried on to show the
effects of working for long periods of time or with extremely
difficult tasks. The effects, known as fatigue, may show
themselves in a decrease in the quantity or quality of the
product. Analogous changes in the character of the per-
formance, which have scarcely warranted the name of fatigue
as the term is ordinarily used, appear in extremely brief
periods of mental work. For instance, Professor Woodworth
in a recent paper,1 showed some changes in speed of perform-
ance which take place during the naming of only ten simple
colors. The present report contains a brief series of records
showing a variation in performance in tasks, none of which
lasted as long as one minute.
In some of the previous work on fatigue one finds references
to the effects of brief work periods, although in few of them
did the matter receive much attention. Voss2 made detailed
studies of an hour's work in addition, taking time for each
separate problem. He concludes that practice tends to
increase the number of rapid additions and fatigue the
number of slow additions. The procedure followed made it
impossible to separate the practice and fatigue effects for
the short periods measured. Hylan and Kraepelin3 measured
the variation in adding one-place numbers in five-minute
periods. They concluded that mental work lasting for only
1 An unpublished paper read before the New York Branch of the American
Psychological Association at Princeton, N. J., Feb. 23, 1914, entitled 'The Work Curve
for Short Periods of Intense Application.'
2 George von Voss, 'Ueber die Schwankungen der geistigen Arbeitsleitung,' Psychol.
Arbeit., Vol. 2, pp. 399~449-
3 Hylan, J. P., and Kraepelin, E., 'Ueber die Wirkung kurzer Arbeitszeiten,'
Psychol. Arbeit., Vol. 4, pp. 454-495.
371
372 A. T. POFFENBERGER AND G. TOLLMAN
five minutes produced an appreciable amount of practice and
fatigue. The ratio of the two factors differed for different
individuals. In the work of Arai1 on the multiplication of a
series of two four-place numbers the time for each problem
was measured. In the author's conclusions (p. 93) we note
that "the difference between the time taken for one example
and that taken for another is greater in the second half than
in the first half of the curve. This fact together with the
evidence of introspection of the subject suggests that fatigue
not only causes decrease of efficiency, but also loss of the
subject's control over herself. For this reason the subject
tends to occasionally relax her original standard of effort."
In this experiment the subject had passed the stage where
practice was an important factor.
In the experiment to be reported two time records were
taken for each task, one at the completion of the first half
and the other at the end of the task. The time was taken
with a stop watch in units of one fifth of a second. From
these two records one may compare the performance during
the first and second halves of the work. Speed was the only
variable factor, for the experimenter announced all errors and
these were corrected by the subject. It is to be noted further
that the subjects were trained in all of the tests, and that
practice effect was thus practically eliminated. If this were
not the case it is quite possible that the improvement by
practice within a single test might have clouded the results.
An intensive study was made upon two subjects, each test
being repeated from 60 to 70 times after the preliminary
practice.
Four tests were used, two of them being taken from the
Woodworth and Wells2 monograph on association tests.
These were the color naming test and the number checking
test, called the cancellation test in this report. Only half
of the regular number checking blank was used for one test,
so that the halves measured were really fourths of the whole
1Arai, Tsuru, 'Mental Fatigue,' Teachers College, Columbia University Con-
tributions to Education, No. 54, 1912.
* Woodworth, R. S. and Wells, F. L., 'Association Tests,' Psychological Mono-
graph, 1911, No. 57.
VARIABILITY IN PERFORMANCE 373
blank. The numbers 3 and 5 were used alternately for can-
cellation. The third test was the opposites test and con-
sisted of 50 words, the opposites of which were to be named.
They were about equal in difficulty to the moderately difficult
series of Woodworth and Wells. The fourth test was the
addition test and consisted in adding 17 to each of 50 two-
place numbers, ranging from 20 to 80 with all of the com-
binations containing zero being omitted. In every case the
test material was divided into two parts by a line, so that the
subject could see the limits of the first and second half. The
subject's work, however, was continuous from the beginning
to the end of the test.
The tests were repeated approximately five times each
day, with an interval of several hours between any two, thus
avoiding cumulative fatigue. All of these tests involve the
process of association and the connection between situation
and response was in each case partially established in the
preliminary tests. It is to be remembered that the test
material was always the same, the only difference being the
order in which the stimuli were presented. There were four
possible changes of order in the color naming test, by rotating
the color blank 90 degrees at each trial; four in the number
checking test, by using different halves of the test and either
one of the numbers 3 or 5 for cancellation; ten in the opposites
and addition tests by means of ten cards containing the same
numbers in a random position on each.
The data of the experiment are presented in the accom-
panying table. The third column of the table shows the
number of cases from which the calculations are made for
each test. The fourth column gives the average speed for
the first half of the test in terms of seconds. The fifth column
gives the gross differences between the first and second halves.
In every case the first half was subtracted from the second
half, so that the absence of minus signs indicates greater speed
in the first half. The sixth column shows the reliability of
these differences in terms of the probable error of the differ-
ence. The seventh column shows the difference in terms of
per cent, of the time of the first half. The column of gross
374
A. T. POFFENBERGER AND G. TALLMAN
differences shows that in every test and for both subjects the
first half is quicker than the second half, while the following
RELATION BETWEEN SPEED OF PERFORMANCE IN THE FIRST AND SECOND HALF OF
THE TESTS.
(In every case the first half is subtracted from the second half, so that absence of
minus signs indicates that the first half is faster.)
Test
Subject
No. Cases
ist Half
Gross Diff.
P. E.
Per Cent.
Diff.
Opposites
T
CQ
16.2
O I
II 7
Addition
P
T
6?
7O
12.9
28 t:
•7
•7
O.I
O 1
13-2
ii 6
Color Naming
P
T
%
18.4
2O.O
.1
.0
0.2
O.2
6.0
o.r
Cancellation
P
T
$
61
17.0
2^.8
.0
2.2
O.2
O.2
5-9
0.2
P
60
20.3
0.7
0.2
34
column shows that the differences are reliable. The data
when examined in detail and when treated in several other
ways, show only one interesting exception. When the tests
are grouped according to the time of day they were performed,
namely 9.30 A. M., 1.30, 3.30 and 5.30 P. M., it appears that
in the case of the cancellation test and at the 9.30 period both
subjects were on the average slower in the first half than in
the second half of the test. No explanation for this peculiar
difference is suggested. The data treated in this fashion
show only one other case where the first half is slower and here
the negative quantity is less than its probable error.
The last column in the table shows the per cent, of differ-
ence in favor of the first half. It should be considered merely
as indicating differences between halves in the tests, for only
two subjects are not sufficient to establish differences of this
type among the tests used. If the tests are grouped, one can
say roughly that in speed of performance subject T was about
10 per cent, less efficient, and subject P about 7 per cent, less
efficient in the second thirty seconds of the work than in
the first thirty seconds.
A series of ten questions was given to each subject at
the end of the series of tests to determine if introspection
could throw any light upon the results. For instance, the
VARIABILITY IN PERFORMANCE 375
subjects were asked if they noticed any difference in speed,
ease of performance, number of errors, etc., in different parts
of the tests, and if so whether they could account for such
changes. A separate series of questions was given for each of
the four tests. Both subjects felt that they made more errors,
and hesitated more in the second half than in the first. In
both subjects there was consciousness of being slow with a
feeling of inability to make the mind work any faster. The
speed was probably judged by the number of errors and hesita-
tions that occurred, according to the answers to some of the
questions. The mistakes were attributed not so much to
lapses of attention as to just getting tired. Neither subject
reached the stage, during the course of the experiment, where
the task became automatic, i. <?., where he no longer felt as
though he were putting forth a great deal of energy.
In drawing conclusions from these records it is necessary
to keep the specific conditions in mind: (i) The test material
was always the same, except for the difference in arrangement.
(2) The process involved in the reaction to the situations was
mainly the recall of previously formed associations. (Formed
or at least strengthened during the preliminary tests.) (3)
There was practically no practice effect within any test, on
account of the preliminary tests, and the use of the same ma-
terial throughout. This is a condition not found in mental
fatigue tests where the records are made at frequent intervals.
(4) The subjects worked at their best speed, with no rest period
between the first and second halves of the test. (5) There
was no rest interval between the performance of separate
items of the tests, such as the relaxation which may take
place at the end of each addition or multiplication problem
usually employed for fatigue work, especially where the time
required for each problem is the unit of measure.
This falling off in performance within such a short period
in the case of two subjects in the mental tests described,
suggests the importance of devising simple mental tests,
which shall approximate the classical ergograph test as
performed upon a trained subject, in its simplicity, forced
regularity of response (e. g., metronome beat in the ergograph
376 A. T. POFFENBERGER AND G. T4LLM4N
test), and in the complete record obtained. For it is probable
that mental fatigue is not so rare as is sometimes supposed,
but that the repair process is so rapid compared with muscle
repair, that as work is usually done, the loss may be com-
pensated for during brief intervals of relaxation.
THE STANDARDIZATION OF KNOX'S CUBE TEST
BY RUDOLF PINTNER
Ohio State University
This paper deals with the results obtained with KnoxV
Cube Test, one of the performance tests used by him for the
mental classification of immigrants at Ellis Island. The test
has been given to 867 normal children and a few adults, and
to 463 feeble-minded individuals. An attempt has been made
to enlarge the scope of the test and to standardize it a little
more adequately. The test appeared to me, after first seeing
it applied, to be an excellent one in many ways. Without
attempting to enter into a useless discussion as to the actual
mental processes involved, we may say in a general way that
it depends largely upon imitation, at the same time affording
every opportunity for other factors involving intelligence to
assert themselves.
The Method of Procedure. — Five blocks are required. The
Binet black cubes were used, but any other cubes of about the
same size would be satisfactory, provided they are all of the
same color. Four of these are placed on the table in front of
the subject at a distance of about two inches apart. The
examiner holds the fifth cube in his hand. He says to the
subject: "Watch carefully, and then do as I do." He then
taps the blocks with the fifth cube in a certain definite order
and at a certain definite rate (about one tap per second),
always beginning with the cube at the child's left or the ex-
aminer's right if he is facing the child. He then lays the fifth
cube down in front of the child equidistant between the third
and fourth cube, but nearer to the child, and says: "Do that."
These oral directions, "Watch carefully and then do as I
do," and "Do that," were given to almost all of our subjects.
1 Knox, Howard A., The Journal of the American Medical Association, March 7,
1914, Vol. LXIL, pp. 741-747-
377
378
RUDOLF P1NTNER
I do not believe this is necessary and I am sure exactly the
same results would have been obtained by saying nothing.
This is borne out by some of the subjects who did not under-
stand English and by deaf children and by some children who
seemed too young to understand such verbal directions. In
these cases all that was necessary was to make some gesture
indicating that the fifth block was to be picked up and the
others to be touched.
The first line of the test, as will be noted presently, is so
simple that even although the subject does not know while
watching the examiner that he will be required to do the same
thing, he can easily remember and imitate what has been
done. There are of course innumerable combinations in
which one can tap four blocks, if one is not limited to touching
each cube only once. If we number the blocks the different
combinations will be readily understood, and the following
diagram should make absolutely clear their position with
regard to the subject and the examiner (if he is facing the
subject).
Subject
Examiner
The following twelve combinations or lines have been used
in this experiment. Number I always refers to the block on
the left-hand side of the child.
A lit
ie
[234.
X
]
[2343
Y
2 3 4. 2
ff
•$24,
c
41 2
r>
A 2 •?
F
32 4, •?
F
4-224.
G
•2124.
H
4. -2 I 2 4.
J
•2 2 4 I 3
1
4 2 3 4. I
STANDARDIZATION OF TESTS 379
Knox uses only five lines. The advantages of my extension to
twelve will, 1 think, be apparent in the light of the results.
The lines are taken up in the order of sequence in which they
appear above. It will be noted that this is roughly an order
of increasing difficulty. A line is never repeated, not even if
the subject begins any line, the A line included, at the
wrong end. This is an error that we have marked W.E.
(wrong end) and will discuss later. The subject is never
corrected, but is allowed to do a line exactly as he chooses.
No hint must be given the subject as to whether he is doing a
line right or wrong. Some children pause and look up at the
examiner waiting for a cue as to the next move. The subject
is encouraged to do his best and is told that he is doing well,
regardless of his actual accomplishment. Each line is marked
plus or minus according as the subject does it correctly or
incorrectly. If the subject corrects a wrong move, thereby
making the whole line correct, he is credited with a plus.
The numbers of the moves must not be shown to the child
and the examiner must not count aloud or indicate in any way
that he is counting. This precaution is very necessary with
intelligent older children and with adults. Nothing in fact
is said about counting and the subject is left free to pursue
that method if he has the intelligence to think of it. This
will of course show in a greater number of lines passed cor-
rectly and therefore the subject receives credit for his in-
telligent adaptation. We asked every subject who did very
well, i. e., succeeded in almost all of the lines, how he did it.
Some replied at once that they were counting. Others were
uncertain. We suggested to the latter that they were count-
ing and if they assented, we asked them and also those who
told us they were counting, to count aloud while we repeated
one of the longer lines or a new combination of moves. The
examiner can then tell whether it is the kind of counting
that will help in remembering the moves. Some children
who said they remembered the moves by counting were
found simply to be saying to themselves, one, two, three,
four, etc., every time. This, to be sure, would help them to
380 RUDOLF PINTNER
remember the number of moves but would not help them to
remember the position of the blocks touched. The other kind
of counting which assigns constant numbers or letters to
each block was found to be very rare. We found it only in
about twenty cases out of our 867 individuals. This kind of
counting did not seem to occur to the majority of adults who
were tested by us.
In many respects this method of giving the test differs
radically from that of Knox, as far as can be determined from
the very brief description of the test in the article cited above.
Knox allows a repetition of a line on the part of the examiner,
in contrast to my procedure where a line is never repeated.
For the first five of his lines he gives three trials ' if neces-
sary,' and for his last and most difficult line involving six
moves he allows five trials. If the subject fails he is evidently
shown the moves over again to the extent of three or five times
if necessary. The drawback of this method seems to lie in
the fact that a varying number of repetitions of any line will
cause unequal practice effect. For example, the subject that
fails twice on the second line and passes on the third trial will
have made up to that point four responses to the test, whereas
the subject that passes the second line at the first trial will
have made only two responses. Each will then start the
third line with different degrees of familiarity with the situa-
tion, and it is possible that the first subject may have gained
an unfair advantage over the second, even although the first
subject is not doing as well as the second as shown by his
failures. A fair comparison of their performances in the
succeeding lines will nevertheless be impossible, however
much the one may be superior to the other. I think that any
difficulty in this respect that may exist can be adequately
overcome by extending the number of lines, as has been done,
by following always the same sequence and by rigidly adhering
o the rule never to repeat a line.
The Subjects. — The subjects included 867 presumably
normal individuals. These were in the main pupils in the
ordinary grade schools of about four or five different schools
STANDARDIZATION OF TESTS 381
in Columbus, and some from a junior high school. Most of
the five-year-olds were kindergarten children and those
below that age were examined in day nurseries and settlement
houses. About half of the adults tested were university
students. Four hundred and sixty-three feeble-minded in-
dividuals were also tested. The vast majority of these were
inmates of an institution for the feeble-minded, but there are
also included in this number several feeble-minded children
that were met with in the juvenile court or in school. No
systematic attempt was made to exclude all cases of suspected
feeble-mindedness from the data for normal children. This
would have involved giving lengthy tests to some hundreds of
children in the public schools. Any child who was obviously
feeble-minded was excluded from the normal group and if his
Binet age had been determined he was included among the
feeble-minded.
The tests were not all given by the writer himself,1 but the
technic of this test is so simple that uniformity in giving it is
very easily attained. I do not think that any possible error
from this source would materially affect the results.
The number of normal children tested at each age is given
in Table I. below, and the feeble-minded in Table IV. The
normal children are grouped according to chronological age
and the feeble-minded according to Binet age. It will be
seen that the number for each age is not uniform, but a suf-
ficiently large number between the ages of five and sixteen were
obtained for each age. The usual difficulties were encountered
in getting children below five and above sixteen years of age.
The feeble-minded differed chronologically a great deal, some
of them were adults and others merely children.
Tabulation of the Data. — The actual recording of the results
while giving the tests is perhaps best done by making out some
such blank as is shown below, which is a copy of some of the
actual data.
1 The writer wishes here to acknowledge the generous help given him in this work
by Mr. Donald G. Paterson, graduate assistant in the department of psychology. He
is also glad to acknowledge the assistance rendered by Miss M. Anderson and Miss A.
Beekman, advanced students in the same department.
382
RUDOLF P1NTNER
Name
Age
Grade
A
X
Y
B
C
D
E
*•
G
H
I
J
Catherine M. . .
Paul C
12
6
55
lA
+
+
4-
+
+
—
+
+
+
+
—
—
+
—
—
Matilda S
14.
7A
+
+
-f
+
+
+
+
-f
__
__
Rosie S
c
Kindg.
W.E.
+
+
__
During the collection of results covering in all about
thirteen hundred cases the following device was employed
to keep an oversight of the results as they came to hand and
also to afford some clue as to when sufficient results had been
collected. As the results came in they were recorded on
squared paper and a curve was slowly built up. This was done
for each separate line and for certain groups of lines. Fig. I
shows the results for two lines correct out of the BCD lines.
The numbers along the abscissa represent the ages. A
mark or dot above the line represents a correct response and a
similar mark below the line represents a failure. The
figure shows the building up process at the stage of completion.
The figures to the left of each column represent the number
of cases, the figures to the right are the percentages of correct
responses for each age at different stages in the growth of the
curve. When each fresh group of results was added, the
percentage of correct responses was calculated and noted on
the growing curve. This device for recording the results was
found to add a much greater interest in the work than could
have been attained by waiting until all the data had been
collected, and it also shows the worker how his results are
developing. If the percentages are calculated every now and
then, he can see whether new results that are coming in are
adding anything new to his work or are merely confirming the
results that have already been obtained. If the percentage
fluctuates a great deal, it is a sign that more results are re-
quired; if it remains more or less stationary for some time, it
may be assumed that additional results are not likely to affect
the shape of the curve. From Fig. I it will be seen that the
percentages for most ages remained more or less constant.
Some slight fluctuation is seen at age eight, where the per-
centage drops from 68 to 65 and then rises steadily to 74.
STANDARDIZATION OF TESTS
383
*
lUnminuimiiuiiimt
e»—
tso
a
IfifttimilHHHMI
H
HfllMlllll
O
J-
Illlllll
oo
Him
intuiuiiit
|
I
S
3 s
8 ^
if
- g
« &
fiiimiuunmmi
UllinilllllllfMllllHUIMtUIHV
<*>
iM|lll1IH|lUllllfUIUUi(U(<lUlHI|IU
ffllfHll
I
384 RUDOLF PINTNER
Still more fluctuation is seen at age seven, where the per-
centage at first drops from 62 to 58 and then rises slowly to
69. From the point of view of standarization this fluctuation
of the percentage becomes important. We obviously cannot
say that a test is a seven- or eight-year-old test, as the case
may be, until we are certain that the addition of subsequent
data is not likely to affect our percentages. At one stage of
our curve it might have seemed that this particular com-
bination of lines would have made a good eight-year-old test,
since 68 per cent, of the eight-year-olds passed it and only 57
per cent, of the seven-year-olds. As the other results were
added, however, it became obvious that a larger and larger
majority of the seven-year-olds were able to accomplish it,
practically as large a majority as with the eight-year-olds,
and since this percentage remained more or less constant for
some time, it may be safe to assume that this group of lines
(to do any two correct out of the BCD lines) is a fair seven-
year-old test.
In a similar manner curves were built up for all lines of the
test and for various combinations of lines, and the compara-
tive lack of fluctuation of the percentages seemed to show that
additional data would not radically alter the results already
obtained, at least with ages five to sixteen inclusive.
Standardization. — After the results had all been collected
in this manner, curves for each line were drawn to show the
percentage of correct responses for each age. These curves
are shown in Fig. 2. The curves show how the lines compare
with each other in difficulty. There are four groups of three
lines each which are, as we mentioned before, about equal in
difficulty. It will be seen also that most of the curves are more
or less irregular and do not show any very decided increase
from one age to the next, and therefore are not very satis-
factory from the point of view of standardization. The actual
percentages from which these curves have been drawn,
together with the total number of children tested at each
age, are shown in Table I.
FIG. 2. Percentage of Passes at Each Age for the Twelve Different Lines of the
Test.
386
RUDOLF PINTNER
TABLE I.
NORMAL CHILDREN. PERCENTAGE CORRECT FOR EACH LINE OF THE TEST
Chronological
Age
Number
Tested
Lines of the Test
^
X
Y
B
c
D
E
F
G
H
7
J
3
II
54
o
0
9
0
IO
0
0
0
0
O
o
4
18
67
22
II
5
17
6
O
0
O
0
O
0
5
56
80
57
36
23
50
ii
4
5
O
0
O
0
6
59
92
74
72
44
63
37
2
H
17
2
0
2
7
67
97
90
81
54 66
69
IO
24
16
O
I
7
8
70
99
95
95
66
66
74
23
21
27
0
6
9
9
10
90
88
99
ICO
96
92
89
90
61
73
64
68
67
84
31
34
Ii
32
49
3
4
6
8
8
4
ii
77
ICO
95
93
73
7i
84
47
35
38
2
ii
18
12
82
99
89
96
81
76
9i
56
24
40
2
15
5
13
76
100
96
96
80 77
88
4i
32
51
9
15
7
14
77
99
97
97
88
82
96
58
26
56
8
26
J7
IS
44
100
95
95
84 89
89
55
21
43
7
27
25
16
30
100
100
100
90
80
IOO
53
40
57
10
23
17
Adult
22
100
98
IOO
9i
77
9i
68
41
50
9
50
38
This brings us to the question of standardization. If we
wish to include these tests in any scale of mental measurement,
it is necessary to decide what line or lines of the test we should
be justified in demanding that a child should pass at certain
definite years in his chronological development. It has been
arbitrarily assumed by some writers that a percentage of 75
or more correct responses is necessary before we are warranted
in placing a test at a certain age. This seems to be taken for
granted by Bobertag, Kuhlmann, Goddard and others.
Stern1 emphasises the 75 per cent, standard, without however
disregarding the amount of advance at each age. Rogers
and Mclntyre2 hold to the 75 per cent, basis. They
say, "The standard for a pass at a given age should be de-
termined on the basis of actual experience; our own results
agree with those of Bobertag and Goddard, as about 70 or
75 percent.," and further on, "A test was considered properly
assigned to a given age when at least 70 per cent, of the child-
ren of that age were able to pass it," which shows a little
weakening from the rigid 75 per cent, standard.
1 Stern, 'The Psychological Methods of Testing Intelligence,' Trans, by Whipple.
Educational Psychology Monographs, No. 13.
2 Rogers and Mclntyre, 'The Binet Simon Tests,' The British Journal of Psy
chology, Vol. VII., No. 3, October, 1914.
STANDARDIZATION OF TESTS 387
Bobertag,1 discusses the matter at length and decides for
the 75 per cent, standard. His discussion is undoubtedly
the most exhaustive that has come within the notice of the
writer. He shows the relation of this question to the normal
distribution of any ability and makes a comparison with
pupils' grades as given by teachers.
On the other hand Binet himself nowhere seems to have
been very dogmatic on this point. As far as one can discover
he seems to have considered a test standardized if passed by
from 60 to 90 per cent, of the individuals tested. Similarly
Terman and Childs2 feel it impossible to adhere rigidly to the
75 per cent, standard. They say that two thirds ought to
pass a given test, but they lay more stress upon "a sharp rise
in ability from the year before." It may be interesting to
note that the latter workers cited, Binet and Terman and
Childs, have all been actively engaged in adding new tests,
whereas the former are more especially those who have merely
worked with the tests. It would seem to suggest that in
actual practice there is some difficulty in arriving at the 75
per cent, standard.
If we now turn to actual results as shown in various stand-
ardizations of tests, we see this practical difficulty. In
GoddardV results we find two or three tests that have been
retained in a given age where less than 75 per cent, have
passed. In BobertagV results we have numerous instances
where the number of passes is not up to the 75 per cent,
standard, although it must be remembered that he is merely
trying out the tests and not attempting to standardize them.
Bobertag, 'Ueber Intelligenzpriifungen,' Zeitschr. f. angewandie Psychologic,
Vol. 6, 1912, p. 495.
2 Terman and Childs, 'A Tentative Revision and Extension of the Binet-Simon
Measuring Scale of Intelligence,' Journal of Educational Psychology, Vol. 3, Nos. 2 to 5,
1912; and Terman, 'Suggestions for Revising, Extending and Supplementing the Binet
Intelligence Tests,' Journal of Psycho- A 'sthenics, Vol. XVIII., No. I, September, 1914,
p. 20.
8 Goddard, 'Two Thousand Children measured by the Binet Measuring Scale of
Intelligence,' The Pedagogical Seminary, Vol. 18, June, 1911, p. 232.
* Bobertag, 'Ueber Intelligenzpriifungen,' Zeitschr. f. angew. Psychologic, Vol. 5,
1911, p. 105.
388 RUDOLF PINTNER
In Winch's1 recent re-standardization of the Binet Scale for
English children a very high percentage of passes is required
before the test is admitted to a given age. But the actual
numbers tested are not shown clearly and it is difficult to
believe that in so many cases 100 per cent, passed the tests.
Tests are shifted from one year to another rather too freely
in the opinion of the present writer. If 42 per cent, pass a
given test at six years, and 52 per cent, at seven, we are not
warranted in assuming that the test is an eight-year-old test
without testing any eight-year-old children. This seems to
have been done.
From the foregoing facts and from my own experience,
I am inclined to believe that it is impossible to lay down a
definite percentage for the standardization of a test. It may
be that theoretically about 75 per cent, should pass a given
test, and probably the greater number of children we test the
nearer to this theoretical standard we may attain. In actual
practice, however, with large groups of unselected individuals
this 75 per cent, standard is difficult to obtain, and for the
practical placing of a test at a given age the crucial point
seems to be the more or less sudden rise in ability from one
age to another. We must require about 60 per cent, passes,
but beyond that the best age for placing a test will depend
upon the shape of the curve showing the percentage of passes
at each age. It is significant for us to know at what age the
ability of the child for a special response arises. If we were
to adhere strictly to the 75 per cent, standard, we might place
a test at a given age where 76 per cent, of the children pass,
and where 71 per cent, of the lower age also pass. We would
then be giving 71 per cent, of a lower age credit for a test of
a higher age. For example, on curve C Fig. 2, we have 76
per cent, of the twelve-year-olds passing the C line, and 71
per cent, of the eleven-year-olds. It would seem to me to be
obviously wrong to give the 71 per cent, eleven-year-olds
credit for a twelve-year-old test standardized by this 75 per
cent, procedure.
1 Winch, 'Binet's Mental Tests; What They Are and What We Can Do with
Them,' Child Study, Vol. VII., Nos. i to 8, 1914.
STANDARDIZATION OF TESTS 389
It will now be clear what I meant by saying above that
the curves in Fig. 2 are not satisfactory from the point of view
of standardization. The curves rise gradually and many of
them do not show any sudden rise that might warrant us in
placing the test at any special age. .From my point of view
we must look for some other way to make our results useful
as actual tests. If we group our lines and mass the results
for various combinations of lines, we get the curves shown in
Figs. 3 and 4.
Curve i XY shows the percentage of those who passed
one correct of the X or Y lines; curve I BCD shows those who
accomplished one line correctly out of the B, C or D lines;
curve 2 BCD those who did two of the B, C or D lines; curve
I EFG those who did one of the E, F or G lines; curve 3 BCD
those who did three of the B, C and D lines, i. e., all three
lines correctly; curve 2 EFG those who did two of the E, F or
G lines; and curve 2 EFGHIJ those who did two out of the
six lines from E to /. All these curves are much more satis-
factory from the point of view of standardization with the
exception of the curve 2 EFG. From this group of lines
(2 EFG) it is impossible to get a test, at least under the age
of sixteen, because the percentage nowhere rises above 50,
and this is too low for the standardization of a test. Other
combinations of the E, F and G lines were not drawn, since it
is obvious that three correct out of the EFG lines would show
still smaller percentages at all ages. Similarly the H, I and/
lines proved themselves too difficult. The highest percentage
of passes for the H line is 10 per cent, at sixteen; for the / line
27 per cent, at fifteen; and for the / line 25 per cent, at fifteen.
Combinations of these lines would obviously not rise above
50 per cent, and so they must be discarded at least as tests
suitable for ages below sixteen.
Discarding the 2 EFG lines, we have six curves that seem
to show all the characteristics essential for the standardi-
zation of a test. Curve i XY shows a marked rise between
ages four and five, from 22 per cent, to 66 per cent, and from
that place onwards to 84 per cent., 95 per cent, and then re-
f S <f 10 II 12. IS
<fy.3lf-St>}8y/OU
FIG. 3. Percentage of Passes for Different Combinations4 of the Lines.
STANDARDIZATION OF TESTS
39'
mains in the nineties until 100 per cent, is reached at age
fourteen. This means that the vast majority of children of
five years and over are able to do correctly any one line of the
t*
€0
7°
30
20
to
2EF6H1J.
Op. 6 / 8 f /O II IZ /3 tif- IS «
FIG. 4. Percentage of Passes for Different Combinations of the Lines.
X and Y lines, whereas very few of those below five years can
do this. Curve I BCD shows a marked rise between five
and six years, from 56 to 83 per cent. Only about half of the
392 RUDOLF PINTNER
five-year-olds can do one line of the BCD lines correctly and
so it is obviously not a five-year-old test, whereas it is well
within the ability of the six-year-olds. Curve 2 BCD shows
a sudden rise between six and seven, from 39 to 69 per cent.
This is obviously a seven-year-old test. Curve I EFG shows
its most significant rise between nine and ten, from 58 to 69
per cent. This is not such a marked rise as in the other cases,
but is probably sufficient to warrant a placing of the test at
ten years. It would seem in general that the higher the ages
the less marked are the rises in the curves, owing to the well-
known fact that one year in the lower ages means a greater
advance in the development of intelligence than in the higher
ages. It would seem to me that this test is correctly placed
at ten years, since the difference between the percentages at
nine and ten, 58 and 69 per cent., is greater than the differ-
ences between the percentages at ten and eleven, 69 and 73
per cent., or between those at eleven and twelve, 73 and 77
per cent. This curve would also illustrate well the fallacy of
adhering rigidly to the 75 per cent, standard. If we were to
do this we would have to consider it a twelve-year-old test
and give such credit to 73 per cent, of the eleven-year-olds
and to 69 per cent, of the ten-year-olds. Curve 3 BCD shows
a marked rise between thirteen and fourteen years, from 52 to
71 per cent. It is to be noted, however, that the twelve-
year-olds with 59 per cent, passes do better than the thirteen-
year-olds with only 52 per cent, passes. But the difference
between 59 and 71 per cent, seems sufficient to place the test
at age fourteen. Again fifteen-year-old children do more
poorly on this combination than do the fourteen-year-olds,
only 65 per cent, passing. This may be explained, perhaps,
by the fact that the majority of these were pupils in the grades
and very probably slightly below normal. The sixteen-year-
olds do well. A good many of them were high school pupils.
It was surprising to me to find that a correct passing of all
the three lines, BCD, is delayed until about the fourteenth
year. Six-year-olds can accomplish one and the one passed
by most of them is the C line. With children between the
ages of six to thirteen we find a large number of them passing
STANDARDIZATION OF TESTS
393
each line regarded singly, (see curves B, C and D, Fig. 2),
but it seems extremely difficult for the same child to pass
all three without a mistake. Perhaps the close attention
demanded is beyond the powers of the younger child. Or
again the similarity of the moves may be confusing to the
child, when line follows line with only a slight difference
between them.
Curve 2 EFGHIJ shows a marked rise between fifteen
and sixteen years, from 55 to 75 per cent., although we have
59 per cent, of the fourteen-year-olds passing this combina-
tion. The rise from 59 to 75 per cent, is no doubt sufficient
to standardize this test and yet I feel some doubt in regard
to the results in this case. The fifteen-year-old children
were not on the whole as typical of their age as the children
of the other ages. They were mainly children from the grades
with a sprinkling of high school students. Almost every curve
shows their deficiency as contrasted with the fourteen-year-
olds. It may be therefore that two out of the E to / lines is a
fifteen-year-old test. This is possible, although in view of
the results obtained not very probable.
TABLE II
NORMAL CHILDREN. PERCENTAGE CORRECT FOR VARIOUS COMBINATIONS
OF THE I.INES OF THE TEST
Chrono-
logical
Age
Number
Tested
Combinations of Lines
iXY
T.BCD
2BCD
T.EFG
$BCD
*EFGHIJ
3
II
0
10
9
0
0
0
4
18
22
22
6
0
0
0
56
66
56
21
II
7
0
6
59
•84'
83
39
20
22
8
7
6?
95
87
69
39
36
16
8
70
98
89
74
53
45
18
9
90
99
91
72
58
30
28
10
88
96
97
86
69
44
38
ii
77
97
97
83
73
50
38
12
82
97
k 96
93
77
59
4i
13
76
99
99
89
76
52
43
14
77
100
99
97
78
7i
59
15
44
100
IOO
93
77
65
55
16
30
IOO
IOO
93
84
80
75
Adult....
22
100
IOO
9i
86
77
Table II. shows the results for the six combinations, which
we believe can be satisfactorily used as tests. In each case
394 RUDOLF PINTNER
the number tested at each age and the percentage of correct
responses is given. It will be seen that an attempt was made
to test large numbers of children at each age. Even although
the experimenter often felt that a given line was far too
difficult for a child, yet it was given in order to get the negative
results, without which no real standardization of a test is
possible. It is just as important to know that the children
below a given age cannot pass the test as to know that those
'of a given age can pass it, and the comparison between the
two ages is not a just one unless equal numbers of both ages
in question have been tested. Much of the work in standard-
ization done up to the present time is open to just this
criticism. In Goddard's figures for the revision of the Binet
Scale we find tests placed at a given year because of a high
percentage of passes obtained from about 100 children,
whereas only between thirty to forty children and sometimes
even fewer of the next lower age had been tested. From my
experience with the results from this test and with others
upon which work is now being done, it seems to me to be
dangerous to take for granted that a few results in the age
below the test age are sufficient negative evidence to exclude
the possibility of the test falling at that age. It is generally
conceded that the tests at the lower end of the Binet Scale
are too easy and an inspection of Goddard's figures will show
that in the six-year-old tests, for example, only about half
and in some cases less than half of the children at the test
age were tested in the age below the test age. In Table II.
it will be seen that the first test of the Knox Cubes — I .XT-
is open to just the criticism that I have been urging. There
are 56 five-year-olds and only 18 four-year-olds. This is
owing to the fact that I experienced great difficulty in getting
four-year-olds. It may be that if fifty four-year-olds were
tested, this combination of lines might prove a four-year-old
test, and yet I do not believe this would be the case and so I
have considered it a five-year-old test. To none of the other
tests can this criticism apply. Large numbers above and
below the test age have been tested and there are sufficient
negative results below the test age in each case.
STANDARDIZATION OF TESTS
395
Total Number of All Lines Passed. — The results of the
test were also tabulated in another manner to show the number
of lines passed correctly at each age. The total number of
lines passed correctly by each child was added together and
the average and average deviation for each age found. These
figures are shown in Table III.
TABLE III
AVERAGE NUMBER OF LINES PASSED AT EACH AGE
Chronological
Age
Av. Number of
Lines Passed
A. D.
3
0-57
0.49
4
.1-43
0.76
5
,2.41
•17
77 % Pass 2 °r more lines.
6
4.22
.25
71% pass 4 or more lines.
7
5-12
.28
70% pass 5 or more lines.
8
5.60
.18
9
5-59
•34
10
6.29
•32
72% pass 6 or more lines.
ii
6.68
.76
12
6.66
.10
*3
6.66
.41
14
15
7-SS
7-45
2
75% Pass 7 °r more lines.
16
8.05
.11
62% pass 8 or more lines.
It will be seen that there is an almost steady increase in
the number of lines accomplished from age four up to age
eleven, where it becomes almost stationary at 6.6, to rise
again at age 14 up to age 16. In no case is the average
deviation larger than 1.7.
The results in this form may not be so suitable for the
placing of a test at a definite age, but they could of course be
incorporated in a scale. They show a very close correlation
with the lines which I have considered standardized. Age
four can do one line, that is the A line, and they are not able
to do one out of XY. Age five can do two lines, that is the
A line and either X or Y. Age six can do four lines, that is
the AXY lines and one out of BCD. Age seven can do five
lines, that in AXY and two out of BCD. Ages eight and
nine cannot do more than five lines, that is AXY and two
out of BCD and not one out of EFG. Age ten can do six
lines, that is AXY and two out of BCD and one out of EFG.
396 RUDOLF PINTNER
Ages eleven, twelve and thirteen all show an accomplishment
of a little more than six but not seven lines, that is they can
do AXY, two out of BCD and one out of EFG, but they cannot
do all BCD nor two out of EFG, as our curves have already
shown us. Age fourteen and above can do seven lines, that is
AXYBCD and one out of EFG. Age sixteen can do eight
lines, that is AXYBCD and two out of the remainder. All
this serves to corroborate our placing of the lines at partic-
ular ages. There is of course no reason why this method
should not be used in crediting for different ages. In that
case we should expect at five years two lines to be passed cor-
rectly; at six years four lines; at seven years five lines; at ten
years six lines; at fourteen years seven lines and at sixteen
years eight lines.
There does not seem to be very much difference in marking
between the two methods. In 33 cases examined nineteen
gave the same results for both methods; the other cases showed
a slightly lower age estimate when the system of marking by
total number accomplished correctly was used.
In twenty cases where the Binet ages were available for
comparison the estimate of age from performance on the Knox
Cubes agreed with the Binet age in seven cases by the group
method and in eight cases by the total number method.
The average amount of difference between the Binet age and
the age as estimated by the Knox Cubes was 2.17 years for
the group method and 2.36 for the total number method of
using the cubes. Apart from the slight differences in results
obtained by these two methods of marking for the test, it is
also interesting to note how well on the whole the estimate of
mental age agrees with that arrived at by the Binet Scale.
It is surprising that any one test should come so near the
result arrived at by a whole series of tests. I do not mean by
this to suggest that the Knox Cube Test alone should ever
be used to compute mental age. It must, of course, take its
place in a series of mental tests.
Perseveration and Reverse Order. — Two definite types of
errors occur in this test to warrant a few words. The first I
have called perseveration. It consists in repeating the A line
STANDARDIZATION OF TESTS 397
after the examiner has continued with other lines of the test.
The child does not seem to notice that the X or B or C line,
as the case may be, is different from the A line with which he
started. There is a perseveration of the tendency to tap the
blocks in the order given first, which also happens to be a very
easy, natural order. The child taps one, two, three, four,
again and again, sometimes adding some other moves to these
first four. This error of perseveration is not common among
the older children. Only one of our fourteen-year-olds did
this, and from his performance in other tests it is very probable
that this child is defective. Of the other children we found
one case at 12 years, one at 9, one at 8, five at 7, three at 6,
six at 5, one at 4, and two at 3, and it is to be remembered
that relatively few three- and four-year-olds were tested. It
would seem then that this error is due to the lower stage of
intelligence of the child, and perhaps one that may be a
clue to possible feeble-mindedness. This idea is strengthened
by the frequency of this error among our feeble-minded cases.
This occurred as follows:
At mental age 2 3 times
3 * "
4 10 "
5 16 "
6 22 "
7 10 "
8 3 "
10 2 "
II I "
This makes a total of 72 cases among the feeble-minded as
contrasted with 21 among the normal children. With the
feebleminded as with the normal, we notice a larger percentage
of cases among the children of lower mental age.
The second error is that of reverse order, i. e., beginning
with the block at the child's right instead of with the block
at his left, which is the one touched first by the examiner.
This was noted as "wrong end." It occurs most frequently
with the A line and rarely is peristed in for more than three or
four lines. It is again an error found more often among the
younger children, but it cannot be said to be particularly
RUDOLF PINTNER
characteristic of the feebleminded,
cases were noted:
NORMAL
Chronological Age
Age 3 i time.
4 4 times
5 7 "
6.. ..4 "
8.
The following number of
FEEBLE-MINDED
Mental Age
Age 2 2 times.
4 2 "
5 i "
6 2 "
7 i "
8 i "
No credit was allowed for this error. It was treated always
as a mistake, even although the line, whether the A line or a
more complicated line, was in other respects correctly per-
formed. In a system of mental classification giving a number
of points for each test or computing by whole or half credits,
it might be justifiable to given this error a half credit or a fewer
number of points.
The Feeble-minded. — Four hundred and sixty-three feeble-
minded individuals were given the same test. These were
inmates of an institution1 and all of them had been graded by
the Binet Scale. They have been classified according to
mental age and the percentage correct for the different lines
is shown in Table IV. and for the various combinations of
lines in Table V. Their performance in some of the com-
binations of lines is shown by curves on Fig. 3. On the whole
TABLE IV
FEEBLE-MINDED. PERCENTAGE CORRECT FOR EACH LINE OF THE TEST
Mental Age
Number
Tested
Lines of the Test
A
X
K
B
C
D
E
F
G
H
/
J
2
6
50
o
0
0
o
O
O
O
o
0
0
0
3
7
17
14
0
0
0
14
0
O
o
O
o
o
4
17
76
17
8
6
12
12
0
o
o
O
o
o
1
£
92
97
l\
21
19
19
27
40
8
5
7
8
6
13
O
0
o
0
o
0
7
85
99
87
14
41
36
4
2
8
0
0
2.5
8
73
99
98
82
55
62
71
20
26
19
1.5
1.5
4
9
75
IOO
97
79
71
67
21
24
20
2.1
1
7
10
69
IOO
89
92
69
67
87
35
33
30
0
IO
9
ii
27
IOO
IOO
IOO
66
70
89
29
29
7
15
ii
1 The writer wishes here to acknowledge the kindness and courtesy of Dr. Emerick,
superintendent of the Ohio Institute for the Feeble-minded.
STANDARDIZATION OF TESTS
TABLE V
399
FEEBLE-MINDED. PERCENTAGE CORRECT FOR VARIOUS COMBINATIONS OF
THE LlNES OF THE TEST
Mental Age.
Number Tested
Combinations of Lines
i XY
x B CD
aB CD
rEFG
*EFG
2
6
O
0
0
0
0
3
7
14
14
0
0
0
4
17
17
12
12
0
0
1
67
39
65
32
49
H
17
14
14
9
3
7
85
92
59
38
II
3-5
8
73
97
82
66
42
22
9
75
100
80
68
49
16
10
69
100
93
78
30
ii
27
100
93
78
63
22
the curves follow those of the normal children pretty closely,
generally remaining a little below. This means that the
normal children as a whole do slightly better on this test than
the feeble-minded of corresponding mental age. In the lower
ages — three and four — the curve for the feeble-minded is very
close to and sometimes even rises above the curve for the
normals. This corresponds to the well-known fact that the
Binet Scale is too easy at the lower end. Only in one curve,
I XY, do we see the feeble-minded curve rising above the
normal curve. In all the other curves it remains below the
normal, although sometimes it is very near the normal. The
feeble-minded individuals of the higher mental ages, ten to
eleven, do not surpass normal children of corresponding chro-
nological age, so that in this test we find no corroboration of
the fact that the Binet Scale is too difficult at the upper end.
This is, of course, not to be taken as a denial of the difficulty
of the higher tests of the Binet Scale, although none of the
tests have been subjected to the rigid standardization as has
been undertaken for this test. It may be that this cube test
is testing something that is not tested by the tests at the
upper end of the Binet Scale. It demands concentration of
attention to a continuously varying task and it requires the
subject to work at a pace that is set for him. The feeble-
minded are not by any means lacking in attention and per-
severance. Many will work for long stretches of time at a
400 RUDOLF PINTNER
task with the greatest concentration of attention, but if their
attention is required for a certain definite period to a changing
stimulus, it seems difficult for them to adjust their attention
to the continuously varying aspects of the problem. It is
probable that many of the feeble-minded would do much
better if the blocks were tapped not at a given constant rate,
but at a rate varying with their ability to shift their attention
from the one block to the other. In terms of rhythm, we
might say that the feeble-minded individual is too dependent
upon his own individual rhythm, and that he lacks the
capacity of adjusting himself readily to external rhythms.
Summary. — The Knox Cube Test given to normal children
and standardized for the different ages gives the following
tests :
i out of XY lines 5 year test.
1 out of BCD lines 6 " "
2 out of BCD lines 7 " "
1 out of EFG lines 10 " "
3 out of BCD lines 14 " " (probably).
2 out of EFGHIJ lines 16 " " (probably).
In actually using this method it would seem well to credit the
child with the age at which the most difficult combination is
passed. The child who passes 2 BCD is credited with a five-,
six- and seven-year-old test, even although he may have
failed in one of the easier combinations, the presumption
being that the failure was not due to an inability to pass
these easier lines, but to some disturbing factor foreign to the
test.
The actual number of lines passed may also be used as an
index to mental age, and if this method is followed we must
give credit in this manner:
For 2 or 3 lines 5-year credit.
" 4 lines 6 "
" 5 " 7 "
" 6 " 10 "
" 7 " 14 "
" 8 " 16 "
In using this test it would be well to follow strictly the
directions given at the beginning of this article, since any
STANDARDIZATION OF TESTS
401
deviation from this method is likely to give different results.
Touch the blocks at a uniform rate, beginning with the cube
at the child's left, never repeat a line and do not give the
child any suggestion of counting.
In conclusion, the justification for a long article such as
this one that deals solely with a single test may be found in
the fact that up to now we have been more or less satisfied
with a very indifferent standardization of the mental tests
that are being widely used in computing mental age. We
have, in fact, been avoiding the hard work and an inadequate
solution of the problem of standardization. Only by a thor-
oughgoing treatment of each and every test, such as has been
attempted here, will we ever arrive at tests that will give us
something more than a mere approximation to a child's
mental age. A scale made up of different tests standardized
in this fashion might lay claim to some exactitude. I think
emphasis must be laid on the number tested before we can
rest satisfied with the standardization of any test, and in this
connection it seems most important to me to have practically
the same number of cases in the ages immediately above and
below the test age, as we have in the test age itself. We dare
not assume that the age below the test age cannot accomplish
a given test unless we have sufficient negative results in that
age. Our knowledge of general intelligence and the develop-
ment of intelligence is so limited that it is very dangerous to
take anything for granted on an a priori basis.
THE ADEQUACY OF THE LABORATORY TEST IN
ADVERTISING
BY H. F. ADAMS
University of Michigan
That it is possible, by means of a simple experiment, to
tell even roughly the relative amount of business which
each of a series of advertisements will bring in, is a revolu-
tionary idea. The feasibility of such a prediction, however,
has been indicated by the writings of Strong and Holling-
worth. Granting this assumption to be true, it opens up an
entirely new field of experimentation for the practical psy-
chologist. Not only that, but it should result in the develop-
ment of a series of principles which would be of the greatest
benefit to the advertising man. The historical summary will
show the evidence upon which the assumption is based.
HISTORICAL SUMMARY
Chronologically, the first experiment which I have found
comparing the results obtained by laboratory and business
methods was performed with a set of five Bullard Lathe
Advertisements.1 Ten subjects, mechanics and engineering
students, were used. They were told to arrange the adver-
tisements 'in the order in which you would buy the machine.'
From the results, the relative position of each advertisement
in the series was obtained. The order as determined in this
way was then compared Svith the actual number of replies
for catalogues received by the Bullard Co. from each adver-
tisement.' It was found that the two orders agreed perfectly.
Another similar experiment was performed with a set of
Packer's Tar Soap advertisements.2 Fifty advertisements
1 Strong, 'The Relative Merit of Advertisements/ IO-H. Hollingworth, 'Ad-
vertising and Selling,' 8-10.
•Strong, 'The Relative Merit of Advertisements,' 11-15: 63-81; Jour, of Phil.,
Psy.t etc., VIII., 600-606. Hollingworth, 'Advertising and Selling,' 11-14.
402
TESTS IN ADVERTISING 403
were arranged by twenty-five subjects 'in the order in which
you would buy the soap.'
"When the order was compared with the order submitted
by Mr. Edward A. Olds, Jr., of the Packer Manufacturing
Company, and with the one from the Blackman-Ross Adver-
tising Agency, we found a high degree of similarity between
the three orders. The resemblance between the experimental
order and either of the other two is equal to a coefficient of
correlation of plus .52. The resemblance between the order
of the Packer Manufacturing Co. and the Blackman-Ross
Agency is equal to plus .64. There is then nearly as great
agreement between the experimental order and that of the
Packer Manufacturing Co. as between the latter and the
agency, which is now handling their advertising business."
Eight advertisements out of the 50 were then selected for a
more detailed study. These advertisements were arranged
by 100 subjects, 60 men and 40 women. The following co-
efficients of correlation were found to exist:
Between 100 subjects and 25 subjects 947
100 subjects and Packer Co 893
100 subjects and B.-R. Agency 866
25 subjects and Packer Co 840
25 subjects and B.-R. Agency 920
Packer Co. and B.-R. Agency 866
A third set of experiments was performed with Electric
Light Advertisements.1 There were originally in the set
five advertisements, but only three were used. For in two
of them there was a difference in one respect which affected
the company's data so much as to render them unsuitable
for the experiment. Thirty-six subjects were used. Working
the results of the three advertisements out by the order of
merit method, Strong found a coefficient of correlation of
plus i.oo between the order as determined by the laboratory
test and by the business returns.
Hollingworth,2 who gives what are apparently the returns
from all five advertisements in the set, finds a coefficient of
1 Strong, Jour, of Ed. Psy., IV., 393-404.
8 Hollingworth, 'Advertising and Selling,' 14-15.
404 H. F. ADAMS
correlation of plus .60 between the laboratory returns and the
business returns as measured by the cost per inquiry.
There have been made, then, three tests, the reports of
which are accessible in the psychological literature, deter-
mining the correlation between the laboratory test and the
business test. The lowest coefficient is plus .52; the highest
is plus i.oo. The average is in the neighborhood of plus .82.
The general conclusion drawn is that the laboratory test is a
satisfactory preliminary for any set of advertisements which
is to be used in business. For by it the poorer advertise-
ments can be eliminated and only the best kept. The
business returns and the laboratory returns agree so closely
that each can, in general, be used as a measure of the other.
Strong2 also worked out the coefficient of correlation
between the results of groups of persons coming from different
walks of life. His general conclusion is: "A group of 50
college students will represent very closely the judgment of
groups of educated business men and women, of young
business men, such as attend evening schools, etc., and of
women of the middle class regardless of age. They will not
represent at all the judgment of groups from small towns and
farming sections such as the regions around Garrison, N. Y.,
from which the data were obtained.
"It is fair to extend the results as set forth in previous
chapters regarding the judgment of college students to groups
of educated men and women in general. But as the data of
this report are mainly concerned with cheap articles in
common use, very little can be postulated concerning the
relation of various groups of individuals with regard to more
expensive commodities."
EXPERIMENTAL RESULTS
In view of the last sentence of the quotation given above,
it was thought that some profitable data might be disclosed
by the study of an entirely different kind of advertising ma-
terial, such as that of a mail order business.
The experimental work was done by John S. Deuble, in the
2 Strong, 'The Relative Merit of Advertisements,' 62.
TESTS IN ADVERTISING 405
psychological laboratory at the University of Michigan. The
results were put in their final form by the writer.
Three sets of advertisements were tested with a total of
161 subjects, 69. men and 92 women. The sets of advertise-
ments were as follows: 4 half-page advertisements of the
American Collection Service, Detroit, Mich., procured from
Mr. William A. Shryer; 10 full-page advertisements of the
American Collection Service; 9 quarter-page advertisements,
Saturday Evening Post size, of the Burroughs Adding Machine
Co., obtained from Mr. William A. Hart of the advertising
department of the Burroughs Company.
The following data concerning the American Collection
Service advertisements were furnished: the number of in-
sertions, the total number of inquiries, the advertising cost,
the cost per inquiry and the profit or loss for each advertise-
ment.
From the Burroughs Adding Machine Co. the following
data were received concerning each advertisement: the
number of inquiries, the number of trials, the number of
sales, the total amount received from the sales.
From these data, it should be possible to determine with
some accuracy which of the advertisements was of the greatest
value from the business standpoint. The settling of this
question is, however, more complicated than appears on the
surface. Any one of the points mentioned may be considered
as a test of efficiency. The trouble is that they may not agree
to any remarkable extent. To indicate this, the various
measurements are put in the form of tables which will show
the order of merit according to the different standards. In
the tables throughout the paper, number I indicates that the
advertisement so designated was the best from the stand-
point of the standard used; number 2, that it was in second
place, etc.
AMERICAN COLLECTION SERVICE
Half-Page Advertisements
Ad.
Average Number of Inquiries
Cost per Inquiry
Profit
A, .
I
I
I
B
2
3
2
c
2
4
D
4
4
3
406
H. F. ADAMS
Full-Page Advertisements
E..
•z
•J
6
F
10
0
c
G .
6
H .. .
g
c
7
/
4~
2
2
T
8
7
•j
L"
7
8
IO
M
IO
8
N
I
I
i
0
2
4
4
BURROUGHS ADDING MACHINE Co.
Ad.
Inquiries
Trials
Sales
Amount
1872 A . . . . ....
c
4.
2
2
1702 A
•j
2
•5
•»
1803 ^
x
7
6
1804 ^
6
6
8
0
1820 A
2
-i
r
1832 A. .
8
c
7
7
184.4. A
7
6
A
1870 A
8
8
1883 ^
I
i
i
i
It is evident from these tables that the measurement of
efficiency which is used is of some importance. The coef-
ficient of correlation1 between the orders as established by
the various possible measurements will show their resemblance.
AMERICAN COLLECTION SERVICE
Correlation between
Half Page Full Page
Inquiries and cost per inquiry 80 .915
Inquiries and profit 80 .430
Cost per inquiry and profit 40 .624
BURROUGHS ADDING MACHINE Co.
Correlation between
Quarter Page
Inquiries and trials 790
Inquiries and sales 834
Inquiries and amount 650
Trials and sales 773
Trials and amount 617
Sales and amount 933
These figures indicate that the busines test which is used
has a considerable influence on the coefficient of correlation
1 The coefficients of correlation are worked out by the formula given by Myers,
'Text Book of Experimental Psychology,' 1909, page 131. The formula is r = I
-[6 sigma (rf)Vn(n»— i)].
TESTS IN ADVERTISING
407
between it and the laboratory test. For reasons which will be
explained more fully in a later part of the paper, it appears
that the average number of inquiries per insertion of the ad-
vertisment is the fairest measure of the pulling power of the
advertisement under actual business conditions. This meas-
ure was taken because it seems to be the fairest test of the
actual pulling power, undisturbed by such things as a follow-
up system, the arguments of salesmen, etc.
The experiment was carried on as follows. The subject
was handed a series of advertisements and told to look them
over carefully. Having become familiar with them, he was
instructed to pick out the one which was to him the most
persuasive. By persuasive was meant the one which would
be most likely to make him answer the advertisement. This
done, he was asked to pick out the second best, the third, and
so on, until he had the series arranged in a descending order
from the most persuasive to the least persuasive. A first
choice was given a credit of I, the next best a credit of 2 and
so on down throughout the entire series. When the 161 sub-
jects had arranged the set of advertisements, the credits
which each advertisement had received were added and di-
vided by 161, thus giving the average place which the adver-
tisement occupied in the opinion of the 161 subjects. The
one which received the smallest average was considered to
have the greatest pulling power; the one which obtained the
largest average was credited with the least pulling power.
In the tables below are given the averages for each ad-
vertisement as worked out in this way, together with the
average deviation (A. D.). The reactions of the men and of
the women are not given separately, for no significant differ-
ences were found.
AMERICAN COLLECTION SERVICE
Half Page
Ad.
Average
A. D.
Position
A
1.84
0.74
I
B
1.93
0.84
2
c
2.8l
0.81
3
D
3-31
0-75
4
408
E. F. ADAMS
Full Page
E
6.13
2 "?!
F
4.4.7
^.31
2 6Q
•3
G
6 OO
2 OO
|
H
4. 83
2IQ
I
6.78
214.
]
5 18
2 Tl
i:::::::::;::::::::
£2
6.6O
•*"*j
I.Q7
g
M
412
2 2O
N
7.71
1. 80
IO
0
2.98
1.67
I
BURROUGHS ADDING MACHINE Co.
Ad.
Average
A. D.
Position
I
3-95
1-93
2
2
6.87
1.88
9
3
4-54
2-34
5
4
4.52
6.70
2.02
1.92
8
6
4.00
1.89
3
7
5-32
2.42
6
8
6.37
1.77
7
9
3-71
1.72
i
The tables show that a more or less definite order of the
relative persuasiveness of the advertisements in these three
sets has been worked out. The average deviations indicate
that some changes in order might occur if more subjects were
tested, but the extremes at least are fairly well defined.
As a check upon the probability of the final order of the
series as determined by the experiment, the following method
was used. The order was determined by averaging the
results of the first 10 subjects, then of the first 20 and these
orders compared. Then the order of the first 20 and the first
30 was compared, and so on for the 161 subjects. If the final
order was established relatively early in the series and per-
sisted without actual change throughout, it was thought that
the final order had a high degree of probability. The
question of the number of subjects necessary in an experiment
of this sort is always an important one and one which it is
difficult to settle off-hand. Such a test as the one described
is an entirely practical measurement which can be applied
at any time in the experiment.
It was found that the final order of the half-page adver-
tisements of the American Collection Service was determined
TESTS IN ADVERTISING 4°9
with 30 subjects. The addition of 130 more did not change
the relative order of merit, though, obviously, there were
changes in the degree of merit. Advertisements A and #,
even at the end of the 160 trials, might have been changed
by the addition of the results of 10 more subjects if they had
all given A fourth place and B first place. In the entire
experiment, however, A was put in fourth place only 10 times,
so the chances of its appearing in last place are I to 16. In
the same way, the chances of B appearing in first place are
I to 2.4. We are justified in concluding, then, that the final
order of this series is determined beyond reasonable doubt
for the class of subjects used.
With the full-page advertisements of the American Col-
lection Service, the final order was determined by the i2Oth
trial. The addition of the next 40 subjects did not affect the
relative order. The probable accuracy of this series, while
not so great as that of the half-page series, is sufficient for all
practical purposes. Some of the advertisements in the
middle of the series might have shifted one place by the ad-
dition of 10 more subjects, but this is unlikely, for in order
to do so, the ten would have to give averages which were not
even approached in the course of the experiment. So we
may conclude that we have obtained here an order which
would be very closely approximated by any laboratory test
conducted upon average undergraduates.
With the Burroughs Adding Machine advertisements, it
may be said a satisfactory final order never was obtained.
An order was established with the i2Oth trial which lasted
through the isoth, but the difference between the credits for
two of the advertisements was so slight that they were inverted
with the addition of the next ten subjects. Two other ad-
vertisements in the series might well have been inverted by
the addition of another 10 subjects. So we may say that an
entirely satisfactory final order never was determined for
this set of advertisements. The extremes are clearly defined,
but the intermediate members of the series are somewhat
variable. It seems probable, in view of these facts, that this
series of advertisements was of a more even, homogeneous
410
H. F. ADAMS
sort than was either of the series of the American Collection
Service. This does not necessarily mean that the Burroughs
Adding Machine advertisements are either better or worse
than the American Collection Service series. It simply
means that it was more difficult to make adequate judgments
between them.
This general condition emphasizes one fact which must be
plainly apparent to any student of the psychology of adver-
tising,— namely, that our laboratory tests have been unable
to tell us whether an advertisement is absolutely a good ad-
vertisement or not. The only thing such a test can do is to
place it relatively in a series. The limits of goodness and
badness are to be found inside of that particular series.
Because one advertisement is the first of one series and an-
other the last in a second is no reason for asserting that the
former is absolutely a better advertisement than the latter.
AMERICAN COLLECTION SERVICE.
Half Page.
Ad.
Order as Determined by
Lab. Test
Inquiries
Cost per Inq.
Profit
A
I
2
3
4
I
2
3
4
I
3
2
4
I
2
4
3
B
c
D
The following coefficients of correlation are found to exist:
Between the laboratory test and the number of inquiries I. CO
" " " " " the cost per inquiry 80
" " the profit 80
AMERICAN COLLECTION SERVICE
Full Page
Order as Determined by
Ad.
Lab. Test
Inquiries
Cost per Inq.
Profit
E..
•j
6
F
•7
IO
Q
G
6
r
I
H
A
I
/
2
2
/. .
C
8
7"
•3
£..
8
7
8
IO
M
2
IO
8
N
IO
I
i
i
0
I
2
4
4
TESTS IN ADVERTISING
411
The following coefficients of correlation are found to exist:
Between the laboratory test and the number of inquiries ,
" «« the cost per inquiry...
" " the profit
.—043
.—0.58
.— O.OI
BURROUGHS ADDING MACHINE Co.
Order as Determined by
Ad.
Lab. Test
Inquiries
Amount Received
I
2
2
9
3
4
I
3
5
5
2
4
4
6
9
5
8
2
5
6
3
7
4
7
6
i
i
8
7
8
7
9
i
9
8
The following coefficients of correlation were found to
exist:
Between the laboratory and test the number of inquiries — 0.43
" " " " " the amount received —0.06
The second step in the experiment is to compare the order
as determined by the laboratory experiment with the order
as determined by the different measures of business efficiency.
This comparison is given in the preceding tables.
Taking these three sets of results, we find the following
correlations between the order as determined by the laboratory
test and the order as determined by the average number of
inquiries.
Half-page advertisements, A. C. S = l.oo
Full-page advertisements, A. C. S = — 0.43
One-fourth-page advertisements, B. A. M = — 0.43
Average = 0.047
We also find the following correlations between the order
of merit as -determined by the laboratory test and the order
determined by the profit or the amount received.
Half-page advertisements, A. C. S = 0.80
Full-page advertisements, A. C. S = — o.oi
One-fourth-page advertisements, B. A. M = —0.06
Average = +0.243
412 H. F. ADAMS
These figures show simply chance resemblance between the
results of the laboratory test and the average number of
inquiries per insertion and very little better than chance re-
semblance between the laboratory test and the business test
where profits are used as the measure. The indication is
that the order of merit method has at least no universal
application to advertising problems. Of course, not enough
series of advertisements have been tested to settle the
question definitely. But since two of the three tests made
have shown a significant negative correlation, a fairly large
number of tests which result in equally strong positive
correlations will be necessary to offset the negative results of
these experiments.
It is possible, too, that college students are not satisfactory
subjects for the kinds of advertisements which were used.
Since a fairly large percentage of the men go into business
upon leaving college, they are possibly more satisfactory than
would appear at first.
As will be seen below, advertisement N of the American
Collection Service series present some queer anomalies. Even
if this advertisement is left entirely out of consideration, the
coefficient of correlation between the laboratory test and the
business test as measured by the average number of inquiries
is — 0.12.
The general conclusion seems to be that the mail order
business appeals to a special and limited class of individuals.
College students, on the average, are not fair representatives
of such a class. With a mail-order business, it is possible to
get returns which are extremely accurate, so such advertise-
ments would make the best material for laboratory tests,
if such tests would only work. On the other hand, it does
seem probable that the order of merit method might be applied
to those commodities which are in more general demand, such
as soaps, foods, and the like. It is, however, impossible to
get accurate returns as to the exact amount of business which
each advertisement has brought in. This renders it impossible
to be sure that the laboratory test actually has a high coef-
ficient of correlation with the business returns.
TESTS IN ADVERTISING 413
There are several interesting things which crop out in
connection with the advertisements of the American Collec-
tion Service. It will be recalled that Mr. Shryer furnished
two sets of advertisements, one consisting of four half-page,
the other of ten full-page, advertisements. Half-page ad-
vertisement D was, as far as possible, an exact copy of full-
page advertisement N. In N there were five pictures in
connection with testimonials, and a return coupon. In the
half-page advertisement D there were but three pictures
accompanying testimonials, and no return coupon. The
wording of the argument in the two advertisements was as
nearly identical as possible considering the change in size.
In fact, only 6 words were changed. The half-page adver-
tisement, however, was printed in smaller type, as necessarily
must be the case, considering the difference in size.
In spite of the great similarity of the two advertisements,
the results of the business returns, taking the average number
of inquiries as the basis, show that the full-page advertisement
N was the best of the set of ten. The half page advertisement
D was the poorest one of its set. With such discrepancies as
this in the business returns, it would be truly remarkable if
the laboratory test did show a high coefficient of correlation
with the business test. The laboratory test is at least fairly
consistent, for it ranked each of the advertisements, D and TV,
as last of the set to which it belonged.
In the business test, the presence of the return coupon
in one advertisement and its absence in the other may have
been a determining factor. Shryer1 ran two half-page ad-
vertisements at different times which were just alike except
that one had a return coupon, the other did not. The one with
the return coupon brought 83 replies, the other 41, showing
that in this particular case, the use of the return coupon
more than doubled the number of inquiries. If N had
brought only half as many replies, it would have ranked sixth
in the series, while if D had brought in double the number of
replies, it would have ranked second. However, if there is
so great an increase in efficiency coming from the return
1 W. A. Shryer, System, December, 1913, p. 579.
41 4 H. F. ADAMS
coupon, the laboratory returns, if they are adequate, should
have indicated it.
It so happened that there was a great similarity between
two other members of the two sets. Half-page advertisement
B was very much like full-page advertisement 0. The word-
ing was practically the same, though different pictures were
used. On the basis of the average number of inquiries per
insertion, B was in second place. According to the laboratory
test, it was second in the set of half-page advertisements.
From the standpoint of business returns, full-page advertise-
ment 0 was in second place; in the laboratory test, it was first.
This shows a fair consistency for both the laboratory test and
the business test.
The similarity between the returns from B and 0 is an
interesting check upon the dissimilarity found to exist between
D and N. For it might have been thought that the half-
page advertisements were much superior in general make-up
and appeal to the full-page advertisements. Logically, D
might easily have been the worst of the half-pages, while its
duplicate N might equally well have been the best of the full-
page displays. This, however, is rendered extremely doubtful
by the results obtained from B and 0. The indication is that
the peculiar results obtained with D and N are due to ex-
traneous conditions which could not possibly be controlled
in the laboratory.
Because of the results of this experiment as well as on
account of theoretical considerations, the writer has been led
to question the application of the order of merit method to
advertising problems. He does not question the true useful-
ness of the method, but does deplore the uses to which it has
been put. In his opinion, the experiments which have been
performed by this method on advertising problems can be
attacked on several sides.
In the first place, the experimenters have in but a very few
instances compared the laboratory results with the business
results. The idea of this comparison is, of course, to show
the dependability of the method as a laboratory technique for
investigating advertising problems. One thing which has been
TESTS IN ADVERTISING 415
done is to compare the order as determined by the laboratory
experiment with the order as determined by the opinion of
certain selected advertising experts, who were practically
put through the same experiment. Since Dr. E. K. Strong,
Jr., was one of the first to apply this method to the psychology
of advertising, a quotation from one of his articles will be
appropriate in bringing out the point. "It is scarcely
necessary to repeat that the results of the Packer Manu-
facturing Company are not based upon carefully compiled
data, but only upon the judgment of the firm based on their
business experience. Any one familiar with advertising knows
that such data have not been compiled for any extensive set
of advertisements, let alone a series of fifty extending over
twenty years of service. If such data did exist, it could not
be used at its full face value, as an advertisement of twenty
years ago might have been very effective then and be out of
date to-day.
"The order of the twenty-five subjects correlates plus .52
with the order of either of the two advertising experts. The
correlations between the orders of the two advertising experts
is plus .64. These relationships are lower than those which
have been obtained with other sets of advertisements. . . .
"It is evident, then, that the * order of merit method' does
give results that correlate high with results obtained in
business."1
Since by results obtained in business, Strong must evi-
dently mean, in the above connection, the opinion of adver-
tising experts, another quotation taken from the same
writer, but in a different article, will be especially interesting.
"At the present time there is no way of estimating which are
the good and which are the poor advertisements except on
the basis of personal judgment; and when the reviews and
criticisms of different advertising men are compared, it is
apparent that this personal judgment is today a very variable
factor."2
The second quotation robs the first of whatever force it
might originally have had.
1 Strong, Jour, of Phil., Psy., etc., VIII., 603, 604.
2 Strong, Jour. Ed. Psy., IV., 393.
41 6 H. F. ADAMS
In order to be of any particular value, the correlation
between the business test and the laboratory test must be
worked out with actual business returns. These are obtain-
able for but few kinds of commodity, since they depend upon
elaborate systems of keying. In order to have the keying
satisfactory, all orders must eventually come to a head office,
labeled in such a way that each advertisement may receive
full credit for its work. Such a thing is an obvious impossi-
bility with such products as soaps, foods, and in general those
things which are procurable at stores.
The advertisements which can be accurately keyed are
ordinarily mail order propositions. With any adequate
system of checking returns, it is possible to figure out from
keyed advertisements the following things: the average
number of inquiries per insertion, the average cost per in-
quiry, the total number of sales, the profit or loss. Some of
these returns obviously depend upon other things than the
advertisement itself, but it was the advertisement which
started the whole process going.
Which of these is the fairest measure of the pulling power
of the advertisement? The number of inquiries indicates
the number of persons who were influenced sufficiently by the
appeal to be incited to action. The weakness of this method
is that the position of the advertisement on the page1 or the
position of the page2 in the advertising section of the magazine
may be detrimental. The same advertisement in some other
position might have pulled many more inquiries. Again, the
time of year is a very important matter. There are good
seasons and bad seasons.3 General economic conditions,
national or sectional eras of prosperity are also modifying
factors.
1 Hollingworth, 'Advertising and Selling/ 80-90.
2 Starch, 'Advertising,' 106-116.
« Shryer, 'Analytical Advertising,' 167-170. Shryer says, on p. 169: "As a whole,
however, it may be said that the three largest months of practically every year are
January, February and March."
See also Starch, 'Advertising,' p. 50. The table at the bottom of the page shows
that, for the commodity mentioned, more advertising was run and more sales were
made during the first half of the year than during the last half. Another table, given
by Starch on p. 93, indicates that the most advertising is carried in May and December;
the least in January and August.
TESTS IN ADVERTISING 417
It seems obvious that the natural procedure in such cases
would be to repeat the advertisement enough times in different
parts of the magazine and at different times selected to take
account of seasonal differences and so on. The objection is
that with successive appearances of the advertisement there
is a fairly constant and regular decrease in the number of
inquiries.1 However, if enough advertisements were used in
this way, either the total or the average number of inquiries
would be a sufficiently satisfactory measure of the pulling
power of the advertisement. It is, in fact, the only obtain-
able measure of the pulling power uncomplicated by other
factors.
A second possibility is the average cost per inquiry. This
method is open to all of the objections noted above, and to a
still further one. The actual cost of the space occupied by
the advertisement does not in any way directly effect the
excellence of the advertisement itself. Even in the same
medium, the charge per page is liable to sudden shifts. It is
unfair to the advertisement to make it suffer the handicap
of the increased rate. The amount charged per page is not an
accurate measurement of the circulation of the medium and
so an approximation of the number of persons who may read
the advertisement.
The number of sales is obviously unfair, for we have to do
there not only with the advertisement itself, but with the
goodness or badness of the follow-up system, the efficiency
of salesmen, etc. Some of the blame may be laid to the ad-
vertisement for it may have been constructed in such a way
as to have interested many who could not possibly have bought
that line of goods. Or they have been misled by the adver-
tisement and when they found out what the product was from
the follow-up system, they lost interest.
The question of profit or loss resulting from the use of a
certain advertisement, while of considerable interest to the
business man, is still not a test of the pulling power of the
advertisement, but is a measure of the pulling power as modi-
1 Shryer, 'Analytical Advertising,' 8l ff., 220-223. Starch, 'Advertising,' 170-179.
Hollingworth, 'Advertising and Selling,' 235. Strong, PSY. REV., XXL, 147.
418 H. F. ADAMS
fied by the cost of the advertisement and the adequacy of the
follow-up system.
Taking it all in all, the average number of inquiries per
insertion seems to be the fairest test of the actual pulling power
of the advertisement. It is, then, the measurement which
should be used in endeavoring to obtain the correlation
between the orders of the business test and the laboratory
test.
Another criticism of the order of merit method as it has
often been used is on the ground of the number of subjects
employed or the number of tests made. Obviously, if rela-
tively few additional tests will change the order of the ad-
vertisements in the series, the experiment is unfinished.
From the experiments discussed above, it appears that the
number of tests necessary depends upon at least two factors.
In the first place, the actual amount of difference in terms of
judgment steps between the contiguous advertisements in
the series is an important consideration. With advertise-
ments far apart, where the judgment is easy to make, the
order will be established with relatively few subjects. But,
as the judgments become more and more difficult, an in-
creasing number of tests will be necessary. Secondly, the
number of advertisements in the series will be a determining
factor. For as we increase the number of advertisements in
the series, we ordinarily must necessarily decrease the
judgment steps, thus rendering a satisfactory arrangement
more difficult.
Shryer1 who was the first to use any considerable number
of persons in an advertising experiment, employed a total of
508 in his efforts to reach practical certainty. In the most
complex of his experiments, in which the method of paired
comparisons was used, the final order was obtained at the
3OOth trial. The addition of 200 more subjects left the
relative order of the advertisements in the series unchanged.
In this experiment he used but five different advertisements.
Had he used more than five, ten for example, he probably
would have had to employ a great many more individuals
1 Shryer, System, XXV., 146.
TESTS IN ADVERTISING 419
before obtaining a satisfactory final order. To be sure, his
material was such that there was a great chance for varia-
bility of response, but this is true of practically all experiments
carried on in the field of advertising.
Since Shryer, in his experiment, used the method of paired
comparisons, his results are not strictly applicable here. The
experiments which have been described in this paper indicate
something about the number of tests necessary. With four
advertisements in the series, the final order was determined
with 30 trials; with ten advertisements, the final order was
determined with 120 trials. With three other sets of ten ad-
vertisements each, which were used to test a different point, a
satisfactory final order was not obtained with 100 trials.
Until enough data have been obtained to work out a satis-
factory mathematical law, it seems that one of the tests of
an adequate number of subjects is the purely practical one
which was given in the first part of this paper. Enough has
been said, however, to indicate that in the majority of tests,
an insufficient number of subjects has been used.
The next point to be considered is whether the order of
merit method can be used to determine the relative pulling
power of a series of advertisements. Before considering this
point theoretically, we may repeat that the experiments which
have been designed and carried on to test the correlation
between the laboratory and the business test have sometimes
shown correlations as high as plus i.oo and sometimes as low
as — 0.60. It would seem, then, that sometimes the method
will work and sometimes it will not.
The instructions usually given in the experiment are,
"Sort these advertisements according to the order in which
you would buy the . . . ." That means that every individual
who performs the experiment makes a definite arrangement
of the advertisements, the order showing the persuasiveness
as far as he is concerned. The assumption is that from his
arrangement of the advertisements, it is possible to tell which
one made him buy the article, for each one experimented upon
is evidently regarded as a purchaser. There is, unfortun-
ately, no way of telling which of the persons experimented
420 H. F. ADAMS
upon would, in actual life, be sufficiently interested in any of
the advertisements in the series to make him purchase the
commodity.
In business, the situation is quite different. The following
figures, taken from Shryer,1 will point out what is likely to
happen in a mail-order business. " Let us assume a circulation
of 100,000 at $100 a page — an honest rate. Let us use a
page of the strongest copy, yielding inquiries at 10 cents. Let
us assume a selling average of 20 per cent., just double the
ordinary. We therefore secure 1,000 inquiries. We there-
fore sell 200 of the 100,000 or one-fifth of I per cent. . . .
These figures are assumed figures, but they represent the
outside limits of actual average results." These figures
indicate that on the average the inquiries are I per cent, of
the circulation of the magazine. It has been estimated that
five persons read each magazine. There is, then, a possibility
that the advertisement will be seen by 500,000 persons.
The estimate of the number of persons who see the ad-
vertisements varies from 10 per cent, to 50 per cent, of the
readers of the magazine. If we take the lower limit, 10 per
cent., that means that 50,000 will see some of the advertise-
ments. The proportion which will see a particular advertise-
ment is pure guess work. As a working basis, we will take
20 per cent. That means that 10,000 will see the advertise-
ment, and a thousand will be sufficiently interested in it to
reply, or 10 per cent. A great many of the other 90 per cent,
who do not inquire are almost interested enough to do so, still
more are slightly interested, others are indifferent, while still
others get a negative reaction. Therefore, the results ob-
tained from the mail-order business test are got from a very
small percentage of the total number of readers. The results
obtained from the laboratory test are arrived at by using the
results of the whole 100 per cent, of readers, instead of the
IO per cent, who would on the average be interested enough
to answer the advertisement. The using of the other 90
per cent, of the persons introduces factors into the experiment
which would quite certainly modify the results so that they
1 Shryer, 'Advertising and Selling,' XXIL, 24.
TESTS IN ADVERTISING 421
would not adequately express the normal results for the 10
per cent. If we only had some way of determining, in our
laboratory experiments, the individuals who make up the 10
per cent, who are sufficiently interested, we probably could
arrive at fairly dependable results.
It must be kept in mind that a mail-order business appeals
to a very small number of persons at best. The same general
situation exists, also, with regard to the more expensive com-
modities, such as pianos and vacuum cleaners. Such ad-
vertisements certainly appeal to a very small and select class.
Consequently, it is doubtful if adequate experiments could be
performed upon advertisements of these commodities in the
laboratory. The cheaper, more frequently used goods, such
as foods, soaps, etc., very probably could be tested adequately
if there were any way of determining accurately the actual
business returns.
Lastly, it is extremely doubtful if the great majority of
individuals can tell which of a series of advertisements would
be most likely to make them buy the advertised product.
It is very much like asking a man what he would do if his
house burned up in the night. The measurement of impres-
sions in relative terms offers considerably less difficulty, as
has been demonstrated in the experimental work upon sensa-
tion, esthetic judgments, and so on. Predicting probable
conduct is a much more hazardous matter. It is extremely
improbable that we can really tell what we will do under a
hypothetical condition unless we have developed a very
definite habit for meeting that situation. Then the chances
are that we will have two or more habits which are about
equally serviceable. Unfortunately for the advertiser, a con-
siderable percentage of the readers of advertisements have
formed the habit of appreciating advertisements and seldom
if ever responding.
The reading of advertisements has become a fixed habit
with many persons, not because they expect to buy anything,
but because the advertisements are an essential part of the
enjoyable features of the magazine. They are looked at for
esthetic appreciation, they are looked at for news value, for
422 H. F. ADAMS
they give information concerning the industrial activities of
the country which could never be found in the body of the
magazine.
The general conclusion which we seem forced to accept is
that the order of merit test is not a very adequate laboratory
method for testing the business value of advertisements.
Where it is possible to obtain accurate business measurements,
the laboratory test, using students as subjects, appears to be
quite inadequate. Where it is impossible to secure accurate
business measurements, the laboratory test may be adequate.
There is no way of telling.
VOL. XXII. No. 6 November, 1915
THE PSYCHOLOGICAL REVIEW
REACTIONS TO THE CESSATION OF STIMULI AND
THEIR NERVOUS MECHANISM
BY HERBERT WOODROW
University of Minnesota
The study of reactions to the cessation of stimuli, although
hitherto largely neglected, nevertheless presents a number of
points of genuine scientific interest. Reactions to the cessa-
tion of stimuli, or cessation reactions, are the same as regards
the reaction movement as ordinary reactions to the beginning
of stimuli, or beginning reactions. While, in the case of
beginning reactions, we have energy acting upon some receptor
of the individual's nervous system, and producing a motor
response, in the case of cessation reactions, we have merely
the discontinuance of energy, which, on the reflex theory,
would readily account for the discontinuance of the motor
response. But reference to more complex processes than
mere reflex conduction is necessary in order to understand
how the mere discontinuance of energy can produce a positive
reaction similar to any produced by the application of energy;
and yet, if cessation reactions are found to be as quick as
beginning reactions, the same complexity of mechanism must
be presumed for both.
Another point of interest, in connection with cessation
reactions, is the question whether the relation of intensity of
stimulus to reaction time is the same as in beginning reactions,
where the times are longer with weak stimuli than with strong.
This relation might be reversed if the reason for the long
reaction times with the weak stimuli was that in such cases a
weak nervous current was meeting with great resistance which
it took a long time to overcome: for then the cessation of the
423
424 HERBERT WOODROW
weak nervous current could not also meet with great resistance,
but, on the other hand, the weaker the current, the more its
cessation would be favored.
That cessation reactions may throw some light upon the
explanation of the differences in reaction time which occur
with variation in the mode of stimulus has already been
recognized.1 The fact, that with moderate intensities, the
reaction time to sound is ordinarily shorter than that to light
is commonly explained on the ground that the stimulation of
the ear by sound, a mechanical process, takes less time than
the stimulation of the eye by light, a chemical process.
That this explanation is as yet a matter of speculation must
be admitted, as has been pointed out by Dunlap and Wells,
who instigated a series of experiments designed to aiford a
more satisfactory explanation.2 Wells, with this same prob-
lem in mind, conducted some experiments in which the reac-
tions occurred upon the disappearance or occlusion of the
stimulus. That is, the stimulus consisted in darkness,
preceded and followed by illumination, with the subject re-
acting at the moment the darkness appeared. From his
experiments, he concluded that the lag in the sensory process in
the case of vision cannot account for as much as 10 a of the
lengthening of the visual reaction time beyond that of the audi-
tory. The argument, by which he arrives at this conclusion,
need not here be analyzed, as it is not based so much upon the
reaction times to the disappearance of the stimulus, as upon
the fact that an interruption of the light for only 10 a was
found to be plainly perceivable.3 As regards the reaction
time to the disappearance of the light stimulus, Wells con-
cludes that it differs little from that to the appearance of the
light. This conclusion is supported by a large number of
sensory reactions, and is in conformity with that indicated
by the data presented below.
Another question concerning cessation reactions is whether
the after-image has anything to do with determining the
1 Wells, G. R., 'The Influence of Stimulus Duration on Reaction Time/ Psychol.
Monog., 5, 1913.
2 'Some Experiments with Reactions to Visual and Auditory Stimuli,' PSYCHOL.
REV., 1910, 319.
3 Op. «*., 65.
• CESSATION OF STIMULI 425
reaction time. Many other questions also suggest them-
selves, but the problems that have been referred to are
sufficient to indicate the nature of the scientific interest of
the present investigation.1
In the case of all the reactions here reported, the in-
structions called for a reaction which would ordinarily be
regarded as a form of "motor" reaction. Reactions are
usually divided into two main groups, those in which, during
the preparatory interval, the attention is mainly directed to
the stimulus, called sensory, and those in which it is focused
primarily upon the reaction movement, called motor. A
more important distinction, however, at least from the point
of view of reaction time, is that between reactions with instruc-
tion to react as quickly as possible, and reactions without
such instruction.2 In the present instance, the subjects were
always instructed to make every reaction as quickly as they
possibly could. Nothing was said about the direction of
attention, but the subjects' introspections indicated that their
attention was, primarily, neither upon the stimulus or its
image, nor upon the reaction movement or its image, but
upon the idea of reacting as quickly as possible. Just how
this idea was carried in the subjects' minds, that is, in what
imagery, or whether in any imagery, I am unable to conclude
from the introspective data. No attempt was made at elab-
orate systematic introspection.
Apparatus. — The reaction times were measured by means of a Hipp's chronoscope.
The chronoscope circuit and the stimulus circuit were separate, but by means of a
double switch both circuits could be closed simultaneously. The stimulus, for light
reactions, consisted in the illumination of a Geissler's tube, while for sound reactions,
the stimulus was the vibration of a telephone receiver. For beginning reactions, both
the chronoscope and stimulus circuits were closed at exactly the same instant by means
of the double switch. Each side of this switch was provided with a platinum wire
which sank into an adjustable cup of mercury when the experimenter tapped upon the
switch handle. For cessation reactions, the stimulus circuit was arranged so that it
was closed when the handle of the switch was raised, and broken when the handle
was tapped down. Thus, in this case, the chronoscope circuit was closed as the stimulus
circuit was broken, though there was an exceedingly slight time between the breaking
of the latter and the closing of the former. This time, when measured, was found to
1The general subject of cessation reactions was first suggested to me by Pro-
fessor H. C. Warren of Princeton, in 1907. The results here reported on sound were
presented in full in a paper read in April, 1912, before the Minnesota Psychological
Conference.
8 See Woodrow, 'The Measurement of Attention,' Psychol. Monog., 1915, Chap. n.
426 HERBERT WOODROW
vary from o to 3 <r, but is regarded as a constant error of 2 cr, and added to all averages
of cessation reaction times.
The chronoscope circuit included besides the chronoscope and half of the double
switch already mentioned, the subject's reaction key, a battery of Edison primary cells
of 12 volts, and a Wundt's fall-hammer. The chronoscope was controlled before and
after each 50 reactions, by means of the fall-hammer, and the fall-hammer itself was
tested each day with a 250 d.v. fork.
The stimulus circuit, when arranged for light reactions, included, as already
stated, a Geissler's tube. This tube was suspended before an oblong aperture in a
dark box, a few feet in front of and slightly below the subject's eyes, as he sat in the
dark-room at his reaction key. The Geissler's tube was actuated by an inductorium
placed in the experimenter's room. The buzzer of the inductorium was kept going
continuously during the experiment, but only the secondary circuit, which included
the Geissler's tube, passed through the operator's double switch, and so the tube was
luminous only when the secondary circuit was closed at this switch. In order to
weaken the intensity of the stimulus, no change was made in the electrical circuits, but
a ground glass covered with white paper was placed over the aperture of the dark box
containing the Geissler's tube.
In the case of sound reactions, the sound used as a stimulus was that made by a
telephone receiver, through which there passed a current which was interrupted both
50 and 250 times per second, by two electro-magnetic tuning forks. All the current
passed through the forks, but the telephone was in a shunt circuit. Consequently, the
-forks ran continuously, whereas the telephone ran only while its circuit was closed at
the experimenter's double switch. By decreasing the resistance in parallel with the
telephone, the loudness of the telephone could be decreased.
The reason for running the telephone current through both a 50 and a 250 fork
was that only in this way could the click which followed the break of the current be
eliminated. This method is noted by Pillsbury in his discussion of 'Methods for the
Determination of the Intensity of Sound,'1 and is one that I have long employed.2
Records of the vibration of the telephone plate were obtained by attaching a light glass
pointer to the plate and having this pointer mark upon a smoked drum alongside a
Pfeil time marker placed in the chronoscope circuit. A sample of these records is
here reproduced.
The vibrations of the 250 fork are seen superimposed upon those of the 50 fork.
The record shows that the sound started in at its maximum intensity with the first
1 ' Report of the Committee of the American Psychological Association on the
Standardizing of the Procedure in Experimental Tests,' Psychol. Monog., i, 1910.
Pillsbury writes as follows: "Dr. Shepard, working in my laboratory, found that the
click could be lessened to a point of not being noticed if two tones were superimposed
upon the telephone. He used the commercial current of 60 cycles and a 250 d.v. fork.
The physical basis for the effect is obscure, but the empirical effect is obvious."
2 See 'A Quantitative Study of Rhythm,' Archives of Psychol., 12, 1909.
CESSATION OF STIMULI 427
vibration of the telephone plate, and that it stopped as suddenly as it began. It will
be noted that no large vibrations follow the breaking of the circuit as they would if
there were a click. A number of careful observers agreed that there was no noticeable
click at either the closing or the opening of the circuit, thus confirming the objective
record.
Sound reactions were taken with three different intensities
of stimulus, called medium, weak and liminal. Physical
measurements of the intensity were not made, but care was
taken to keep the current the same throughout the work
with any one intensity. The reaction times themselves offer
a sufficient index of the intensity. The intensity given in
Table I. as liminal, in reality exceeded the true limen by an
extremely small amount. It was determined as follows: Four
series of minimal changes in intensity were used, two of which
passed from an audible to an inaudible sound and two from
an inaudible to an audible one. This procedure was just
what one might employ to determine the sensation threshold.
The sound was then placed not at the average liminal inten-
sity obtained, but at the lowest intensity which did not fail
to produce a sensation in any of the four series.
Since an object of the present study is a comparison of
cessation reactions with beginning reactions, both kinds of
reactions were always taken at each sitting of one hundred
reactions. On alternate days the middle fifty reactions were
cessation reactions and the first and last twenty-five were
beginning reactions; while on the remaining days the middle
fifty were beginning reactions and the first and last twenty-
five were cessation reactions. In the tables below, averages
for beginning and cessation reactions given on the same line
of the table were obtained at the same sitting. Each average
is the average of fifty reactions. No reactions were thrown
out on account of their deviation from the average, except in
the case of the liminal sound. However, all reactions which,
by a prearranged signal, the subject indicated were mistakes,
as well as all cases of error of manipulation on the part of the
experimenter, were rejected. In the case of the liminal sound
there were a number of cases where no reaction at all occurred
and some that were very long, from 1,500 to 5,000 a. These
long reactions occurred both in the case of beginning and
HERBERT WOODROW
cessation reactions, but all reactions over 1,500 cr were not
counted. The total number of reactions to sound and light
stimuli reported in the following tables is 10,000.
As a warning signal, in the case of beginning reactions the
click of an electric sound-hammer was used, and given always
two seconds before the reaction stimulus, the interval being
indicated to the experimenter by means of a pendulum. In
the case of cessation reactions the beginning of the stimulus
itself acted as the warning signal. The cessation of the
stimulus always occurred two seconds after its beginning.
Five subjects were used, three of which, subjects Ww, Ht,
and St were quite practiced in beginning reactions before the
work began, while the other two, subjects Vs and Sz, were
altogether unfamiliar with work in reaction time.
The results obtained for reactions to sound and to light
are presented in Tables I., II. and III. These tables show
the results of each sitting in the order in which they were
obtained. The results are summarized in Table IV. In this
latter table, the mean variation given is the average mean
variation for a series of fifty measurements.
The conclusion to be drawn is clear and simple. In the
case of sound and light reactions, there is no appreciable
difference between reaction time to the beginning of a
stimulus and reaction time to its cessation, no matter what the
intensity of the stimulus, and no matter what its mode. In
view of the variability in reaction times, such a close corre-
spondence between the cessation and the beginning reaction
times as shown by Tables I. to IV. could not be regarded
as a reasonable expectation unless the true values of both were
in all cases substantially the same.
It is to be regretted that results with other senses than
light and sound were not obtained. Difficulties of technique
were largely responsible for limitation of the work to these
modalities. Some reactions to touch were taken, however,
using an intermittent current passing through two fingers of
the left hand, each finger being placed in a salt-water elec-
trode; but so much difficulty was experienced as a result of
sensory adaptation and the after-tingling of the fingers, that
N, for each av. = 50'.
CESSATION OF STIMULI
TABLE I
SOUND REACTIONS
Total No. of reactions = 4,000.
429
Intensity
Subj.
Series
Beginning Reactions
Cessation Reactions
Av.
M. V.
Av.
M. V.
^tedium . .
Ht
II
«
((
<«
M
«
Fj
«c
<«
M
(1
««
(I
«<
M
Sz
<«
M
M
«
«
M
M
M
ffj
||
Pj
«
«(
<(
M
• Sz
u
II
M
«(
Ht
M
u
F>
«
«
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
Av.
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
Av.
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
Av.
I.
II.
III.
IV.
Av.
I.
II.
III.
IV.
Av.
I.
II.
III.
IV.
Av.
I.
II.
Av.
I.
II.
Av.
US
H3
121
I36
123
116
106
122
119
143
123
141
135
140
145
132
137
137
I46
143
162
137
140
154
147
IS7
I48
172
202
175
I85
I84
1 88
196
160
150
174
215
198
222
2O I
209
829
728
779
755
995
875
II
19
20
20
13
14
12
15
16
3
18
i5
20
12
15
I?
17
20
17
20
23
14
22
26
14
19
34
%
24
26
18
28
II
20
31
34
?
33
146
114
130
137
315
226
121
125
128
127
"5
121
H3
1 2O
121
144
142
155
147
136
144
138
140
143
157
155
143
151
I5J
136
147
144
148
170
190
170
201
I83
186
X55
155
171
167
226
204
223
218
218
682
807
745
735
909
822
15
15
17
15
15
II
15
13
IS
16
20
18
21
14
%
14
17
24
12
22
18
23
17
25
25
21
44
26
27
36
33
20
32
20
^7
22
35
27
40
29
33
no
147
133
177
359
268
«
«
«
u
«
a
«
«
«
«c
M
<(
«c
«
«
<«
«c
M
«(
M
«(
M
M
«
M
Weak..
<
i
c
(
«
«
I
e
t
(
i
<
1
<
Liminal
«<
M
M
in
M
43°
HERBERT WOODROW
N, for each av. = 50.
TABLE II
REACTIONS TO BRIGHT LIGHT
Total No. of reactions = 3, 600.
Beginning
Reactions
Cessation
Reactions
Subj.
Series
Av.
M. V.
Av.
M. V.
Ww
I.
1 60
je
IC2
17
K
II.
14.2
U
14.7
16
«
III.
IC4.
J?
y
160
16
(i
IV.
I4.O
17
TC2
21
<(
v
A7
1 60
IO
TP-7
ry
it
VI.
1 68
24.
JAA
AJ
/?
«
VII.
JCQ
17
IAC
i6
(i
VIII.
1 4.O
If
157
Id
«
Av.
154
16
I"?I
16
Ht
I.
156
18
1 60
18
u
II.
163
2S
JCg
2f
u
III.
1 68
IO
171
12
«
IV.
160
21
171
jr
((
Av
162
21
l67
18
St . .
I.
IQO
17
187
2O
((
II.
IQ2
22
IQO
J7
««
III.
187
JC
1 80
2O
«(
IV.
178
21
IQ2
~"
16
«
V.
187
22
176
26
<(
VI.
178
10
187
16
((
VII.
176
II
I 7O
22
II
VIII.
178
28
101
1$
(1
Av.
X83
21
184
19
Vs. .
I.
2OO
24
196
21
(i
II.
2O7
22
177
12
<(
III.
IQO
18
1 80
11
«
IV.
IQ3
14.
IQ3
18
«
V.
1 04
17
IQ2
IQ
u
VI.
170
20
178
16
«
VII.
1 86
18
1 80
16
«
VIII.
1 86
21
184
20
«(
Av.
192
19
185
i?
Sz..
I.
196
IJ
200
23
(f
II.
200
23
193
27
tl
III.
207
21
193
18
u
IV.
177
23
2O I
14
«
V.
IQI
2$
2OO
22
u
VI.
IQO
21
2OO
23
te
VII.
*y?
216
28
2O C
21
«
VIII.
221
24.
212
26
(C
Av.
2OI
22
201
22
the results cannot be regarded as reliable. The few that
were obtained showed a marked prolongation in reaction
time with decrease in the intensity of the stimulus, both with
beginning and cessation reactions. The cessation reactions,
however, were uniformly about 30 <r longer than the beginning
reactions, which fact may be attributed to the after-tingling
CESSATION OF STIMULI
431
of the fingers set up by the irritation of the electric current.
Some experiments were also tried, in which the stimulus
consisted in the fall and rise of an electric hammer arranged
to strike the back of the finger. In this case, while the
stimulus was adequate, one could not be certain that the
N, for each av. = 50.
TABLE III
REACTIONS TO WEAK LIGHT
Total No. of reactions = 2,400.
Subj.
Series
Beginning Reactions
Cessation Reactions
Av.
M.V.
Av.
M . V.
Ht. .
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
Av.
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
Av.
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
Av.
179
I96
211
216
204
221
211
201
205
221
217
252
240
244
*P
269
241
243
254
239
269
296
284
258
264
280
268
35
44
37
34
20
%
33
33
25
23
%
22
29
19
28
26
2?
33
30
22
%
24
27
28
192
I78
205
2I4
194
217
217
209
203
230
222
214
234
255
229
24I
243
234
244
250
%
268
255
228
230
250
28
30
26
22
18
24
ii
24
21
15
21
30
21
18
22
17
21
15
22
31
29
30
34
21
23
26
tt
«
ii
tt
tt
tt
tt
tt
Vs
tt
tt
tt
tt
tt
tt
tt
n
St..
ti
it
tt
tt
tt
tt
tt
tt
removal of the pressure was a change equivalent in intensity
to its occurrence; but the results obtained in this case were to
the effect that the beginning and cessation reaction times are
equal. Thus, with one subject, the average reaction time
for 100 reactions to the beginning of the stimulus, was 119 cr
and for 100 reactions to its removal, 1200-. With another
subject, the average for 100 beginning reactions was 128 <r
and for 100 cessation reactions with the same stimulus,
132 <r.
432
HERBERT WOODROW
The simplicity of the conclusions drawn above from the
results presented in Tables I. to IV. should not blind us to
the important bearing that they have on current theories of
the action of the nervous system. The explanation of the
results here presented, by means of the theories of nervous
action that are at present most in vogue, is extremely dif-
ficult,— so difficult, in fact, as to suggest that these theories
are themselves incorrect, and that they must either be given
TABLE IV
A SUMMARY OF TABLES I., II. AND III
Beginning
Reactions
Cessation
Reactions
Mode
Intensity
Subj.
Av.
Av. M. V.
Av.
Av. M. V.
Sound
Medium. .
Ht
110
16
121
JC
«
Vs
137
17
14. -j
17
(1
Sz
*j/
14.8
IQ
H8
21
w$
Weak
Ht
184.
26
183
yy
«
Vs
174.
2O
**
167
22
«c
Sz
2OQ
31
218
??
Liminal
Ht
770
I1O
74. c
/??
ii
Vs
87?
*
226
822
168
Light
Brieht .
Ww
ISA
16
ici
16
8
f"
Ht
*5t
l62
21
167
18
«
«
St
i8<?
21
184.
IQ
H
M
Vs
IQ2
IQ
185
17
«<
«
Sz
2O I
22
2O I
22
«
Weak. .
Ht
2O C
??
203
24.
II
<(
Vs
24, 3
26
234.
21
«C
M
Sz
268
28
250
26
up or seriously modified. In particular, it can be shown that
certain phenomena which it is customary to explain by ref-
erence to the latent period of sensory stimulation or by ref-
erence to certain hypothetical effects of the synapse, are not
adequately explained thereby. Their explanation requires
either a theory which is incompatible with the statement that
the reflex is the type of all nervous activity, or else a theory
which involves some modification of the ordinary concept
of the reflex.
In the first place, let us consider the explanation of the
increase in reaction time with decrease in the intensity of
CESSATION OF STIMULI 433
stimulus. Pieron,1 in a recent able discussion of this matter,
concludes that the main factor, in determining the increase
in reaction time, is the increase in the latent period of ex-
citation of the first sensory neurone with decrease in intensity
of stimulation. His discussion of other possible explanatory
factors, however, is unprejudiced. He correctly regards as
very doubtful the proposition that the rate of transmission
of nervous energy along a neurone varies with the intensity of
stimulation.2 Likewise, and with equal correctness, he rejects
the idea that the effect can be explained by reference to phe-
nomena on the motor or centrifugal side of the process.3 He
does not deny, though, that when the excitation is weakened
there may be an increase in the time of transmission from
one neurone to another. He thinks, however, that the vari-
ations in time, due to this last-mentioned factor, are not very
considerable, and that they do not suffice to explain the
variations obtained experimentally. Pieron also raises the
question whether the central or brain phase is not the seat
of the principal variations. "La partie la plus longue, dans
cette phase, correspond a la circulation de 1'influx associatif
dirige, et sa brievete depend surtout de Petat d'attention,
c'est-a-dire de Paiguillage prealable qui assure la conduction
de cet influx par les voies les plus rapides."4 He argues,
however, that since the reactions are always accompanied
by a state of intense attention, the fact that a more intense
1 'Recherches sur les lois de variation des temps de latence sensorielle enfonction
des intensites excitatrices,' Ann'ee psycho!., 1914, 17-96.
1 In confirmation, see the following: Gotch, Journ. of PhysioL, 1902, 395. Koike,
Ztsch. f. Biol., 1910, 310. Lucas, Journ. of Physiol., 1911, 46. Adrian, Ibid., 1912,
389; 1913, 384-
* In this connection, one may well cite the work of Moore, 'A Study of Reaction
Time and Movement,' Psychol. Monog. Sup., i, 1904. Moore writes, "In one and
the same series the reaction time undergoes considerable change, but the movement time
is fairly constant. If you introduce factors which increase the difficulty of attention,
the reaction time is lengthened and rendered still more variable, but the time of move-
ment remains about the same," p. 58. "That the efficient path from cortex to muscle
is not affected by the disturbance of the attention and that reaction time is not
lengthened by any changed conditions along the path, seem to be conclusions warranted
by the fact of constancy in the time of the movement by which the reaction was exe-
cuted," p. 59.
* Op. cit., 73-74.
434 HERBERT WOODROW
excitation better calls out attention cannot be of very great
importance. Concerning his final hypothesis, that the ex-
planation of the relation between reaction time and intensity
of stimulus lies chiefly in the latent period of the sense-organ,
Pieron himself admits that it is an hypothesis, the truth of
which he cannot demonstrate. One of the principal argu-
ments which he suggests in favor of his conception is the
fact that the data at hand indicate that the relation between
intensity of stimulus and reaction time differs in the case of
different senses.
In spite of the lack of evidence in support of the con-
clusions of Pieron concerning the role of the sensory latent
period, they have a certain plausibility which would no
doubt lead many to agree with him. Others, however, like
Sherrington, would give more stress to the role of the synapse.
Sherrington calls attention to the fact that the latent period
of reflexes, that is, the time between application of stimulus
and appearance of end-effect, increases with decrease in the
intensity of the stimulus.1 Thus, he finds that the latent time
of the scratch-reflex varies from 140 a with intense stimula-
tion to 500 (7, or even several thousand cr, with weak stimu-
lation. He writes: "This slackening of propagation speed
under weak stimuli is, I would urge, a more significant differ-
ence between reflex-conduction and nerve-trunk conduction
than is the mere greater slowness of the former than the
latter."2 This difference between conduction in reflex-arcs
and nerve-trunks is referable to that part of the arc which
lies in the gray matter. The slower conduction in the gray
matter as well as the increase in reflex time with weak inten-
sity of stimulus is supposed to be due to phenomena of
transmission occuring at the synapse, where a surface of
separation acts as a resistance or barrier to the passage of the
nervous current from one neurone to the next.
The weaker stimulus, then, is to be thought of as resulting
in a longer reflex time because it is delayed longer in trans-
mission from one neurone to the next, than is the more intense
1 'The Integra tive Action of the Nervous System/ 1906, 21.
2 Loc. cit.
CESSATION OF STIMULI 435
stimulus. Sherrington, apparently, does not regard the
effect of intensity on reflex time as due in any appreciable
degree to the latent period of the sense-organ.
Now, while the prolongation in reflex time with decrease
in the intensity of the stimulus is a different phenomenon
than the prolongation in human reaction time with decrease
in intensity, yet it seems highly probable that the essentials
of the explanation are the same in both cases. At any rate,
there is just as much evidence that the effect of intensity
upon reaction time is to be explained by reference to the
impeding action of the synapse, as that the effect of intensity
upon reflex time is to be so explained. Fully as many synapses
are involved in the voluntary reactions of human beings as in
the reflexes of the dog, and any effect which results from the
nature of the synapse must be present in both cases and must
be explained in both cases by the action of the synapse.
The inadequacy of both the sensory latent period and the
synapse theories may be shown, to a large extent, by the same
argument: for after all, both theories offer much the same
explanation of the prolongation in reaction time with decrease
in intensity of stimulus. In both cases, the concept of re-
sistance to be overcome is the fundamental one; in one theory,
the resistance is thought of as occurring at the first sensory
neurone met with by the stimulus and in the other, as occur-
ring at every synapse. It is the same concept that is constantly
used in attempts at explaining the physiological side of all
manner of psychological phenomena.1 The excitation prc^
duced by the stimulus has to exceed some minimal limit, or
threshold, before it is sufficient to bring about a reaction. A
small force has to act longer to produce a given effect than a
large force : consequently, it is argued, the weaker the stimulus,
the longer the time to produce the required degree of excita-
tion.
Now is such an explanation as the above tenable? We
1 Ladd and Woodworth write as follows : " In fine, it seems possible to conceive
the action of the nerve-centers as a process of the transmission of nerve-impulses that
is subject to the peculiarities of central conduction. Most of these peculiarities can
be stated in terms of resistance — resistance in general high, but variable with many
conditions." 'Elements of Physiological Psychology,' 1911, 287.
436 HERBERT WOOD ROW
must conclude that it is not, if for no other reason than that
it does not hold for cessation reactions; for, if we argue that
a large resistance increases the time required for any stimulus
to produce its effect, we cannot also argue that this same large
resistance also increases the time required for the disappear-
ance or the cessation of the effect. On the contrary, the dis-
appearance of the effect should be hastened. A few analogies
will make this clear. If we make a slight impression on the
skin, it will disappear, or fall below a prescribed threshold,
quicker than if it is deep. The after-image of a bright light
lasts longer than that of a faint light. A string vibrating
with large amplitude requires a longer time to come to rest,
or to fall below any given threshold, than one vibrating with
small amplitude. The plasma membrane of a nerve cell
would recover from a slight increase in permeability more
quickly than from a great increase in permeability.1 Such
illustrations, which could be multiplied indefinitely, all show
how improbable it is that any resistance offered by the first
sensory neurone, or even by the synapse, would act in such
a way that it would not only impede the action of a weak
stimulus more than that of a strong, but would also impede
the cessation of the action of the weak more than that of the
strong.
One might object, and not without some justification, that
this argument leaves out of consideration the fact that the
resistance offered by the synapse may not be resistance to
the rise or fall of the excitation but merely to its conduction.
This resistance, it might be alleged, merely results in the weak
current traveling along the nerve fibre more slowly than the
strong. In all cases, the awareness of the cessation of the
stimulus would be delayed by the time required for the
very rearmost part of the stream of excitations produced by
the stimulus to pass from the sense-organ to the cortex. Now,
1 For an excellent discussion of the electric phenomena in nerve and muscle cells,
see Hober, 'Physikalische Chemie der Zelle und der Gewobe,' 4th ed., 1914, especially
Chap. XIL, 'Elektrische Vorgange an physiologischen Membranen.' Cf. also Lillie,
'The Relation of Stimulation and Conduction in Irritable Tissues to Changes in Per-
meability of the Limiting Membranes,' Amer. J. of PhysioL, Vol. XXVIII. , 1911, 197-
222.
CESSATION OF STIMULI 437
one might argue that this time would be greater for weak than
for strong nervous excitations. Against this view, however, we
have the results of recent physiological investigations, which
show that the rate of transmission of nervous impulses is
independent of their magnitude.1 More decisive is the fact,
apparently well established,2 that the 'all-or-none' law applies
to the normal nerve-fiber. This law is to the effect that any
stimulus which excites a nerve-fiber at all will produce a
maximal excitation. It follows that differences in intensity
of excitation of sensory nerves is due to difference in the
number of nerve-fibers (or conducting elements) stimulated,
or else, as indicated by the results of Frohlich,3 to variation
in the frequency of secondary excitation waves. Now, if
the ' all-or-none ' law be true, it is impossible for any number
of synapses to delay the excitation produced by a weak stim-
ulus more than that produced by a strong one. Consequently,
the assumption of resistance offered by the synapse cannot
account for the marked prolongation in either reaction time or
reflex time with decrease in intensity of stimulus.
It is true that fatigued nerve-fibers do not follow the ' all-
or-none' law. It might be held that, in a similar way, a
sensory nerve-fiber subjected to stimuli such as light or sound
would not follow this law. The high frequency of the
secondary oscillations, in the case of the excitations produced
in an optic nerve by a light stimulus, might result in the
nerve taking on the properties of what Verworn calls an
'heterbolic' system,4 and consequently in its showing variation
in magnitude of excitation with variation in intensity of
stimulus. However this may be, it is very improbable that
such differences in intensity of excitation of single nerve-
fibers (assuming them to exist) could account for the increase
1 In confirmation, see the following: Gotch, Journ. of Physiol., 1902, 395. Koike,
Ztsck. f. Biol., 1910, 310. Lucas, Journ. of Physiol., 1911, 46. Adrian, Ibid., 1912,
389; 1913, 384.
2 Gotch, Journ. of Physiol., 1902, 392. Symes and Veley, Proc. Roy. Soc., B.
kxxiii., 1910, 431. Verworn, Ztsch. f. allgm. Physiol., 1912, 277. Veszi, Ztsch. f.
allgm. Physiol., 1912, 321. Adrian, Journ. of Physiol., 1913, 389. Lodholtz, Ztsch. f.
allgm. Physiol., 1913, 269. Lillie, Amer. Journ. of Physiol., Vol. XXXIV., 1914, 410.
8 Ztsch. f. Sinnesphysiol, Vol. XLVIIL, 1914, 28-165.
4 Ztsch. f. allgm. Physiol., 1912, 289.
438 HERBERT WOODROW
in time of cessation reactions with decrease in intensity of
stimulus. This becomes very evident from a consideration
of reactions to a decrease in intensity, of which, after all,
cessation reactions are merely a limiting case.
The assumption of greater slowness of conduction of weak
excitations than of strong, even were it warranted, would
not account for the fact1 that a slight decrease in intensity of
a strong stimulus results in a longer reaction time than a
larger decrease in the same stimulus: for it cannot be held that
a slight decrease in excitation is conducted more slowly than
a large decrease. The decrease itself is not conducted at all,
but merely the excitation as it exists before and after decrease.
It may be urged that this argument is not pertinent, but I
think there is little doubt that just as the reaction time to the
beginning of a stimulus equals that to its cessation, so does the
reaction time to an increase in intensity equal that to a de-
crease of the same size. We know, at least, that the reaction
time, whether the reaction is to an increase or a decrease in
intensity, is lengthened as the size of the change in intensity
is decreased. Reactions to a decrease in intensity should, then,
be considered along with cessation reactions.
Another point to be considered, in addition to the effect
of intensity of stimulus upon reaction time, is the difference
in reaction time between sound and light. Reactions to a
moderately bright light are longer than those to a moderately
loud sound. This fact is not uncommonly explained by as-
suming that the latent period of stimulation of the sense-
organ is longer in the case of light.2 Against this assump-
tion, we may urge the following considerations:
First, since cessation reaction times are the same as be-
ginning reaction times in both sound and light, we would
have to assume that in both cases the latent period for the
1 See Woodrow, 'The Measurement of Attention,' Psychol. Monog., 1915, Chap.
IV.
2 Ladd and Woodworth write as follows : " . . . there seems good reason to suppose
that the reaction time of sight is necessarily longer than that of hearing or touch, on
account of the photochemical nature of its more immediate stimulus." "On the
whole, the suggestion which probably is most generally entertained ... is that already
adopted by us, — namely, that the inertia or latent time of different sense-organs differs."
'Elements of Physiological Psychology,' 1911, 472.
CESSATION OF STIMULI 439
dying out of the excitation just equalled that for its rise, that
is, not only assume that the photochemical process was longer
in getting started but also equally longer in dying down.
Second, if we assumed any considerable latent period of stim-
ulation, it would seem reasonable to suppose that such latent
period would increase markedly with weakened intensity.
But as has been pointed out in preceding sections, such an
increase in latent period is incompatible with the fact that
reactions to the cessation of a very weak stimulus are longer
than to the cessation of a strong one. At any rate, the effect
of intensity, which cannot be explained by reference to the
latent period, is sufficient to completely overwhelm the effect
of the latent period, since a very weak sound may give
much longer reaction times than a very bright light. Third,
in the case of reflexes, we know that different reflexes may vary
greatly in time even when they are elicited by direct stimula-
tion of the sensory nerve, so that the latent period of stimula-
tion of the sense-organ is not involved. Similarly, a great
variation in the time of different reflexes occurs where the
sense-organ stimulated remains the same. For example,
the latent time of the scratch-reflex is, on the average, very
much longer than that of the flexion-reflex of the same limb,
although the distance of nerve fiber conduction is not greater.1
Fourth, and lastly, it can be shown, as will be pointed out
below, that the degree of attention is less with light reactions
than with sound, and that this fact offers an intelligible
explanation of the variation in reaction time with mode of
stimulus. It is true that reflexes as a rule do not involve
attention, but since the work of Sherrington, showing the
important role of the central nervous system in determining
the characteristics of reflexes, it may very well be assumed that
a nervous mechanism analogous to the nervous mechanism of
attention, though less complex and without conscious ac-
companiments, is involved in every reflex.2
One other point may be mentioned as one which should be
1 Sherrington, op. cit., p. 21.
2 "The interference of unlike reflexes and the alliance of like reflexes in their action
upon their common paths seems to lie at the very root of the great psychical process of
' attention.' " Sherrington, op. cit., 234.
44° HERBERT WOODROW
illuminated by knowledge of cessation reactions. I refer to
the long time required by a weak intensity of stimulus to
produce its maximal intensity of sensation. The data here
presented do not deal exactly with this point, but the experi-
ment with the liminal sound as stimulus is of interest in this
connection. It is known that a sound of constant physical
intensity increases in apparent intensity up to a duration,
in the case of weak sounds, of 1.5 sees.1 It ought, then,
to be possible, by using a sound so weak as to be in-
audible until . nearly its maximum effect is produced, to
obtain a sound so weak that it could not be heard at all for
a large fraction of a second. The reaction time I actually
obtained for such a sound was about three fourths of a second,2
and the reaction time to its cessation was fully as long. In the
case of a liminal sound such as here used, there can be no
doubt that the reaction does not occur until the subject is
aware, in the case of beginning reactions, of the presence of
the sound, and in the case of cessation reactions, of its absence.
It is well known, apart from the present data, that it takes
longer to become aware of a weak stimulus than of a strong.3
It may be said, then, that the very long reaction time in the
case of beginning reactions shows how very long it takes to
become aware of a liminal stimulus. Moreover, in view of the
approximate equality of the beginning and cessation reaction
times, it may be concluded that just as it takes longer to
become aware of a weak stimulus than of a strong, so does it
take longer, and about equally longer, to become aware of the
cessation of a weak stimulus than of a strong.
Here again, we may raise the question of explanation by
resistance. And there can be little doubt that the explanation
of the period required for the rise of the excitation up to the
awareness threshold is closely connected, if not essentially
1 Kafka, 'Ueber das Ansteigen der Tonerregung/ Psychol. Stud., 1907, 256-292.
See also Sander, 'Das Ansteigen der Schallerregung bei Tonen verschiedener Hohe,'
Psychol. Stud., 1910, 1-38.
2 The great discrepancy between this value and the corresponding values obtained
by Wundt (337 o-) and by Pieron (361 a) is probably due to the stricter insistence upon
liminal intensity of stimulus in the present instance.
8 See Minneman, 'Untersuchungen uber die Differenz der Wahrnehmungsgeschwin-
digkeit von Licht und Schallreizen,' Psychol. Stud., 1911, 1-82.
CESSATION OF STIMULI 441
identical, with the explanation of the period required for the
further rise to a maximum. In the explanation by resistance,
in the present instance, the resistance would probably be
thought of as at the synapse; but no mode of behavior on the
part of the synapse has yet been suggested which would
explain how it can offer resistance both to the rise and to the
fall of the excitatory process. The cessation of the excitation
is not a new excitation as we know from the introspection that
the sound sensation merely ceases, — no new sensation is set
up, when the stimulus is cut off. So it is evident that neither
the long period required for the rise of the excitation up to the
sensation or awareness threshold, nor, in all probability, the
further rise to a maximum, can be explained by reference to
the concept of a surface of separation at the synapse or to
any other form of resistance to conduction along nervous arcs.
A consideration of the facts of cessation reactions, then,
leads to certain negative conclusions concerning the explana-
tory value of the concept of resistance, whether this resistance
is placed at the point of stimulation, at the synapse, or in the
whole neurone. Such resistance cannot explain the fact
that cessation reactions to a weak stimulus are longer than
to a strong one, but on the other hand would tend to cause us
to expect the reverse. Moreover, since both cessation and
beginning reactions follow the same law as regards variation
in time with variation in intensity of stimulus, we should
expect the same explanation in both cases; and so, if resistance
to conduction does not explain the one, it probably does not
explain the other. Again, as regards the difference in reaction
time between sound and light, this, likewise, does not seem
explicable by reference to the greater resistance of the retina
to stimulation. Further, since it takes as long to become
aware of the cessation as of the beginning of a very weak
sound, the long time required for the awareness of either can-
not be due to resistance to conduction. And, moreover, this
latter fact renders it probable that such resistance has nothing
to do with the explanation of the long time required for a
weak stimulus to produce its maximal sensory effect.
While the facts of cessation reactions thus serve to call
442 HERBERT WOODROW
attention to the shortcomings in the explanation of the be-
havior of the nervous system by means of the concept of
resistance, it must be admitted that there exists a large accu-
mulation of other facts which likewise greatly impair the
plausibility of such explanation. Among this mass of evi-
dence, we may cite the great disproportion and lack of cor-
relation between energy of stimulus and energy of response.
Even more conclusive is the temporal discrepancy. This dis-
crepancy is especially striking in cases where psychocerebral
processes are involved, but it is very evident even in reflexes,
in the phenomenon of the after-discharge. Concerning this
phenomenon, Sherrington writes as follows: "Tetanic contrac-
tion of the knee-flexor muscles of the dog induced by brief far-
adization of the motor-nerve usually ceases within 150 cr of the
cessation of the stimulation of the nerve, if crude condition of
fatigue, etc., be avoided. The contraction of these same
muscles, when induced reflexly by a similar brief stimulation?
often persists for 5,000 <r after cessation of the stimulus.1 Now
this great duration of the after-discharge certainly cannot be
explained by resistance offered at the synapse; and it is equally
certain, that resistance at the synapse cannot explain the rela-
tion between intensity of stimulus and both duration of after-
discharge and reflex latent time, since the duration of the
after-discharge increases with intensity of stimulus, while
the length of reflex time decreases. It would be easy to con-
tinue, indefinitely, these instances of the unsatisfactory nature
of the theory that the action of the nervous system is ex-
plicable merely by the assumption of a network of reflex
paths, which offers more or less resistance to the passage of
currents from the sense-organs to the muscle. However,
the following two citations will suffice to indicate the dissatis-
faction with such a theory.
Titchener writes as follows: "The assumption that the
reflex arc is the unit of nerve function evidently makes the
brain nothing more, in principle, than a mass of superposed
reflex arcs; the central is assimilated to the peripheral mech-
anism; the office of the brain is to perceive, to couple up, and
1 Op. tit., 26.
CESSATION OF STIMULI 443
to send out. But this view that the nervous system is a
system of conduction, a sort of glorified telephone exchange,
is in the author's opinion wholly inadequate to explain the
phenomena of mind. The theory of conduction, with ob-
stacles or easements between cell and cell, must, he believes,
be replaced by a theory of intracellular change, or change
within the cell-body; and if this is the case, the cortex must
be regarded rather as a disjunction of the reflex arc than as a
switchboard for the manifold connection of afferent with
efferent process."1
Reference may be made also to the views of the physi-
ologist, T. Graham Brown. From experiments which de-
monstrated that the phenomenon of "narcosis progression"
in the cat may occur at a depth of narcosis at which the spinal
reflexes are abolished, and from a general consideration of the
facts of rhythmic motor phenomena, Brown concludes: "The
fundamental unit of activity in the nervous system is not that
which we term the spinal reflex." He says that his experi-
ments "show the independence of the efferent neurone, and
suggest that the functional unit is the activity of the inde-
pendent efferent neurone; or rather, that it is the mutually
conditioned activity of the linked antagonistic efferent neu-
rones (' half-centers') which together form the 'center,' and
they also suggest that the primitive activity of the nervous
system is seen in such rhythmic acts as progression and
respiration."2
We may now take up the specific and difficult question of
how the results obtained with cessation reactions, discussed
above, are to be explained. We have seen that the concept
of resistance is inadequate. Is it possible to offer a more
satisfactory explanation? While in the present state of
nerve physiology it would require an unprofitable degree of
speculation to attempt a detailed explanation, it does seem
feasible to indicate at least the general direction which such
an explanation must take.
1 'A Text-Book of Psychology,' 1911, 489.
2 T. Graham Brown, 'On the Nature of the Fundamental Activity of the Nervous
Centers, Together with an Analysis of the Conditioning of Rhythmic Activity in Pro-
gression, and a Theory of the Evolution of Function in the Nervous System,' Journ. of
Physiol., Vol. XVIII., 1914.
444 HERBERT WOOD ROW
In the first place, it seems clear that a greater and more
distinctive role must be given to the central nervous system,
a role such that the activity of the central nervous system
cannot be said to be assimilated to the peripheral. To sub-
stantiate this proposition I shall refer to some results I have
already published elsewhere, on the effect of the preparatory
interval on reaction time.1 These results show that a much
longer reaction time is obtained with a set of totally irregular
preparatory intervals of varying length, or with very long
preparatory intervals, than with a regularly repeated pre-
paratory interval of two seconds. We may speak, then, of the
prolongation in reaction time produced by the unfavorable
preparatory intervals. Now I have shown that with weak
intensity of stimulus this prolongation is very much greater
than with strong or moderate intensity; and I have pointed out
at length,2 that this result can be explained only on the assump-
tion that in the case of the weaker intensity we are dealing with
a lower degree of attention. Moreover, it is known that with
weakened attention we have lengthened reaction time. We
have, then, experimental proof that a large part, at least, of
the increase in reaction time with decrease in intensity is due
to a decrease in the degree of attention. This conclusion is in
general agreement with the explanation long since advanced
by Wundt3 and by Martius.4 The basis of the explanation is
the fact that intensity of stimulus is a condition determining
the degree of attention; and I shall attempt to show how, on
this basis, we can explain the laws of cessation reactions as
well as those of ordinary reactions.
Another investigation which I desire to mention, is one on
the measurement of attention as it occurs in reactions to
sound, light, and touch, using as the measure of attention the
reciprocal of the prolongation in reaction time produced by
unfavorable preparatory intervals. While these results may
not be published for some months yet, it is perhaps permissible
1 'The Measurement of Attention,' Psychol. Monog., 1915.
2 Op. «'/., Chap. III.
3 ' Physiologische Psychologic/ 5th ed., 1903, Vol. III., p. 430.
4 'Ueber den Einfluss der Intensitat der Reize auf die Reactionszeit der Klange,'
Philos. Stud., 1892, 469-486.
CESSATION OF STIMULI 445
to state that, in the case of nearly all of the twelve subjects so
far measured, with whom over 15,000 reaction times have
been taken, the greatest prolongation occurred with light,
the next greatest with sound, and the smallest with touch;
from which I conclude that the degree of attention is in general
lowest in the case of light reactions, and highest in the case of
touch reactions. I am convinced from an analysis of the
data, that the difference in degree of attention in these cases
is sufficient to account for the difference in simple reaction
time that occurs with difference in mode of stimulus.
Befor making use of these results in the explanation of the
phenomena here under discussion, I wish to emphasize the
fact that the reactions in human reaction time experiments
are not at all simple reflexes. Of course, cessation reactions
can never be what we ordinarily think of as reflexes, for in
this case there is no stimulus, merely the cessation of a
stimulus. One cannot mince words in such a fashion as to
say that the cessation of stimulation is stimulation. The
reflex theory would seem entirely inadequate to account
even for the existence of cessation reactions. A similar point
could be made in the case of reactions to a mere decrease
in intensity. Moreover, the reaction is only in a very
secondary manner a reaction to the change in stimulus. It
is primarily brought about by the subject's intention to react,
which in turn is the result of acceptance of instructions given
the subject a considerable time before the reactions. It is
further evident, at least in the case of weak intensities, that
the reaction does not occur until the subject becomes aware
of the change in stimulus. The process of inhibition, so far
as can be judged either from introspection or from the ob-
jective reaction times, is present equally in cessation and
beginning regctions. In both cases the subject is prepared
to react when the right time comes, but this preparation re-
mains merely a preparation until then. In one case the 'right
time' means when the subject becomes aware of the stimulus,
and in the other case, when he becomes aware that the stimu-
lus has ceased. No satisfactory account of the reaction
process can be given without taking into consideration this
intention and preparation to react.
446 HERBERT WOODROW
In the present state of our knowledge of the physiology of
the nervous system, it is impossible to say definitely in what
the preparation to react consists. We know that, in practiced
subjects, it is often unaccompanied by any appreciable ten-
sion on the part of the muscles of the fingers used in the
reaction. It appears, therefore, not to consist in any actual
innervation of the specific reaction movements. We know,
too, that this preparation is constantly undergoing certain
irregular variations in degree, that it takes about two seconds
to reach its maximum, and that thereafter, in addition to
irregular oscillations, it shows a definite general tendency to
weaken. It should be emphasized, however, that a certain
degree of adaptation may be maintained throughout a con-
siderable period of time. We know, further, that this
preparation can vary in degree or intensity, and that the
greater the intensity of such preparation, the quicker the
reaction. Without attempting to go into details, it is evident
that the preparation to react means, on the nervous side, the
presence of certain nervous energies which are held in check,
so far as a reaction movement is concerned.
There is energy, the intimate nature of which we must
admit is as yet unknown, present in the central nervous sys-
tem and temporarily inhibited in some way from producing
the reaction. It is probable that this energy is of an electri-
cal nature, and that it always involves many, if not all, of
the neurones of the central nervous system. The condition
within any one neurone is to be thought of as interrelated
with the conditions within all the others, so that there is
always involved an immensely complicated and widespread
system of energies, including, perhaps, both potential chemi-
cal energies and electrical activities. So long as the subject's
state is merely one of preparation for the reaction, it must be
supposed that the interrelated cortical energies, while by no
means entirely suppressing each other (for the subject would
then be unconscious), yet, interact in such a way that some
sort of equilibrium occurs at the point of exit of the motor
pathway for the reaction movement. The existence of this
equilibrium, as respects the reaction movement, is evident
CESSATION OF STIMULI 447
merely because the reaction movement does not occur, ordi-
narily, until after the stimulus which comes at the end of the
preparatory interval.
The mere fact that both beginning and cessation reactions
are possible, shows that this equilibrium may be upset either
by the beginning of some new disturbance, or by the cessation
of one. There is no theoretical difficulty here such as we
encountered in attempting to explain both the laws of cessa-
tion and beginning reactions by reference to the concept of
resistance. Different instructions set up different central
adjustments, and while these adjustments in both cases con-
sist in a system of activities which are in equilibrium so far
as the reaction movement is concerned, the pattern of the
adjustment must be different in the two cases; so that, in one
case, the occurrence of a new excitation destroys the existing
equilibrium, while, in the other, the new excitation merely
serves to increase the intensity of activity of a different sys-
tem, the equilibrium of which, however, is not upset until
this excitation ceases. The widespread system of cortical
energies takes on such a form, as the result of the instructions,
that upon the occurrence of the proper sensory disturbance,
some of this cerebral energy is released for the work of inner-
vation of the reaction movement.
The energy used in the innervation of the reaction move-
ment is not the energy of the excitation set up in the sensory
nerve by the stimulus; it is energy already present in the cen-
tral nervous system before the stimulus acts. In a reflex,
as commonly described, the outgoing energy is merely the
redirected incoming energy. This remains true, even though
we further add that reflexes may inhibit or reinforce one
another. The present view, however, regards the central
nervous system as not merely a network of paths, but also as
a vast system of interrelated energies, potential and other-
wise. A disturbance produced in this system is not con-
ducted through it; it causes a readjustment of the system,
which readjustment may (or may not, so far as is known)
result in the release of energy for the work of motor inner-
vation. It is difficult to see how, on any other theory, we
448 HERBERT WOODROW
can account even for the fact of cessation reactions. It has
already been pointed out that it is impossible to think of the
cessation reaction as a simple reflex, for in this case there is
no stimulus, but merely the cessation of a stimulus. On the
other hand it is not difficult to understand that a change in
any part of a system of energies may upset the balance of the
system as a whole, whether the change in question be an in-
crease or decrease of activity.
The process by which the change in sensory excitation
disturbs the preexisting cerebral system is in part, no doubt,
the physiological process corresponding to becoming aware of
the change. Even if actual awareness were not a necessary
condition of the reaction, it seems altogether probable that a
nervous process identical in nature to that which underlies
awareness, though less in degree, must precede the reaction
movement. The process which underlies awareness of the
cessation of a stimulus must consist in an effect produced by
this sensory change upon more or less widely distributed
parts of the cerebral cortex. It could not consist in the local
sensory effect, for this is merely a cessation. It is probably
no less true that the awareness of the occurrence of the stimu-
lus involves changes in the entire cerebral activity. We
know from introspection that any new perception coming
into the field of consciousness ordinarily modifies the pattern
of consciousness as a whole.
It is clear that these central nervous processes in which
alone is to be found the explanation of the phenomena dis-
cussed in the preceding pages, are in large part identical with
those of attention or awareness; but that these central proc-
esses are all such that they may be subsumed under this
heading is improbable. They are quite likely very complex.
It may be, for instance, that in addition to the physiological
process underlying awareness, there occurs another process
which consists in the release of so-called determining tend-
encies.
We may now ask why a small change in the sensory ex-
citation, whatever this may mean in terms of nerve physiol-
ogy, should bring about the required disturbance in the pre-
CESSATION OF STIMULI 449
existing system of cerebral energies so much more slowly
than a large one, regardless of the direction of change. This
is satisfactorily explained as due to the inertia of the pre-
existing cortical adjustment. Of course we do not know the
intimate nature of the cortical changes brought about by the
sensory change; but that there is work to be done in bringing
about these charges, and that this work would be done more
slowly by a small local disturbance than by a big one, seem
to me unquestionable propositions. I have already argued
that the cortical processes in question are largely those under-
lying the phenomenon of attention. It is well known that
these processes require considerable time; e. g., it is well es-
tablished that maximal adaptation of attention requires about
two seconds. There is, moreover, no difficulty in under-
standing why the reaction time to the cessation of the ex-
citation equals the reaction time to its occurrence, no matter
what the intensity of the stimulus. The size of change in
sensory excitation is of course the same, whether the change
be due to the occurrence or the cessation of the stimulus. The
direction of change of the excitation need not affect either
the size or rate of change. The rate of change in the
sensory cortical excitation, then, we assume to be prac-
tically the same at both the occurrence and cessation of the
excitation; and, further, we assume that so long as the in-
tensity of the preparatory adjustment remains constant, the
time required for the disturbance or release of the preadjusted
system of cortical energies varies with the size and rate of change
in excitation, which brings about this release. The difference
in direction of the change in excitation is counterbalanced
by the difference in the nature of the preadjustment.
The next question to be answered is why light reactions, at
moderate intensities of stimuli, are longer than sound re-
actions. This is easily accounted for on the hypothesis that
the cortical preadjustment is less effective in the case of
light than in the case of sound. The lesser effectiveness of
preadjustment, in the case of light reactions as compared with
sound or touch reactions, may possibly be accounted for by
the fact that the adjustment of the sense-organ in the case of
45° HERBERT WOODROW
vision — including the control of convergence, accommodation,
winking, proper direction of head, etc. — is more elaborate
than in the case of hearing or touch, where no sensory ad-
justments of any great consequence are required. The ex-
penditure of energy on the sensory adjustment may detract
from the adequacy of the central cerebral adjustment as a
preparation for the reaction movement. Whatever be the
cause of the better cerebral adjustment in the case of sound
and touch reactions, the fact of such better adjustment is
evident; for upon no other assumption can we explain the
result already mentioned, that light reactions are prolonged
more by the use of unfavorable preparatory intervals than
are sound or touch reactions. I have shown elsewhere, that
the prolongation produced in reaction time by unfavorable
preparatory intervals is due, solely, to the effect of such as a
detractor of attention. This detraction effect occurs in ac-
cordance with the law that the prolongation produced varies
inversely with the degree of attention acted upon. Accord-
ingly, since the prolongation in reaction time produced by
unfavorable preparatory intervals is greater in the case of
light reactions than in touch or sound, it follows that the
degree of attention in the former is less than in the latter.
On the cerebral side we may substitute effectiveness of cerebral
adjustment for degree of adaptation of attention.
CONCLUSIONS
1. The experiments here reported show that the reaction
time to the cessation of a sound or light stimulus is in all
cases the same as the reaction time to the beginning of that
stimulus. This statement includes the following two corol-
laries, (i) The same difference between reaction times to
sound and light exists in the case of reactions to their cessa-
tion as in the case of reactions to their beginning. (2) The
lengthening in reaction time due to a decrease in intensity
of stimulus is equal in amount, and follows the same law, in
the case of both beginning and cessation reactions.
2. The above mentioned similarity, in the variation of
both beginning and cessation reactions with mode and with
CESSATION OF STIMULI 451
intensity, cannot be explained either by the hypothesis of
variation in the latent period of stimulation of the sense-
organ, or by the hypothesis of variation in the resistance
offered by the synapses. Likewise, this similarity cannot be
adequately explained on the commonly accepted hypothesis
that all nervous action consists essentially in conduction of
nervous impulses from the sense-organ to the muscle along
arcs which offer varying amounts of resistance. On the other
hand, the facts on beginning and cessation reactions are in-
compatible with such an hypothesis.
3. The explanation of the experimental data presented in
the preceding pages, while at present impossible in detail,
seems to require us to regard the central nervous system as not
merely a network of paths, but also as the seat of a complex
system of interrelated activities and potential energies which
is disturbed throughout by any change in any part of the
system. The fact of cessation reactions cannot be adequately
explained without postulating such a central system of ener-
gies, the balance of which may be upset by either an increase
or decrease of activity in any part of the system. The
pattern of the preexisting system differs in beginning reactions
from that in cessation reactions in such a manner that, in
beginning reactions, the same effect is produced by the rise
of the excitation as is produced in cessation reactions by its
fall. A small change in excitation disturbs the preexisting
cortical system so as to bring about the reaction movement
more slowly than does a large one — not because of resistance
to its conduction, but because of the inertia of the preexisting
central system. The reaction time to light, both for cessation
and beginning reactions, is longer than for sound because of an
inferior preadjustment of the cerebral mechanism. In all
cases the reaction results from the release of energy already
within the nervous system before the occurrence of the
stimulus; and is not due to the mere redirection or modifica-
tion of the incoming sensory excitation.
4. The physiological disturbance of the central nervous
system here involved is in large part, though not entirely,
that which underlies the process of becoming aware of a
45 2 HERBERT WOODROW
stimulus or of attending to it. This is shown by the fact that
the degree of attention, in the case of reactions to a weak
stimulus, is less than that in the case of reactions to a strong
stimulus, and also, less in the case of reactions to light than
in the case of reactions to sound or touch.
FACILITATION AND INHIBITION OF MOTOR
IMPULSES
A STUDY IN SIMULTANEOUS AND ALTERNATING FINGER
MOVEMENTS
BY HERBERT SIDNEY LANGFELD
Harvard University
DESCRIPTION OF EXPERIMENTATION
The purpose of the investigation1 was to ascertain the
facilitating and the inhibitory effect of successive and simul-
taneous muscular impulses in the movements of the several
fingers of both the right and left hands. The data have also
been arranged to show the effect of practise and fatigue and
the relation of the movements of the fingers of the right hand
to those of the left hand.
The instrument used was similar to that of the Whipple
tapping-board2 as will be seen from the cut, p. 476. In place
of a metal stylus, a ring was worn on the finger. This was
insulated by rubber tubing except on the under part where
a metal peg protruded and made an electric contact with the
metal base of the board. A light flexible wire ran from the
ring to an electric marker which registered the contact on a
kymograph.3 In order to obtain a free and natural movement
of the fingers, the hand rested on a curved block of wood,
and was thus slightly vaulted. The tips of the fingers and
cushions of the palm of the hand rested on the board. The
arm rested on the table, several of the subjects using a cushion
for greater comfort.
It was deemed of importance that the height of the finger
1 The experiments were conducted in the Harvard Psychological Laboratory
during the academic year, 1913-14. Trial experiments, not here reported, were made
the previous year.
2 'Manual of Mental and Physical Tests, Simpler Processes,' p. 131.
8 It was under similar conditions that one of the first tapping tests — that of von
Kries — was performed, 'Zur Kenntniss der willkiirlicher Muskelthatigkeit,' Arch. f.
Anatomic u. Physiologie, 1886, Sup. Band, pp. 1-16.
453
454 H. S. LANGFELD
movement should be constant so that there could be no doubt
that differences in the rate of tapping were not due to changes
in the length of stroke. A bar was therefore placed 4 cm.
above the board and the subject was instructed to hit the
bar at each stroke. The bar met the finger slightly below
the second joint, and care was taken that it should always
strike at approximately the same spot throughout the tests.
This arrangement worked very well, the subject soon becom-
ing used to the task. The movement thus remained more
nearly voluntary than would probably have been the case
if the finger had not had to touch the bar.1 The hand was
not strapped to the board as it was soon evident that the
subject could keep his hand and arm still, and if there did
happen to be a slight movement, his attention was called to
it at once; nor was the tapping continued to that period of
fatigue when the subject in seeking relief begins to use other
muscle groups of the arm. Any slight error that might have
crept in could not be as great as would have been the danger
to the result caused by the binding of the muscles. The
only muscles used, therefore, were the extensors and the
flexors of the phalanges. These muscles pass from their
origin in the forearm over the wrist joint to the phalanges
and at the end of the grasping movement there is a tendency
for them to flex the hand, which is inhibited by the synergic
muscle of the wrist.
The time in two-second periods was marked on the drum
at the beginning of each set of trials by a marker actuated
1 In regard to the length of stroke, von Kries writes, loc. cit., page 4, "Der Umfang
der Bewegungen ist auf die Dauer von nur geringem Einfluss; doch scheint es, dass die
Bewegungen von einem gewissen mittleren Umfamg am schnellsten ausgefiihrt werden
konnen und sowohl die sehr kleinen als die sehr grossen ein wening langer dauern."
Bryan ('On the Development of Voluntary Motor Ability,' Amer. Jour, of Psychol.,
5, 1892, p. 150) is in agreement with von Kries as to the slight effect of change of am-
plitude. Max Isserlin states that "... die Tendenz besteht, trotz abnehmender
Geschwindigkeit die Bewegungszahl konstant zu erhalten. Diese wird zuletzt herab-
gesetzt" ('Ueberden Ablauf einfacher willkiirlicher Bewegungen,' Psych. Arbeiten, 6
1910, p. 186). From this we may conclude that the change in amplitude conceals the
fatigue as measured by the rate of tapping alone. It was also found in the preliminary
tests before the control bar was used that there was a strong tendency for several
subjects to execute a series of quick reflex movements similar to a tremble which greatly
increased the tapping rate and seemed difficult at times to avoid.
FINGER MOVEMENTS 455
by the laboratory clock. When both hands were used, a board
was employed for each hand.
There were four subjects who will be referred to as A, B, C,
D. A and B were advanced graduate students; C and D under-
graduates. A and D were very athletic, B less so and C the
least strong of the four. The experiments were made in the
morning, the subjects coming always at the same hour. As
a rule, there was a week's intermission between each set of
trials. The period of tapping for all fingers and all combina-
tions of movements was 30 seconds with a two seconds'
pause between the members of a series and a five minutes'
pause between the series. The finger movements examined
were as follows: During the first half year, the subject began
by tapping with his right index finger (Ri). This was
followed by the second finger of the same hand (R2). Then
these two fingers were tapped alternately (A), a movement
similar to the walking movement, that is, Ri was raised as
R2 was lowered, the two fingers passing each other in the
middle of the stroke. There then followed what may be
termed complete alternation (CA). Ri made a complete
stroke up and down before R2 began. The signal for a finger
to begin was the return of the other finger to the starting
point in the manner of a relay race, Ri went up and down
then R2 went up and down, etc. Finally Ri and R2 tapped
simultaneously [S(Ri R2)]. This completed the series. After
a five minutes' pause, the series was repeated in the same
order. On alternate weeks, the order of the series was
reversed, beginning with S and ending with Ri. At the
beginning of the year the left-hand fingers were tapped in
the same manner as the right and again for one series about
six weeks later, in order to ascertain if there was any transfer
of practise. During the second half of the year, both hands
were used. The series began with the right index finger
alone. Then followed the left index finger alone (Li) and
then both simultaneously [S(Ri Zi)]; and then the second
finger of the right hand (R2) alone for thirty seconds, followed
by the second finger of the left hand (Z,2) alone and then both
simultaneously [S(R2 L.2)]. This series was repeated and
456 H. S. LANGFELD
the order reversed on alternate weeks. A few series were
also made with Ri and L^ and [S(Ri £2)] and Ri and L\
and [S (Ri £4)].
SIMULTANEOUS MOVEMENT
The final averages have been gathered of the simultaneous
tapping of the pairs of fingers of both hands and have been
placed together in Table IV.1 for convenience of comparison.
In the second and fourth columns for each subject are the
rates of tappings for 30" for the fingers separately, and in
the sixth column, the rates when tapped together. In the
seventh column is the difference in rate of the two fingers,
and in the eighth, the difference between the rate of the
simultaneous tapping and the rate of the slower finger when
tapping alone. Examining first the two fingers of the same
hand, J^i and .#2, we find the interesting fact that, with all the
subjects the two fingers are moved more rapidly together
than the slower finger alone and in the case of A faster even
than the faster finger by a considerable amount. With B, S
is one stroke faster than the faster finger. A similar relation
holds with the symmetrical fingers of the left hand (second
horizontal column). Here in fact in the case of all but J?,
who shows little change, the two fingers are tapped faster
together than either of the separate fingers. In Tables I.,
II. and III. the maximal rates of tapping are in heavy type and
may be readily compared. For the right hand the highest S
is greater than the highest Ri or R2 for all subjects but C.
For the left hand the highest S is greater than the highest Li
or L2 for all the subjects. The explanation which suggests itself
is that the two fingers being very closely related, an impulse
to one tends to influence the other, since a strong coordination
has probably been induced by the grasping movement. In
the single movement the other finger must be voluntarily
held down and this slows up the action of the moving finger.
When both fingers are moved together this inhibition is
removed and they both move faster, unless one is much slower
than the other, when it acts as an inhibition. It is possible
1 These averages have been taken from Tables I., II. and III.
FINGER MOVEMENTS
457
that with the left hand, the inhibition of the idle finger is
more difficult than with the right hand due to less practise.
This would account for S being greater than the single tapping
TABLE I
Subject A
Subject B
Subject C
Subject D
No.
Ri
Ra
A
c,|,
Ri
Ra
A
CA
S
Ri
Ra
A
CA
S
Ri
Ra
A
CA
5
I
154
I6S
120
37172
137
116
36
163
178
I48
44
23
160 139
161
106
25
172
2
ISS
170
121
39
192
168
I4I
131
34
170
161
162
24
174 149
158
112
34
i74
3
153
I6S
117
34
187
H3
I38
131
4i
161
119
121
68
27
H4 137
152
1 08
32
151
4
I84
IQO
126
40
196
172
161
136
40
iSS
115
117
31
132 147
162
112
36
157
5
H5
153
118
40
178
173
152
140
35
170
174
162
57
29
175 HI
I48
I04
34
154
6
146
115
46
187
170
164
H3
42
188
193
ISI
67
27
190 136
I46
102
36
138
7
159
171
119
43
167
178
160
133
39
166
191
168
56
26
168 133
152
109
35
135
8
170
166
136
47
197
166
170
141
42
164
194
165
66
26
174 136
153
103
40
134
9
137
145
1 20
43
163
160
151
127
40
181
178
161
54
24
167 121
I46
103
35
148
10
154
143
130
44
179
189
171
168
59
187
159
59
25
186
145
105
36
148
ii
ISS
159
131
37,150
178
178
133
37
173
142
168
H3
40
146
12
160
150
137
40 168
192
186
154
45
181
141
160
IOQ
40
146
13
141
136
139
41 176
169
161
142
33
!69
145
156 105
42
148
14
159 ISO
137
46169
170
170
138
40,177
129
158114
34
148
IS
157
161
133
45
161
178
178
ISS
44
1 88
129
151 106
36
138
16
IS4
150
134
52
182
1 80
175
153
53
178
135
148 101
32
149
17
1 86
177
iSi
4i
187
139
162 103
38
158
18
185
185
144
40
198
143
153 "0
35
155
Av.
ISS
158
127
42
176
173
164
141
41
174
169
iSi
59
26
164
I38
154
107
36
ISO
Ri
toS
149
iS3
125
42
177
174
160
140
40
178
178
157
56
25
175
135
153
106
35
154
S to
Ri
161
164
129
42
176
173
168
142
42
170
159
142
63
27
147
I38
156
108
36
144
m.v.
7
10
8
4
ii
9
13
9
4
9
22
H
6
2
17
6
5
3
3
8
of either finger for the two subjects who did not show this
with the right hand.
Is, however, the release of inhibition due to the close
relationship of the fingers the only factor which causes the
TABLE II
No.
Subject A
Subject B
Subject C
Subject D
Li
L2
A
CA
5
Li
La
A
CA\ S
Li
La
A
CA
5
Li
La
A
CA
5
I
1 10
128
46
27
135
134
131
III
32*131
"5
"3
44
24
116
134
120
90
30
156
2
"S
140
55
36
164
139 H3
119
35 HO
III
H7
58
27
128
137
132
105
34
H3
3
107
128
56
30
133
143,142
119
34 131
114
117
8z
22
125
137
135
104
38
127
4
"3
125
68
37
141
154 146
114
36 142
H3
"3
71
25
120
151
139
106
35
HI
5
107
122
59
32
133
149 I48
US
44 155
1 10 109
S3
28
118
137
I2S
97
34
134
6
H3
119
68
36
140150149
H3
49 164
no
I2O
59
28
121
I36
133
98
38
142
Av.
112
127
59
33
141 145
143
US
38 144
112
"5
61
26
121
139
131
IOO
35
I40
m.v.
5
5
6
3
7l 6
4
2
5 10
2
3
10
2
3
4
5
5
2
7
H. S. LANGFELD
increase especially in the slower finger? Does not one
impulse directly influence the other when discharged simul-
taneously, not only exciting an inhibitory effect in the case
of the slower movement, but a facilitating effect in the case
of the faster? To answer this question symmetrical fingers
of the two hands, Ri and Li and R2 and Z,2 were tapped simul-
taneously. Here there cannot be the same strong natural
tendency to move the two fingers simultaneously as is prob-
TABLE III
Subject A
Subject B
Subject C
Subject D
No.
Ri
LI
s
Ra
La
5
Ri
Li
s
Ra
La
^
Ri
Li
S
Ra
La
s
Ri
Li
s
Ra
La
I
I46
117
137
H7
124
1.1.1
172
146
111
170
143
163
179
107
118
168
114
1.33
130
124
128
111
128 I
2
117
119
110
111
123
184
1.13
167
177
111
174
191
114
127
167
116
122
131
I2S
128
111
1321
3
164
128
140
117
ii;6
173
144
163
174
I48
ill
190
122
171
124
139
128
127
112
1301
4
119
Ht;
149
168
140
H8
193
146
169
1 8O
I48
114
200
116
123
1 80
126
147
140
133
I64
1341
S
131
129
1 16
131
174
110
177
I48
161;
173
123
131
169
117
131
133
132
131
113
6
160
131
146
173
141
112
1 80
H8
162
174
114
161
192
in
124
184
124
141
138
136
130
H8
140
7
169
117
1.12
170
1.1.1
192
159
17.1
18.1
1.14
1.19
180
H.1
123
169
114
120
142
132
132
1.13
127
8
177
130
144
160
132
110
203
113
167
186 160
170
172
116
123
120
IOI
134
142
144
136
113
131
9
167
133
163
163
1.10
1 80
i.17
1 86 149
1.16
182
123
134
I69
12.1
134
1.38
132
128
1.17
132
10
162
138
160
132
1 66
187
148
1.17
172
146
1,1.3
204 119
122
162
III
118
138
129
130
1.13
I.3I
ii
IS8
H9
T43
167
123
151
193
1.13
170
1 80
148
162
176
121
120
1.16
118
124
1.3.1
140
i33
12
162
125
110
i?4
i,11
200
148
178
185
112
182
179
124
132
164
135
133
143
133
141
163
137
J3
171
141
162
173
112
161
H
1 86
152
162
172
I58
160
Av.
160
125
H3
161
129
155
185
150
164
178
ISI
162
185
117
125
165
118
130
138
133
131
155
1331
Ri toS
155
129
140
158
I3I
156
179
150
161
I7S
ISO
162
187
116
126
170
118
130
135
129
129
154
1331
S toRi
166
122
146
164
129
154
192
ISO
170
180
152
163
181
118
124
160
118
130
140
136
i34
156
1321
m.r.
7
6
7
8
5
4
8
4
5
S
4
6
9
4
4
10
6
7
4
S
3
3
4
ably the case with the fingers just examined. In Table IV.
we find that with three of the subjects for both sets of fingers
there is an increase in the tapping of the slow and a decrease
in that of the fast finger. In the case of both A and B the
simultaneous tapping approaches the average of the two
fingers. For C the effect of the faster finger is not so great
and the increase of speed of the slower finger is below that of
A and B. A further peculiarity of this subject to be discussed
later offers an explanation for this. For D the simultaneous
tapping for both pairs is about the same as the slower finger.
The difference between Ri and Li, however, is very slight
FINGER MOVEMENTS
459
and between R2 and Z,2 less than with any of the other sub-
jects. This difference is an important factor as will be seen
below (p. 460). It is evident, however, that with three of
the subjects the rapidity of the movement is increased by the
simultaneous exercise of a more rapid movement taking place
in a symmetrical part of the opposite side of the body.
The more rapid movement, on the other hand, is to an extent
inhibited.
Will this phenomenon occur if the fingers moved are not
symmetrical? To investigate this point Rl and Z,2, R2 and
Zri, Ri and £4 and Li and R^. were tapped simultaneously.1
TABLE IV
Subject A
Subject B
Ri
155
R^
158
S
176
3
21
Ri
173
R2
164! 5
174
9
10
Li
112
L2
127
S
141
IS
29
Li
145
L2
143
S
144
2
i
Ri
160
Li
125
S
148
35
Ri
185
Li
150' S
164
35
14
R2
161
L2
129
S
155
32
26
R2
178
L2
151! 5
162
27
ii
Ri
161
L2
140
5
132
21
— 8
Ri
181
L2
150; S
160
IO
R2
157
Li
132
S
132
25
0
R2
174
Li
149
S
158
25
9
Ri
160
1*4
89
S
93
71
4
Ri
177
L4
133
S
144
44
ii
R4
107
Li
140
S
109
33
2
R4
H5
Li
152
S
150
7
5
Subject C
Subject D
Ri
169
R2
151
S
164
18
13
Ri
138
R2
154
5
150
16
12
Li
112
L2
lie
S
121
3
9
Li
139
L2
S
140
8
9
Ri
185
Li
117
S
125
68
8
Ri
138
Li
133
S
5
— 2
R2
I6S
L2
118
S
130
47
12
R2
155
L2
133
S
133
22
O
Ri
193
L2
I2C
S
129
64
4
Ri
135
L2
139
S
127
4
—8
R2
I67
Li
126
S
120
41
— 6
R2
152
Li
139
S
138
13
— i
Ri
195
L±
98
S
95
97
— 3
Ri
141
L±
84
S
in
57
27
R4
Li
119
S
112
6
— i
Ri
98
Li
139
S
106
8
In the combination Ri 1,2 both B and C and in R2 Li, B still
show an increase in the rate of the slower finger. A has now
dropped below the single tapping for Ri Li and D is below
for both. With Ri £4 and R^ Li all the subjects but C show
an increase in the slower movement. C's simultaneous move-
ment is slightly below that of the slower finger.
In order more readily to examine and analyze these results
the difference between the tapping rates of the two fingers
1 It has not been thought necessary to give a complete table of these tests. The
averages were taken from fewer series than the previous ones, but as the general re-
lationship does not vary materially from series to series they can be safely used.
4^0 H. S. LANGFELD
has been placed in the seventh column of Table IV. and the
increase in the rate of the slower finger during the S move-
ment in the eighth column. A minus sign, of course, indicates
a decrease.
Examining first .5's result we find that in the asymmetrical
pair Ri Li there is less of an increase of the slower movement
than with the symmetrical pair Ri Li, an increase of ten
as compared to fourteen, but the pair Ri £4 which is more
asymmetrical than jRi Li shows a slightly greater increase,
i. <?., eleven compared to ten. In this latter case, however,
the difference between the rates of the two fingers Ri and £4
is greater than between the former pair Ri Li. The former
is raised one third of the difference, the latter only one fourth
of the difference. The result suggests that there are two
factors influencing the rapidity of the simultaneous movement,
the degree of symmetry and the difference in the rapidity of
the two members of the pair. These two factors should act
in opposite directions; the difference between the rate of the
two fingers increases as a rule with the decrease in symmetry,
and the greater this difference in rate the more should the
slower finger be aided by the faster in simultaneous tapping,
but the greater the assymmetry the less is the advantage of
simultaneous tapping. The relation of these two factors very
likely differs in individuals. When the coordination is not
good asymmetry probably plays an important role in slowing
the movement. When the coordination is good the rate
differences have more effect. Let us examine the data further
with this suggestion in mind. Take for example B's Ri Li
and R2 L2. The asymmetry is the same but the difference in
rate of the R2 Li pair is less than that of the Ri Li and con-
sequently the increase of the slower movement is less. With
Ri £4 the asymetry is increased but the difference rate is also,
so that the actual increase in rate remains the same as the
other pair. In the case of A with the same two pairs it is
true that the results are in the opposite direction, but in the
next pair there is a drop in both symmetry and difference and
there is in consequence, a falling off in the rate of simultaneous
tapping in the one case even below the slower of the pair.
FINGER MOVEMENTS
In the most asymmetrical pair, Ri £4, the difference is very
large and there is again an increase of the slower movement
notwithstanding the great asymmetry. In £4 Li, a pair of
like asymmetry, there is less difference and less increase.
Turning to Z)'s record, we find that although, like the other
subject, he showed an increase when the two fingers were of
the same hand whether the right or the left, as soon as the
movements are on opposite sides there is often evidence of an
inhibition. In the Ri Li pair we should not expect much
change for there is little difference between the rates, but
with the R2 Lz pair, although the difference is twenty-two
taps there is no increase of the slower finger, and in the Ri L^
pair the lack of coordination actually causes an inhibition of
the slower movement amounting to eight taps. The results
for Ri L\ and R^ Li taken in connection with the foregoing
results of this subject, speak strongly for the assumption of the
above mentioned opposing factors. The pairs are the most
asymmetrical but the differences in rate are very large, being
for one almost three times as great as the largest previous differ-
ence. Examining the rate for simultaneous tapping we find
that the slower movement increases by twenty-seven taps
for Ri £4, that is, the facilitation is 31 per cent, of the rate
when the finger is moved alone, and for R\ Li there is an
increase of only eight but the difference is less, forty-one
compared to fifty-seven. The inhibitory effect of asymmetry
which, judging from the previous results, is most probably
present, has been more than overcome by the facilitating
effect of the rate difference. This explanation also fits C's
result although he differs in type from the other subjects.
It will be found when we come to the further test performed
that C showed much more pronounced lack of coordination
of different muscle groups than the other subjects. There
should therefore be less facilitation and probably even in-
hibition. But he also showed the largest rate difference.
Therefore, although the facilitation is less than in the case of
A and B, nevertheless it is present in some instances. In
the pair Ri L\ the large difference of ninety-seven was not
sufficient to overcome the inhibition of asymmetry and when
462 H. S. LANGFELD
we notice that with the much more symmetrical pair Ri L.2
the large difference of sixty-four was only able to cause a
facilitation of four taps, and with R2 Li there was a difference
of forty-one and yet a decrease of six taps, this result is
rather to be expected. The only figures which do not readily
offer themselves to the explanation here attempted are those of
the pair R2 L.2. The symmetry of R2 Z,2 is the same as Ri Li
but a difference of forty-seven in the former pair causes a
facilitation of twelve taps while a difference of sixty-eight
in the latter pair only causes a facilitation of eight. It is
true that with none of the subjects is the relation of facilitation
to rate difference constant. This ratio also varied with the
different subjects. Two points might be mentioned in this
regard. First, equality in symmetry does not necessarily
mean the same amount of coordination of different muscle
impulses. Thus in the results of C, although Ri Li and R^ L^
are both symmetrical pairs the coordination between R2
L.2 is better than between Ri Li, and perhaps for that
reason the smaller rate difference has a greater facilitating
effect. This explanation could also be offered in regard to
the similar results of A. Secondly, although the change in
symmetry between any two pairs is naturally the same for
all subjects, yet one subject will probably have a different
change in coordination in going from one pair to another,
than a second subject, and this will readily explain individual
differences in the above mentioned ratio. It should also be
mentioned that investigations on other subjects revealed a
difficulty to synchronize, which retarded the simultaneous
movement.
ALTERNATING TAPPING
As was stated above the alternating tapping was performed
in the same series as the simultaneous and the figures for Ri
and R2 may again be used. A is the alternation in which
one finger ascends while the other descends and CA the com-
plete alternation in which one finger does not begin to move
until the other has returned. The figures express only the
rate of one finger, the number of actual taps made being
twice that number. In Table V. these alternations are
FINGER MOVEMENTS
4*3
expressed in per cent., of the average of the two fingers tapping
singly. Let us first discuss the A results. If the fingers were
alternating levers of a mechanical machine there would, of
course, be twice as much work done in the same time as one
lever working alone would perform. In the human machine
when two different movements, and in this case opposite
movements, are carried on simultaneously we look for some
TABLE V
Right Hand
Left Hand
A
CA
A
CA
Rl. + /?2
Ri + R2
Li + Lv
Li + L*
a
2
2
2
Subject A . . . .
Subject B
Subject C
Subject D
.81
.83
•37
•73
.27
.24
.16
•25
• SO
.80
•54
•74
.28
.26
•11
inhibition. The amount of this inhibition is expressed in
the per cent. It will be seen that for three of the subjects
A, By and Z), with the right hand there is only a loss of about
one quarter of the speed of one finger when working alone.
In other words by carrying out simultaneously two movements,
though opposite in nature, there is a gain of fifty per cent, as
compared to the amount of work done if only one movement
was performed in the same time. The other subject, C,
shows a much lower figure. It is only thirty-seven per cent.,
which means that there is an actual loss in work accomplished
by alternating simultaneous movements of thirteen per cent.
In his present state of muscular coordination he would accom-
plish more with one finger moving alone than he would by
moving two fingers. Fifty per cent, would mean that the
work accomplished by the two fingers is the same as if one
had moved alone. The actual figures make the above
perhaps clearer. Alternating he only taps fifty-nine times
for each finger, or one hundred and eighteen taps in all, while
the mean of the rate of the two fingers tapping alone is 160.
The loss is forty-two taps or twenty-six per cent, for the two
fingers. This subject is the one referred to on page 461 as
having poor coordination between different muscle groups.
464 H. S. LANGFELD
The above figures make this evident. In this connection it
is very interesting to note that C's rate of tapping with one
one finger is faster than any of the other subjects with the
exception of B. From this it would seem that the inhibition
is a variable independent of the rate of movement of the
separate muscle groups. This assumption is strengthened
by the results of the left hand. The separate tapping rates
are much lower but the A rate is about the same as before,
consequently the per cent, is higher. Subject A shows this
same independence. The A rate for the left hand has dropped
relatively much lower than the separate rates and the per cent,
is consequently lower. In fact it is now about the same as C's
rate. That is, his coordination in the left hand is worse than
in the right. Subjects B and D show almost identically the
same amount of coordination for both hands. To repeat,
the above results substantiate a fact which from what we
already know is rather obvious, that the degree of coordina-
tion between several muscle groups does not bear any definite
relation to the degree of efficiency of the separate muscle groups
concerned.
Particularly in regard to the coordination of these move-
ments it seems of interest to inquire into the musical training
of the subjects. A is proficient with the violin, B has played
the organ since boyhood, C has just begun to take piano
lessons, and D has played the piano for years for his own
amusement. $'s ratios of 83 per cent, and 80 per cent, are
the highest of any subjects and one is disposed to say that
this is due to greater practice and that C's low ratio is due to
lack of training. A's figures are what one would expect.
Being a violinist the fingers of the left hand are trained to a
different set of movements. In playing the fingers are bent
and one is held down while the other taps. There would
therefore be an inhibition when the fingers were forced to tap
alternatingly. This would account for the 50 per cent, which
is even lower than that of the untrained C. These results,
then, seem to indicate degrees of practice but they are too
few to be more than a suggestion.1
1 O. Raif argues that the fastest rate required by any piece of music is slower than
the average rate of tapping, and therefore piano practice does not increase the rate
FINGER MOVEMENTS 465
There remain to be examined the results of complete
alternation (CA). Again using the illustration of a mechan-
ical machine the one lever begins to move after the other has
stopped. The work done in a given time is the same as if
there were only one lever which moved continuously instead
of alternating with the second lever. Each ringer carrying
out such a movement should do fifty per cent, of that which
it would do if working continuously. Instead it will be seen
that all the subjects with the exception of C do only about
twenty-five per cent, whether with the right or left hand.
C's loss is again greater than that of the other subjects for the
right hand. This difference is less than in the A rate and with
the left hand he shows almost the same per cent, of loss as
they do. These figures mean that in this alternation there is
a fifty per cent, loss in muscular work done. This amount of
loss seems to be relatively the same for both the right and
left sides.
INDEX OF RIGHTHANDEDNESS
The indices of righthandedness of the first and second
finger in the performance of the different combinations here
investigated are given in Table VI. As has been done both
by Woodworth and by Wells1 the index is obtained by dividing
the rate of tapping of the left hand by that of the right, thus
giving the ratio of the efficiency of the two sides. .As has
been found in the tapping with the whole hand there are great
individual differences. The range is even greater than that
found by Wells. It is significant that subject C, who showed
poor coordination in the more complex movements has also
the lowest LijRi index which means that he is also relatively
inferior to the others in this simpler coordination for the left
but rather the proper timing of the movements. He says: "Nicht in der Bewegung an
sich, sondern in der Rechtzeitigkeit der Bewegung, d. h. in dem Zeitverhaltniss von
einer Bewegung zur anderen liegt die Schwierigkeit. Diese Rechtzeitikgeit kann zwei-
fellos nur ein Product unseres Willens sein, wir haben also den Ausganspunkt fur die
Fingerfertigkeit in den Centraltheilen unseres Nervensystems zu suchen, etc." (' Ueber
Fingerfertigkeit beim Clavierspiel,' Zeitschrift fur PsychoL, 24, 1900, p. 354.) While
upon the subject of characterization it is worth mentioning that B has the fastest
tapping rate and D the slowest, and both are very athletic, as was mentioned above,
which means that here there is zero correlation between strength and rate of tapping.
1 ' Normal Performance in the Tapping Test,' Am. Jour, of Psychol., 19, 1908^.446.
466
H. S. LANGFELD
hand. The indices for the efficiency during the first and second
half year's work have been given separately under each
subject, and we notice that the change is not great. Of these
indices five are slightly lower, two are the same, and one higher.
That is, there is an indication that practice has had somewhat
more effect upon the right hand than the left.2 The
m-
TABLE VI
Sub-
Li
Lz
(L)A
(L)CA
Sub-
Li
£2
(L)A
(L)CA
ject
Ri
R2
WA
(R)CA
ject
fa
J?a
(R)A
(R)CA
A. ..
.72
.80
.46
.78
c.
.66
.76
I O3
I O±
.78
.80
.63
•71
*>VT
B...
.83
.87
.82
.92
z>....
1. 00
.8<J-
•93
•97
.81
.85
.96
.86+
dividual characteristics, however, remain unaltered, the
ranking of the subjects according to the size of index being
the same. With three of the subjects the Li/Ri index is
somewhat lower than the L^\R^ index due to the superiority
of the efficiency of the index finger of the right hand. This
is not the case with subject Z), whose R2 is throughout de-
cidedly the most efficient finger.1
In the A and CA movements subject B, who had the best
coordination, shows an index similar to his index for simpler
movements. Subject A has the same for CA but as shown
above his coordinated wth the A movement on the left side
was poor and his A index is therefore much lower than his
other indices. Subject C has the same great difficulty with
both hands and consequently has practically no index.
Subject D has a higher index for the CA movements. From
2 Wells's results are similar to these. He writes: "Again, in neither subject does
the left hand show an improvement relative to the right. In Subject 1 the index of
right-handedness remains practically the same. In Subject II. the right hand may
even improve more than the left." Loc. cit., p. 454. Our subject A with one of the
fingers of his left hand showed an improvement relative to the right-hand finger.
Whipple remarks, loc. cit., p. 143, that "practice affects the left hand no more than the
right; consequently the index of right-handedness is unaffected by repetition of the test."
This generalization is not borne out by all of the subjects in this experiment, nor by all
of Wells's subjects.
1 The subject could give no reason for this. He had never to his recollection exer-
cised this finger more than the others, and believes it must be an inate characteristic.
FINGER MOVEMENTS 467
these results we again see the low correlation of these complex
coordinations with the simpler coordinations of the single
movement as was shown before.
PRACTICE EFFECT
The experiments were not arranged with the idea of
investigating practice and fatigue, but it does not seem amiss
to discover what evidence there is of their effect under the
conditions described. An examination of the tables I., II.
and III. shows as was to be expected from the results of
previous work that the improvement is not a steady one. If
curves were plotted they would reveal the characteristic
fluctuations. As stated, the maximum rate for each series is
in heavy type. See Tables I., II. and III. It will be seen
that it may occur at almost any point of the series, nor is the
Ri maximum necessarily obtained on the day of the R2
maximum or the S maximum on the day of the maximum of
either finger concerned, nor does the A maximum always
occur on the day of the AC maximum. In short there is a
low correlation as regards the days of the maximum results
for the different fingers of the different movements.
Table VII. has been arranged to show the general change
between the first and second half-year's work, and between
the first and second part of the first half year. The figures
in the first and second horizontal columns are the averages for
the first and second half of the series given in Tables I. and II.
The third, fourth, and fifth horizontal columns contain the
result of the second half year's work.
It will be seen that .with the separate tapping of the fingers
of the right hand in the case of two of the subjects, B and C,
there is decided evidence of the effect of practice, not so much
with B in the difference between the first and second half
year as in the difference of the halves of the first half year.
The other two subjects do not show this difference, in fact A
in the first half year shows a falling off. With the left hand
fingers B and C again show the effect of practice, but it
is not so marked as with the right hand. Subject A shows
a practice improvement in the index finger of the left hand,
468
H. S. LANGFELD
and subject Z)'s results again show no evidence of practice
effect. In the alternating movements B shows decided
improvement as does also A, while C, who had great difficulty,
shows a considerable loss. D remains about the same. In
the complete alternation there is no significant changes. The
change in the simultaneous movement follows the change in
the individual movement.
The most evident fact in these results is the wide individual
differences which preclude any general statement. It is
TABLE VII
Subject
Ri
#2
A
CA
5
Li
L*
A
158
168
121
4.1
184
•**•
112
127
J
152
1 60
149
161
133
T
43
168
125
•*/
129
160
161
157
B
169
154
136
41
168
141
14.3
180
176
4s
T^
41
181
TJ
150
*TJ
ISI
185
178
177
181
174
c
157
143
68
27
158
112
lie
187
f
163
59
*5
J
174
117
J
III
185
I6S
193
I9S
I67
D
I4.O
ICQ
106
14.
IC2
I3Q
131
**r
136
0 -7
155
107
«JT
37
0
I48
Js
133
* j *•
133
138
155
141
135
152
interesting to note, however, that practice does affect the
result in some cases even though the daily amount of work of
each finger is slight and there is a week's interval, if not more,
between each series.1 Another fact to be noticed is that the
practice gain in the more complex coordination, as in the case
of the A movement, may be independent of the progress of a
less complicated movement. In the case of subject A both
^i and R2 show a loss in rate in the second half of the
series, J?2Js loss being considerable, and yet the A movement,
1 The difference of practice gain between the two hands appeared in the difference
of indices for right-handedness, p. 466.
FINGER MOVEMENTS
469
which is a coordination of these two, shows a decided gain.
Subject C shows a gain by practice in Ri and R2, and a loss
in the A movement. Finally the fact that the more voluntary
movement, CA^ shows practically no change in rate, should be
emphasized.
The two horizontal columns next to the last column in
Tables I. and III. are arranged to compare the averages of the'
two orders of procedure. The first one of these horizontal
columns for each subject gives the averages when a single
finger movement precedes the double, simultaneous move-
ment, the second when the reverse order is used. In both
series the rate for the R finger for all subjects except C and
one figure for B is more rapid when the R succeeds than
when it precedes the S movement. In the majority of cases
the difference is as large or larger than the m.v. Only in
two instances is the rate for the L appreciably affected
by the order and the results are of opposite nature. The
simultaneous movement in three cases is decidedly more
rapid when it starts the series, in two instances the difference
is greater than the m. v. The only results, then, that seem
at all significant are those that show the R movements more
rapid when they succeed the other movements. The possible
explanation is that with this finger the warming up effect was
greater than the fatigue. It must be remembered that there
was a pause of two minutes between the 3O-second tests
which could very well be sufficient for recovery of this finger
but not for the others. In the wrist-tapping test Wells used
two and a half minute pauses and his results show an increase
in rate as the series progressed.
FATIGUE
The amount of fatigue in each 3O-second series is calculated
by finding the relation of the difference in rate of the first and
second 15 seconds to the entire 30 seconds. These were found
for the results in Tables I. and III. and are given in decimal
form in table VIII. The absolute differences are also given.
For instance, the first figure in column 3, i. <?., .047, was found
by dividing the rate of Ri which is 155 into 7.3 which is the
470
E. S. LANGFELD
difference of the two 15-second halves.1 The averages for all
the subjects are given in column 10.
Most striking is the fact that fatigue is greatest for the A
movement and that there is no fatigue in the CA movement. It
must be remembered that the A movement is the rapid antag-
onistic movement, the CA movement is very slow and one set of
muscles rests while the other reacts. This readily explains
TABLE VIII
Subject A
Subject B
Subject C
Subject D
AT.
Ri .
n
-%
7.5
4
5-5
2-5
5
I
.047
.041
.071
.042
.025
.044
.017
.031
•039
•039
6
4-5
— 1/18
5
5
4
5-5
B
5
5
•034
.028
•OS
.028
.027
.027
•034
.028
•033
.031
12
8
-4/5
5-5
ii
5
4-5
7
2
6
.071
•053
.07
•033
•059
.043
.036
.042
.017
.046
6
9
6
-i/3
6
6
6
10
6
3-5
.044
•059
.056
.04
•043
•053
.046
.065
•045
.026
.049
.045
.062
.036
•039
.042
•033
.041
•033
•035
R^
A
CA
S
Ri..
Li. . .
S
R2
L^
S
the absence of fatigue in 30 seconds. This is not only true
for the averages but with few exceptions for all the subjects.
For subject C the A movement fatigue is about the same as the Ri
fatigue. The S movement for Ri and R2 for two subjects is the
least fatigueable, for the other two it is the same as the lowest
index of the single finger. In the S movement for Ri and
Li and R2 and Z,2 the index is the lowest in three instances
and only twice is it higher than both of the single movement
indices.
It may be said, therefore, that in general the simultaneous
movement of two fingers for 30 seconds does not show more,
and very often less fatigue than one finger. A comparison
of Ri and R2 in the two halves of the table shows as other
experiments have before, that practice has the tendency to
reduce the fatigue. Only in one instance out of eight is it
1 In view of the manner in which the results were recorded, it was thought better
to divide the periods in halves rather than to compare the first five seconds with the
averages of the ist, 2d, 3d, 4th, 5th and 6th five second periods as Wells did. Loc. cit.,
p. 469. His index is higher probably because the initial spurt has thus more influence
on the result, it being reduced in the above index by the results of the 2d and 3d five-
second periods.
FINGER MOVEMENTS 47 *
greater with practice. As has frequently been found the
decrease of fatigue is the important factor in practice gain.
No correlation can be found between the fatigability of the
two hands. Nor can a general statement be made as to the
fatigue index of the right as compared to the left hand. In
three subjects the relation between R2 and Li and Ri and Li
in regard to fatigue is in the same direction, except that subject
B shows the same fatigue for Ri and Li while R2 is less than
Z,2, but subject D shows opposite results. For him the fatigue
for Li is greater than for Ri and for L.2 less than R2. One can-
not say, therefore, that if the Ri finger is more easily fatigued
than the Li finger, that the R2 finger will be more easily
fatigued than the L2 finger.
VARIATIONS.
There is less variation in the left-hand movements than
in those of the right hand. This is what both Bryan1 and
Wells2 found for wrist movements. The small m. v.'s ac-
company the slower reactions but there are even indications
of a less relative variation on the left side. The simultan-
eous movement shows about the same variations as the single
movements. The CA movement shows the least variation
of all the movements. This movement being very slow
(fewer taps per 30") the relative variability is higher than with
the other movements. No general statement can be made in
regard to the ^-movement. There is a tendency for it to
be relatively larger than that of the single movement. Ab-
solutely it is sometimes larger and sometimes smaller.
SUMMARY AND CONCLUSION
If the index and second finger of the right hand are tapped
simultaneously as rapidly as possible the resulting rate
according to the results of four subjects of varying degrees of
motor ability, is faster than the rate of the slower finger and
may even be more rapid than the faster finger when tapping
singly. There is doubtless a more or less innate coordination
1 Loc. cit.y p. 163.
2 Loc. cit., p. 480.
472 H. S. LANGFELD
between the movements of the fingers of the same hand caused
by the biologically important grasping reflex. When the
extensor of the index finger is innervated a tendency for the
symmetrical extensor of the next finger to move is also ob-
served.1 Ths impulse must be inhibited and it is probable that
this inhibition also extended to the motor half-center of the
first finger causing a loss in rate of movement. An inhibition
somewhat similar in nature has been demonstrated by Sher-
rington in his experiments on the stepping reflex when he
simultaneously stimulated two afferents which are antago-
nistic in their effect. He says: "Of the two afferents concur-
rently stimulated, that one which when stimulated alone causes
flexion of the joint excites the flexor half-center and inhibits
the extensor half-center; and the other afferent, which when
stimulated alone causes extension of the joint, excites the
extensor half-center and inhibits that of the flexor. When
both afferents are stimulated simultaneously with appro-
priate intensity, the discharge from the flexor half-center
represents the algebraic sum of the opposed excitation and
inhibition which the two afferents individually exert on it,
etc."2 In our experiment it is the inhibition of one flexor
half-center which is communicated to the other flexor half-
center and this inhibition is then compounded algebraically
to the excitation of the extensor half-center. When both
fingers move simultaneously this inhibition is removed.
This at least seems to be a plausible explanation of the fact
that the simultaneous movements were faster than either
single movement. Whether it was the only factor could
readily be tested by tapping simultaneously symmetrical
fingers on two hands. It was found that under these con-
ditions the movement of the slower finger was facilitated,
the degree varying with different individuals. The two fingers
together, however, never tapped faster than the faster finger.
1 L. Huismans states "Homolaterale M-B. (Mit-Bewegungen) sind in den weitaus
meisten Fallen auf eine Irradiation des Bewegungsimpulses in der Hirnrinde Zuruck-
zufuhren" ('Uber Mitbewegungen,' Deutsche Zeitschrift fur Nervenheilkunde, 40, 1910,
P- 233)-
2 'Reflex Inhibition as a Factor in the Coordination of Movement and Postures,'
Quarterly Journal of Experimental Physiology, 6, 1913, page 269.
FINGER MOVEMENTS 473
Numerous examples are to be found in the literature relative
to the influence of the movement of one side of the body upon
that of the other. In the simple reaction experiment Paul
Salow found that the reaction time for simultaneous move-
ments of two hands was longer for each hand respectively
than when they reacted separately.1 M. L. Patrizi found in
his ergograph tests that in simultaneous action both hands
did less, but when they worked alternately the right hand
action reinforced that of the left hand. He believes that
one cannot give maximal attention to the two simultaneous
acts.2 W. W. Davis found that in general the right hand
tapped more rapidly alone than in connection with either the
other members. Two of his subjects were able to tap more
rapidly when all four members were tapping. He remarks
that "with a longer practice the right hand, in multiple
tapping, would undoubtedy excel in rapidity its record while
tapping alone." He also concludes from his results that
"there is a close connection between different parts of the
muscular system through nervous means. This connection
is closer between parts related in function or position." In
this work Davis was not interested in the influence of the
faster on the slower member.3 Mention should also be made
of the fact that a paralyzed member may be moved by moving
a healthy member.4
1 'Untersuchungen zur uni- und bilateralen Reaktion. II. Einige Versuche am
Chronographen,' Psychol. Stud., 8, 1913, pp. 506-540).
2 Patrizi writes: "Les recherches que j'ai rapportees dans ce memoire nous font
admettre une uncompatibilite d'etats psychiques, meme quand il s'agit de la coincidence,
dans le meme instant, d'impulsions maximales destinees a des mouvements symetriques
et homogenes et qui sont habituellement accouples." In regard to Fere's results,
which indicate that when the left hand is almost fatigued its capacity can be increased
by the movement of the right hand, he says that an increase will not occur if the two
movements are exactly simultaneous ('La simultaneite et la succession des impulsions
volontaires symetriques,' Arch. Italiennes des Biol., 19, 1893, p. 138).
3 'Researches in Cross-Education,' Yale Studies, 6, 1900, pp. 6-50.
4 See J. Grasset, 'L'action motrice bilaterale de chaque hemisphere cerebral/
L'annee Psychol., II., 1904, pp. 434-445. It is interesting to note that W. P. Lombard
said some years ago: "Not enough work has been done to admit definite statements
concerning the strengthening or weakening effect of the action of the one hand upon the
other. . . . The few observations which have been made with reference to this question
favor the idea that if one hand acts simultaneously with the other it tends to weaken
rather than strengthen its movements. This effect is not a constant one, however, as
474 H. S. LANGFELD
In searching for an explanation for this contralateral
facilitation of the slower movements by the faster, the fact
of sympathetic movement seems the most significant. Every-
one has had the experience when the member which one
desires to move is held or when it has become fatigued, of
moving the opposite member. In this regard Ch. Fere re-
marks that the examples drawn from his research "indiquent
que, lorsqu'il existe un obstacle a un mouvement volontaire
unilateral, Pinflux nerveux a une grande tendance a prendre
la voie symetrique du cote oppose." This tendency is greater
in children, being later more or less suppressed.1 In the
simultaneous movement of the two fingers the faster finger
must be held back to synchronize with the slower. According
to the above if the one finger were not already moving while
the other finger was being held back from its full movement
the sympathetic movement of the former would probably be
evident. When it is moving at the same time as the finger
which is being somewhat retarded, its movement is facili-
tated by the surplus energy of the faster finger. There is also
another explanation or, perhaps, a second factor in conjunc-
tion with the above and that is the increased amount of
peripheral stimulation, i. e.y the contact with the board and
bar. We know from the work of Sherrington, Alexander
Forbes, T. Graham Brown, and others that the afferent
impulse on one side of the body may cause a contralateral
reflex. Now the afferent path on the ipsolateral side may
be fatigued and the stimulus on the contralateral side may
become more effective. Alexander Forbes says: "The fact
that central fatigue induced through one afferent nerve
usually does not impair the reflex involving the same muscles
induced through another afferent nerve supports the con-
clusions of Lee and Everingham that this fatigue does not
involve the moto-neurones, and accords with the view of
Sherrington that its seat is the synapse."2 This explanation
many exceptions occur" ('Alterations in the Strength which occur During Fatiguing
Voluntary Muscular Work/ Jour, of PhysioL, 14, 1893, p. 116).
1 L'alternance de 1'activite des deux hemispheres cerebraux,' UAnnee Psychol., 8,
1901, p. 148.
1 'The Place of Incidence of Reflex Fatigue,' American Journal of Physiology, 31,
1912-13, p. 122.
FINGER MOVEMENTS 475
can also be applied to simultaneous tapping on the same hand.
Owing to the fact, however, that these movements were never
carried on sufficiently long for any great fatigue this factor
must play, if at all, a very small role.
Thus far we have been discussing simultaneous movements
of symmetrical members. This facilitation was also observed
between asymmetrical members but it was not so great.
There can be little doubt that the coordination is less perfect
the more asymmetrical the members are. Even the transfer
of practice is greater to symmetrical parts.1 There is not,
however, a very high degree of correlation between increasing
symmetry and increasing facilitation with all the subjects.
In order to explain the results an antagonistic factor has been
suggested, i. e., that, ceteris paribus, the greater the difference
within limits between the rates of the two fingers the greater
the facilitation of the slower. This difference increases with
the asymmetry and at times causes an asymmetrical pair
to show greater facilitation than a more symmetrical pair.
An explanation which includes two opposing factors can ex-
plain anything and should not be used without strong proof.
There are, in this series of experiments, instances where the
degree of symmetry is a constant and we see here the effect
of difference in rate in the direction just mentioned. We
have also cases of similar degrees of difference in rate and
here the effect of symmetry can be seen. The results are
not sufficiently consistent to be conclusive. They are, how-
ever, suggestive and the explanation founded upon them
must necessarily also bear that adjective.
2. An examination was also made of the simultaneous
movement of the index and first finger of the right and left
hand when these movements were in opposite directions, a
combination of movements similar to the stepping reflex and
spoken of in this paper as alternating movements. Here
there is double reciprocal innovation according to Sherrington.
The four subjects were divided into two groups in respect to
the amount of coordination. With three of the subjects each
finger was able to do about three quarters as much in this
1 See W. W. Davis, loc. cit., page 49.
476 H. S. LANGFELD
combination as when tapping alone, except for one of the
subjects with the left hand. The loss for each finger due to
this combination of impulses was only 25 per cent., so that
in the same time the pair was able to do one and a half times
as much as a single finger. The fourth subject showed a
much poorer state of coordination. Two of the three subjects
with good coordination were piano players and the other who
had a low efficiency with the left hand had his fingers of that
hand trained for another system of coordination, for violin
practice. Judging, however, from the great differences be-
tween these subjects and from observations on others, it
seems probable that this test shows some fundamental dif-
ferences in motor coordination. It would appear profitable
to gather further data with special reference to the effect of
practice, the tests here described being too few to draw any
conclusion from them on this point.1
Another fact which appears clearly in the results is that
the amount of coordination between the two reflexes does not
run parallel with the development of these reflexes. The
subject who has the worst coordination has one of the highest
1 Sherrington, in speaking of grace in walking, says that "the proper execution of
the act ensures a moment of complete rest to each of the opposed motor centers en-
gaged." Loc. cit., p. 271. T. Graham Brown found very decided differences in the
reflexes of the cat "... these individual variations are probably due in great part
to more fundamental differences in the constitution of the nervous centers. Some cats
are 'walkers.' They exhibit in a marked degree the phenomena of 'narcosis progres-
sion.' Other cats are 'scratchers.' In them the scratch-reflex is peculiarly excitable"
('Studies in the Physiology of the Nervous System, XIV. Immediate and Suc-
cessive Effects of Compound Stimulation in Spinal Preparations,' Quarterly Journal of
Experimental Physiology, 7, 1914, p. 200).
FINGER MOVEMENTS 477
rates in single tapping. He is a subject with little piano
practice, so that this ability in single tapping substantiates
Raif's assumptipn (vide supra, p. 464) that piano players do
not show any special ability in rate of tapping.
3. The term complete alternation has been applied here
to successive movements of the two fingers; one finger com-
pletes both the extensor and flexor movements before the
other begins. With all the subjects and with both right and
left hand the two fingers in this combination tap only one
quarter the amount that they do in the same time in tapping
continuously and alone. The loss in this form of alternating
movement is about 50 per cent. This combination is a system
of successive reflexes. The loss in efficiency is probably in
great part due to the necessary inhibition of the previous
reflex. As Sherrington says ". . . there will persist during
the new reflex activities belonging to the old with, in result,
confusion of the two. Rarely, indeed, can it happen normally
that the reflex machinery in executing a train of different
reflexes is actuated by a train of different stimuli, each one
of which abruptly ceases just as the next one begins."1 And
further: "For orderly and unconfused sequence of reflex
acts — also of willed acts — central inhibition is a necessary
element of coordination in the transition from one muscular
act to another."2 Another reason for the loss is that in the
single tapping movement inhibition of one of the antagonistic
movements increases the tendency for that reflex to discharge,
causing what might be termed a rebound. If, however, the
finger must rest after the completion of each flexor movement
until the other finger has completed its movement, most of
this post-inhibitory effect is lost.3
4. Even with an interval of one week between the tests a
practice gain in most of the movements is noticeable. An
exception must be made with the most voluntary movement,
namely, the complete alternation. The subject who had the
worst coordination in the alternating movement even showed
a loss. The practice effect in the alternating movement does
1 Loc. cit., p. 275.
2 Loc. cit., p. 276.
3 See Sherrington, loc. cit., p. 278.
478 H. S. LANGFELD
not run parallel to the practice effect of the fingers tapping
separately. A gain in the former may be accompanied by a
loss in the latter. Practice in the simultaneous movement,
on the other hand, does follow that of each finger when working
separately. There is some evidence that practice affects the
right hand more than the left.
5. Fatigue is noticeable for all the movements except the
complete alternation.1 It is greatest for the alternating
movement and generally least for the simultaneous move-
ment. That is, during thirty seconds there is less fatigue when
two fingers are working simultaneously than when one is
working alone.
6. The index of righthandedness is not necessarily the
same for all the movements, nor is it always the same between
the different pairs of symmetrical fingers.
7. The variations in tapping rate are less for the left hand.
1 Wells says: "To sum up, the maximum rate of repeated voluntary movements is a
function that practically every investigator working with sufficiently accurate methods
has found to be subject to fatigue effects, though the degree of this subjection has
differed considerably" ('A Neglected Measure of Fatigue,' Am. Jour, of PsychoL, 19,
1908, pp. 3S2-3-)
RETINAL FACTORS IN VISUAL AFTER-MOVEMENT
BY WALTER S. HUNTER
The University of Texas
The present paper is a continuation of a previous study
made by myself on the after-effects of visual motion.1 On
pages 255-257 of that article, comments are to be found
bearing upon retinal factors effective in the production of the
illusory motion which occurs after gazing for some time at a
series of moving bands. Again on pages 275 and 276 an
experiment is described in which a moving area was used of a
size sufficient to cover most or all of the visual field. In this
test an after-movement was observed which went in the
same direction as the real or stimulus movement. Ordinarily
the after-effects seen are secured with small stimulus areas
and move in a direction contrary to the real movement. In
the earlier study, it was held that the normal after-effect was
a result of some or all of the following factors: eye-muscle
strain, association factors and retinal processes which were
probably fading after-images. The present work is primarily
concerned with a determination of the nature of the effective
retinal factors. The conclusion reached is that this factor
is a streaming phenomenon which moves through external
space in a direction opposite to that taken by the stimulus
area. The evidence for this is necessarily of an introspective
and theoretical nature.
The data presented were obtained largely from three
subjects, two of whom had served in the earlier tests. A
number of observations were made on untrained subjects.
The apparatus used was the large striped curtain (six feet
long by four feet high) described in the earlier paper. This
curtain could be made to move either up or down. A small
square of white paper was held close in front of the curtain
by a thread and served as a fixation point. A number of
1 Hunter, Walter S. 'The After-effect of Visual Motion,' PSYCH. REV., Vol. 21,
pp. 245-277, 1914.
479
480 W. S. HUNTER
newspapers were tacked on the wall above the curtain to
serve as a projection field for the after-effects. Other fields
will be mentioned later. Tests were also made with rotating
black spirals and with a striped Scripture drum.
EXPERIMENTAL DATA
The immediate problem from which the present tests
took rise was that of the occurrence of an after-movement
(abbreviated as a.-m.) in the same direction (an s.-a.-m.)
as the stimulus movement. It will probably conduce to
clarity, if the observed facts are grouped accordingly as this
type of after-movement is or is not seen. The facts concern-
ing streaming can then be presented.
Observations on S.-a.-m. — If an observer is seated about
one meter from a moving curtain of the type used here and
fixates a point in front of the curtain for twenty seconds,
upon turning his eyes to a series of newspapers tacked upon the
wall above the curtain, he will usually see the following phe-
nomenon: the fixation point of the projection field and prob-
ably about one or two square feet of the immediately adjacent
paper will be seen to move in the same direction as the real or
stimulus movement. This movement is most rapid during
the first second and is always a drifting bodily movement of
the projection field. It is identical in quality or kind with
the usual after-movement seen when a small stimulus area
(parallel stripes on the Scripture drum or the Archimedian
spiral) is involved and when the projection field is the stopped
stimulus area. The difference between the two phenomena
is one of direction only.
I have never secured an s.-a.-m. using the stopped curtain
as a projection field. (The same is true for spirals and for
striped fields of small extent.) It is all but impossible to
secure it if the observer sits within eight inches of the moving
curtain even though he projects the after-effect above the
curtain as before. When the observer sits some two and a
half meters from the moving curtain, he can secure the
s.-a.-m.; but in the tests reported here, it has not been so
easy as from the distance of one meter.
VISUAL AFTER-MOVEMENTS 4Sl
The following test has also been made. A white card-
board ii in. X 14 in. with an aperture 4 in. X 7J in. was
placed in front of the moving curtain. An observer seated
at from one to three meters fixated a point on the edge
of the aperture for twenty seconds and then fixated a dot
on the papers above the curtain. A stationary black negative
after-image of the cardboard was seen. In this a.-i., there
was a rushing movement either upward or downward in the
same direction as the real or stimulus movement. Outside
the black a.-i. and in the area corresponding to the aperture
of the card, a movement is usually to be seen which goes in a
direction opposite to the real movement. (This we shall term
an op.-a.-m.) Some observers reported the whole projection
area to be involved in an s.-a.-m. When the after-effect was
projected upon either a black or a white cardboard, no move-
ment was seen in the negative after-image of the cardboard
which had been in front of the curtain.
Observations on Op.-a.-m. — An after-movement in the
opposite direction to that of the real movement is the usual
and "normal" after-effect save under the circumstances just
described. When the moving area of large extent is used and
the after-effect is projected upon newspapers on the wall
above the curtain, an op.-a.-m. is seen particularly below
and to the sides of the fixation point. Simultaneously the
s.-a.-m. described above is seen around the fixation point.
At times the op.-a.-m. appears later than the s.-a.-m. Some
observers see a drifting op.-a.-m. around and over the fixation
point also. In this case the movement seems to be between
the observer and the paper. These central and peripheral
op.-a.-ms. are nearly always described as a film moving over
the projection field and not as a movement of the field itself.1
It is a radically different type of movement from the s.-a.-m.
seen about the fixation point. The op.-a.-m. is a film
which is either described as a series of shadows or as a 'rain-
fall' or 'sleet' or 'dust film.' Frequently this film will drag
the projection field along with it; but even in this case there is
no drifting bodily movement as is seen in the s.-a.-m.
1 See Hunter, op. cit., pp. 247-8 for results of earlier tests.
482 w. S. HUNTER
If the projection field is the stopped curtain, an op.-a.-m.
is seen which is a slow drifting movement of the field itself.
A film has never been seen by my observers under these con-
ditions in the present tests. This is also true when stimulus
areas of small extent are used, whether they are spirals or
systems of parallel lines. In the earlier tests just referred to in
the footnote, such a film was suggested by the observers.
The absence of this in the present tests is probably due to
the fact that the subjects were mainly familiar with intense
films.
Observation on Streaming. — It will be well to preface these
observations with a few historical comments. Our interest
lies chiefly in the observations of Pierce,1 Szily2 and Ferree3
who wrote in that chronological order. The present writer
does not care to examine any claims as to originality. It is
quite probable to his mind that the phenomenon here under
consideration has long been known. Schilder,4 e. g., quotes
from Purkinje's 'Beobachtungen und Versuche' passages on
the streaming phenomenon that parallel Szily's. Ferree
accepts Pierce's explanation as a basis for distinction between
their respective observations.
Pierce's observations were made with a stationary black
and white striped area. The projection field was a plain
black ground of cloth or card. After some twenty seconds'
fixation, upon turning to the projection field, "The appearance
is that of a thin cloud of fine white dust moving across the field
of vision" The direction of moving is always perpendicular
to the striped lines. The same observations were verified
with concentric circles. Again, "if the usual field of fixation
be divided by a vertical strip of some uniform color, no 'drift '
will be seen in that portion of the field corresponding to the
strip." In explanation Pierce says "it seems probable on the
whole that the ultimate explanation of this as of all after-
1 Pierce, A. H., 'Studies in Space Perception,' pp. 331-8, N. Y., 1901.
2 Szily, A. v., 'Bewegungsnachbild und Bewegungskontrast,' Ztsch. f. Psych, u.
Physiol. d. Sinn., 1905, Bd. 38, S. 124.
3 Ferree, C. E., 'The Streaming Phenomenon,' Amer. Jr. Psych., 1908, Vol. 19.
4 Schilder, Paul, 'Uber auto-kinetische Empfindungen,' Arch. f. ges. Psych., 1912,
Bd. 25, S.
FISUAL AFTER-MOVEMENTS 483
images of motion, will be somehow formulated in terms of
impulses to movement aroused by the particular stimulation
that precedes. Perhaps the experiments here recorded may
contribute their mite towards this final explanation, if that
ever comes."
Szily's essential statement is made in connection with his
discussion of the s.-a.-m. and is as follows: "Die hier geschild-
erten Nebenerscheinung begleitet, die dem minder unsich-
tigen Beobachter allerdings erst dann auffallt, wenn sie unter
gewissen Versuchsbedingungen in erhohtem Masse zur Gel-
tung gelant. Ich selbst sehe unter alien umstanden in der
Peripherie das Regelrechte [op.-a.-m.] Bewegungsnachbild,
zumeist in der Hiille eines strahligen Nebels, in entgegengest-
zter Richtung ablaufen. Bedient man sich eines noch
dichteren Streifenmusters, verlangert man die Dauer des
objektiven Eindruckes, setzt man die Beleuchtung des Pro-
jektionsgrundes herab, so verbreitet sich diese Zone des regel-
rechten Bewegungsnachbildes mit gleichzeitig erhohter In-
tensitat immer mehr zentrumwarts, so dass sie selbst dem
Ungeiibten auffallen muss. An den Konturen im Bereich
des direkten Sehens aber, solange sie als wahrgenommen
werden, vollzeiht sich auch dann noch stets die paradoxe
[s.-a.-m.] Scheinbewegung."
From Ferree we may take the following: "When one sits
with lightly closed lids, which must be kept from quivering,
before a bright diffuse light such as that of a partly clouded
sky, and looks deep into the field of vision thus presented,
beyond the background as usually observed, one sees about
the point of regard, after the field of vision has steadied, slowly
moving swirls. These swirls have the appearance of streams
of granules1 moving in broad curves now this way, now that,
seemingly without order, unless a noticeable eye-movement
occurs, or is made voluntarily, when the direction of streaming
changes to that of the eye-movement." Ferree is inclined to
identify the phenomenon with a diffusion of lymph over the
retina. Aside from the above general description, our interest
centers in the drawings which his subjects made. These
1 Italics mine.
484 W. S. HUNTER
represent the qualitative character of the phenomenon and are
identical in everything save direction with those made by my
subjects and represented in Fig. I.
The quality of the streaming observed by the subjects of
the present tests was either of two kinds: (i) 'a shadow-like
succession of clouds' of rather medium velocity; or (2) a
'sleeting,' ' snow-fall' or 'dust streaks' which went usually at
a high velocity. In any case, this streaming is always in a
direction opposite to that of the real or stimulus movement.
Neither type appeared in these tests when the projection
field was the stopped stimulus area. The first type has never
appeared when the projection field is a plain black or white
cardboard. It has been observed in the periphery where the
after-effects have been projected upon the paper above the
large curtain, and all over the visual field when the after-
effect has been projected upon floors and upon plain gray
walls.
The greater theoretical significance attaches to the 'sleet'
film. This will appear as the discussion advances. I have
assumed that the best method of detecting what takes place
in or on the retina is to project the phenomena upon black or
white surfaces (cardboard, shadows or the black of the retinal
field). Under these conditions the phenomena appear un-
modified by variations of the external world.
A 'dust' or 'sleet' film — which is best represented by A
in Fig. I — is always seen if the after-effect of the moving
curtain is projected upon plain surfaces as just described.
In these cases it is pure, i. e., unmixed with the first type.
Pierce's test was carried out with the present subjects. The
stimulus areas were the large striped cloth, the striped
Scripture drum and a black spiral on a white disc. In each
case after a fixation of twenty seconds, the observer turned
his eyes to a black or white ground and described the phe-
nomena seen. Drawings of the after-effects were requested.
When the stimulus area was one of parallel stripes, the after-
effect was a 'dust film' as described by Pierce. It differed
from that secured with a moving stimulus area only in its
fainter intensity and in its less certain direction. By the last
VISUAL AFTER-MOVEMENTS
485
phrase, I mean this: the Pierce dust film was perpendicular
to the lines, but the film could be interpreted as moving either
up or down. With the dust film produced by a moving stimu-
lus area, there was no doubt but that the film moved in a
direction opposite to that of the stimulus area. The drawing
B in Fig. I represents the phenomenon seen on a plain card
after having fixated either a stationary or a rotating spiral for
twenty seconds. The difference just described for parallel
FIG. i. Patterns of Streaming Observed in the Present Tests. A. Seen with either
a moving or stationary area of parallel stripes when the projection field is a plain
surface. B. Seen with either a moving or a stationary spiral when the projection field
is a plain surface.
lines is the only one that holds here also. In each case the
film has appeared before the negative after-image of the
stimulus area.1
Earlier in this paper, tests were described where a white
cardboard with an aperture was placed in front of the large
moving curtain. In order to analyze the nature of the retinal
factors producing the peculiar after-effects already described,
the phenomenon was projected upon a black card. In the
black negative after-image of the white card (where the
s.-a.-m. had been under other conditions), there was no
movement; but in the remainder of the card which cor-
responded to a stimulated retinal area, the dust film was seen
(moving of course in a direction opposed to that of the stimulus
movement). There can be little doubt, then, but that the
1 See Hunter, op. cit., p. 258.
W. S. HUNTER
s.-a.-m. seen in the black after-image when this was projected
upon the newspapers was due to association factors. The
movement of the film produced the impression of a movement
of the objects within the after-image in an opposite direction.
Szily1 describes a similar experiment under the caption 'Kon-
trast im Bewegungsnachbilde.' I am uncertain how he would
explain it. However other cases of 'contrast' are accounted
for on the basis of a higher threshold for the perception of
movement in central than in peripheral vision. This, I
think, in all of the cases described, is a needless hypothesis.
The author has repeated some of Ferree's tests upon
himself and one of his subjects. It has been possible to
confirm the existence of a normal streaming activity in the
eye. This when represented by drawing is exactly similar
to the pictures of dust films found in the present experiments
upon stationary and moving striped areas.
THEORETICAL CONSIDERATIONS
To keep theory close to observed fact, it seems to the
present writer that the choice of explanatory retinal processes
lies between a "streaming phenomenon" and fading after-
images. The latter is the more conventional and therefore
the more apt to be favored. The evidence in its favor,
however, is purely hypothetical and non-observational.
Hence as an hypothesis it may well be faulty. What is
actually seen when the eyes are either closed or turned toward
a uniform field is a dust like film in constant movement. To
pass from Ferree's streaming to Pierce's streaming, it is
necessary to assume: (i) that the after-image effects of the
striped field make the dust film more vivid; and (2) that in
some manner the striped field gives definite direction to the
streaming. I can offer no solution for the last statement,
although it is by no means absurd and impossible. Con-
cerning the first assumption, there can be no difficulty for
experiment shows that the film can be more readily seen upon
certain backgrounds than upon others. This will also account
for the fact that the after-effect is largely confined to (i. e.,
1 Szily, op. cit., S. 126.
VISUAL AFTER-MOVEMENTS
487
visible on) an area corresponding to that stimulated by the
real movement.
Let us examine each proposed retinal factor with the aid
of Fig. 2. Of the phenomena described above, it is necessary
to bear in mind particularly the s.-a.-m. about the fixation
point and the dust-like film which is seen to pass over the
projection field in a direction opposite to that taken by the
stimulus area. Our discussion will be further aided by a
quotation from Wundt giving the conventional statement of
the after-image theory. "Indem ein schwaches Nachbild
der gesehenen Bewegung im Auge zuruckbleibt, scheint ein
fixiertes Objekt infolge der Relativitat der Bewegungsvor-
stellung in entgegengesetzten Sinne bewegt zu sein. Das
Nachbild, in der Regel zu Schwach um selbst gesehen zu
werden, geniigt doch um auf das Objekt die zu seiner eigenen
entgegengesetzte Richtung zu iibertragen."1
In Fig. 2, 0 represents the curtain moving up. i. is its
FIG. 2. Relations between External Movements and Retinal Processes.
image on the retina where the movement passes down,
a.-i. stands for the fading after-images which move in the
same direction as i. and not up with the curtain, 0. Let us
assume that twenty seconds have passed and the striped
curtain or drum is stopped. Upon whatever objects the eye
is turned the stationary images of these objects will be cast
upon the area i. over which the fading after-images (a.-i.)
are passing downward. We have two possibilities now: (i)
The fading after-images do not enter consciousness. In this
case a.-i. passing down over i. gives the impression of i.
moving up, i. e., is the same as though a.-i. were stationary
1 'Physiol. Psych./ Bd. 2, S. 622, 6. Aufl., Leipzig, 1910.
488 W. S. HUNTER
and i. were moving up. i. projected, then, into the external
world would be seen as 0 in movement downward. (Here I
assume that a sub-liminal retinal process is not projected into
space. If it were it would be discussed as though the a.-i.
came into consciousness and hence would.be treated under
(2) below.) The present case can explain an s.-a.-m. about
the fixation point by claiming that for some reason the move-
ment of a.-i. on the retina is, relatively, non-effective in central
vision. Szily does this by appealing to the high central
threshold for the perception of movement. The truth in
this point lies only in the observable fact that the direction
of apparent movement differs in the periphery and in the
center. The explanation of this lies in the fact that the clear
bold contours in central vision dominate over the moving
process as compared with the peripheral contours. When the
after-effect is projected upon plain cardboard the dust film
is seen all over the field. The effectiveness of contours in
inhibiting the perception of after-movements is shown by
Szily,1 and is continually verified in tests on after-movements.
The validity of that investigator's threshold hypothesis is
further impugned in the following case. (2) // the fading
after-images do enter consciousness, they do so as projected
into external space, a.-i. is then seen as F' moving up. 0
seen through this appears to move down. The damning
fact, however, is that the film or 'sleet' is actually seen not
as F1 but as F and moves down. It cannot therefore be the
projection into space of fading after-images. The retinal
equivalent of the film which is seen to move down over the
external objects must be the passage of a stimulation upward.
However before ruling fading after-images out entirely, it is
well to consider the following possibility: May it not be that
a.-i. passing down over i. is the equivalent of a wavering or
filmy passing of i. upwards on the retina? This when pro-
jected would give F. Such an hypothesis appears fairly
plausible when F is projected upon print or other equally
diversified objects. In that case a cloudy or wavy film is
often seen to pass down. The plausibility of the theory,
1 Op. dt., S. 1 10.
VISUAL AFTER-MOVEMENTS 489
however, vanishes when one bears in mind the cases where a
dust film is seen even when the after-effects are projected
upon the printed sheets and particularly upon dark shadows
or cardboards. This film has been sufficiently described al-
ready. It is exactly like fine particles of dust moving rap-
idly through space. One would not expect fading after-images
to be of this nature.
A final objection to the fading after-image theory is as
follows: The continued passage of a series of stimuli across
the retina fatigues the retinal elements. If the elements can
recover between stimulations, a movement will be seen.
This is what actually occurs. Now after the stimulation
has ceased, it is conceivable that one wave of recovery
(fading a.-i.) would sweep across the retinal area. I see no
reason why successive waves should do so. And yet this
would be necessary in order to secure such temporally ex-
tended after-effects as are actually observed.
In the light of what has gone before in this paper, we are
therefore led to interpret visual after-movement in terms (so
far as the retina is concerned) of a streaming phenomenon
which passes across the retina in a direction opposite to the
image of the objective movement. I confess my ignorance of
what this streaming may be. It may be lymph currents, as
Ferree supposes. It may be an electrical phenomenon.
The uncertain status of its exact nature cannot, however,
overcome the necessity of its postulation.
The theory readily explains the phenomena described:
The dust films are the projections into space of the streaming.
The s.-a.-m. is due to the invisibility of the film because of
the clear bold contours of central vision as compared with
peripheral vision. This results in the interpretation of the
external objects as moving in a direction opposite to the film
and occurs under special circumstances only. In the regular
after-movement with the large areas, the film is itself visible
and drags the external objects along with it. With small
areas the same thing occurs with a minimum of film visibility.
EXPERIMENTAL DATA ON ERRORS OF JUDGMENT
IN THE ESTIMATION OF THE NUMBER OF
OBJECTS IN MODERATELY LARGE SAMPLES,
WITH SPECIAL REFERENCE TO PERSONAL
EQUATION
BY J. ARTHUR HARRIS
Carnegie Institution of Washington
I. INTRODUCTORY REMARKS
In attempting to estimate the number of a considerable
group of objects of the same kind, the observer can seldom
state the true number but generally gives one which is either
too high or too low. If a series of such estimates by the same
observer be considered it may be found that there is no ten-
dency for the errors in excess of the true value to be more
numerous or greater in amount than those in defect. In
such a case the average of the deviations of the estimates from
the true number of objects will be sensibly zero.1 The
individual making the estimates may then be said to have no
personal equation. Other individuals, however, may have a
definite tendency to err on one side of verity in their evalu-
ations. Such may be said to have a positive or a negative
personal equation, as the case may be.
Personal equation is not the only factor which should be
taken into account in determining the rank of an individual
among a number passing judgment upon the value of any
magnitude. An observer with no consistent bias towards
over or under valuation may be characterized by very erratic
judgment — assigning sometimes a value far in excess, at other
times a value far in defect of the actual. Thus consistency
or steadiness of judgment is also a characteristic of impor-
tance which should be taken into account in the comparison
of individuals.
In this paper, I have presented several series of experi-
1 The mean actually determined by experiment would be o plus or minus a smal
amount due to the errors of sampling.
49°
ERRORS OF JUDGMENT 49 l
mental data bearing on these questions. Such materials,
properly analyzed and interpreted, should be of value to the
psychologist. The data, which are a by-product of long
routine processes in experimental breeding, are presented
solely for their intrinsic value as experimentally determined
facts. The arrangement of the material is that suggested by
the view point of a quantitative biologist, a biometrician.
Comparison, criticism and interpretation are left to those hav-
ing the necessary training. I trust, therefore, that the pro-
fessional psychologist will overlook crudities in terminology,
and accept the experimental data for what they may be worth
when interpreted from his own point of view.
The circumstances leading to the collection of these data
were the following:
At various times, I have found it necessary to obtain very
large series of countings of bean seeds, either for germination
tests or for determining the mean weights for different series.
Various considerations (which need not be reviewed here)
led to the conclusion that the counting could most easily,
accurately and advantageously be carried out in units of 25,
50 and 100 seeds. By the slightest addition to our labor we
could determine the accuracy of judgment in the estimation
of these lots of 25, 50, 100 or 200 seeds. The advantages of
so doing were threefold: (a) It gave a means of testing the
personal equation and the steadiness of judgment of the
three assistants who are responsible for a large part of the
routine work in my laboratory, (b) Competition (for first
place in accuracy of estimation) added a little spice to a long
task which would otherwise have been the most monotonous
drudgery, (c) It gave an extensive series of data on errors of
judgment.
II. EXPERIMENTAL METHODS
The method of the experiment was as simple as can be
imagined.
The observer took from a container a handful of beans and
poured as nearly as possible a specified number of seeds (25,
50, 100 or 200 according to the experiment) on the table.1
1 The seeds were poured upon a dark gray felt paper mat. This was chosen
492 ]- A. HARRIS
If there appeared to be too few, more were added, if too many,
a portion were put back. The work was carried out so rapidly
that there was no possibility of counting assisting in the
estimate.1 The number was then at once determined by
counting and the deviation of the sample from the desired
value was noted and recorded by the observer who made the
estimate. A persistent effort was made by each observer to
improve in succeeding estimates. The influence of this con-
stant checking upon personal equation and upon steadiness of
judgment seems to me the most interesting phase of the
present study. It will be considered in a subsequent paper.
Except in one special case fifty estimates were made by
each observer in the morning and another fifty in the after-
noon. In general, this was attended to early in each half day's
work. Each lot of fifty may be designated as a period. The
periods were, with a single exception, consecutive except for
a break over Sunday. Ample details concerning the indiv-
idual experiments are given below. Four individuals took
part in the work.2
The following series of experiments were made:
A. Our first series of experiments was made in May, 1912.
The attempt was in every instance to lay out a sample of fifty
seeds. Observer A made only 141 estimates and these with
because the color selected was easy for the eye although affording sufficient contrast
with the (generally) white seeds and because the seeds do not roll about as badly on
the felt surface. These conditions count for rapidity and comfort of work. Each of
the observers occupied her own table, so that light and other conditions were perfectly
familiar through long experience.
1 The validity of this statement will be apparent from the fact that in the first
experiment the average time required for pouring out the samples of fifty seeds, counting
them twice, recording the deviation of the guess from the true number, and replacing
slightly wrinkled or weevil eaten seeds by perfect ones was 86 minutes for Observer B,
83 minutes for Observer C and 80 minutes for Observer D.
2 My own observations were too few and too continually broken into by extra-
neous matters to be of particular value. The other three are Miss Edna K. Lockwood,
Miss Margaret G. Gavin and Miss Lily J. Gavin. Each of them has been with me for
six years or more. It would be difficult to express adequately my obligation to them
for their patient, conscientious and highly efficient assistance in the onerous routine,
observational, clerical and arithmetical work of a biometric laboratory. For conven-
ience the observers are designated hereafter by letters: Mr. Harris, Observer A or
merely A\ Miss Gavin, Observer C; Miss Lockwood, Observer B; Miss Lily Gavin,
Observer D.
ERRORS OF JUDGMENT 493
numerous interruptions. Observers B-D made each 700
attempts at laying out the required number. These are
grouped in 14 periods of 50 trials each. In the case of B and
C these were made in the mornings and afternoons of seven
days which were consecutive except for Sunday. Observer
D worked on the same schedule but was necessarily absent one
afternoon, hence the 14 periods were distributed over 8 days.
The data of this series will be designated as \A^ IB, 1C, and
IZ), the Roman numeral referring to the series and the letter
to the observer.
B. The second series of experiments was made in Novem-
ber, 1912. Again the White Navy bean seeds were used and
the attempt was to lay out samples of fifty seeds each. The
work covered a period of two weeks, in which there was a
morning and an afternoon period of 50 estimates each. Thus
there were altogether 1,200 estimates by each of three ob-
servers, By C, D. The estimates logically fall into two major
periods of six days each, separated by Sunday. They are
therefore designated as series II. and III. The appended
letters designate the subjects, B-D.
C. The experiments were again taken up with the White
Navy beans in February, 1913. Again two weeks were devoted
to the work and 1,200 trials, in daily morning and afternoon
periods of 50 each, were made by the three individuals B-D.
This time 100 instead of 50 seeds was the number aimed at
in the laying out of the samples. The first of the two weeks
may be designated by IV., the second by V.
D. The last four days in July, 1913, trials at estimating
200 seeds were made by B, C, and D. Four days' work of
50 trials per day completed the countings that were necessary
for the masses of Navy seeds then to be weighed. Because
of the time required for the counting and recounting of samples
of 200 seeds it was not feasible to do more than 50 estimates
in a day. These were made consecutively and usually
required a full half day's concentrated work. Doubling the
number would have meant an abnormal mental and physical
effort for those making the estimates. The experiment is
designated as VI.
494 /• A. HARRIS
E. The Tuesday following the preceding set of estimates,
(Aug. 5) which were closed on Thursday, a set of trials at
laying out 25 beans, of a larger-seeded brown bean — "Ne Plus
Ultra" — was begun. Sets of 50 estimates in the morning
and afternoon were made daily except for Saturday afternoon
(Aug. 9). Thus there were 9 periods each by C and D. These
will be referred to as VII.
F. The week immediately following the preceding trials
(Aug. 11-16) Observer D did one full week (12 periods of 50
estimates each) on the "Ne Plus Ultra" seeds; again the
attempt was to lay out 25 seeds. Series VIII.
G. For the two weeks beginning August 25, 1913, Ob-
servers B and C made trials at the estimation of lots of 25
seeds, using a White bean somewhat smaller than the Navy
on which the main bulk of these experiments was based.
The first week embraced n periods, Saturday afternoon being
out. The second week included only 9 periods, Sunday and
Monday (Labor Day) separated the two lots. The estimates
were closed with Saturday morning of the second week,
when the supply of seeds which required weighing was ex-
hausted. Note that Observer B had made no estimates since
the end of July when she was estimating at lots of 200 seeds.
Observer C had made no estimates since August 9, when she
was working with large seeded brown beans, and estimating
at fifties. The first week forms series IX. and the second X.
H. During the period of May 4 to May 9, 1914,
inclusive, Observers B-D made estimates twice daily at 50
seeds of a large brown bean, Burpee's Stringless. Thus these
estimates were made after a lapse of several months (August,
1913— May, 1914) since the last trials. For convenience these
will be designated as Experiment XL
Thus there are altogether 28 sets of experiments, carried
out by three observers, distributed over a period of two years,
and comprising a total of 15,200 estimates.
The analysis of the data is carried out by the modern
higher statistics, the notation of which is very generally
familiar or easily accessible.
ERRORS OF JUDGMENT
495
III. PRESENTATION AND ANALYSIS OF DATA
I. Personal Equation
Data Tables A-E give the deviations of the number of
seeds actually laid out from the desired number (25, 50,
100 or 200), i. e., seeds actually drawn less seeds intended to
be drawn. Note that the attempt was in each case to lay
out a definite number of seeds, say 50. +15 indicates,
-IS -10 -f 0 -f^ 4/0 4 Af 42 O
EXPERIMENT I D
-/O -S 0 4-5" +/<?
EXPERIMENT I B
DIAGRAMS 1-2. Distribution of Errors of Estimation in Experiment I. The
heights of the ordinates indicate the frequency of deviations of different grades.
therefore, that 65 instead of 50, and —6 indicates that 44
instead of 50 were actually drawn.
Certain of these series are also represented graphically in
Diagrams 1-2.
Two distinct problems are presented by these distributions
and graphs. The first is that of personal equation properly
so called; the second is that of steadiness of judgment.
/• A. HARRIS
By personal equation, we understand a bias in a given
direction — a tendency to estimate too high or too low. If
there be a personal equation, one observer will tend regularly
to pour out too many seeds, just as another will tend to make
the sample too small. By steadiness of judgment, we mean
consistency in estimation. One observer may be more erratic
than another, estimating now far too high, now far too
low.
These points may to some extent be illustrated by the
first two diagrams. We note that for Observer B the fre-
quencies (represented by the heights of the bars) of the grades
above and below o are about equal, while in the case of the
Observer D the frequencies above o are distinctly greater than
those below . Indeed in the cases of Z), there are six groups of
errors of observation above o which contain more cases than
any class below o. Apparently Observer B has little personal
equation, while D has a pronounced tendency to lay out too
many seeds — that is to underestimate the number of objects
in a group. The diagrams also show somewhat the relative
steadiness of judgment of the two observers. The deviations
appear to be less widely scattered about the mean in the case
of B than in that of D — judgment is apparently steadier, less
erratic.
Numerically the existence of personal equation may be
most simply tested for by determining the relative numbers
of estimates in excess and in defect of the true value. For
convenience the tables have been broken up into three com-
partments. At the head the frequency of cases in which the
error was o (i. e., in which the experimenter actually succeeded
in laying out the number of seeds desired) is indicated. The
frequencies of + and — deviations of various magnitudes are
shown side by side in the two parallel columns. The totals
of these columns give the data needed in answering in the
most rough and ready manner the question as to the existence
of a personal equation.
Of the 28 experiments made, the totals of the tables show
that in only 3 cases is the frequency of minus deviations greater
than that of plus deviations. Or in other words, in 25 cases
ERRORS OF JUDGMENT
497
out of 28 the experimenters in the long run made the error of
laying out too many seeds.1
In comparing the frequencies of + and — deviations of the
same magnitude (which have been placed in the tables in
parallel columns to facilitate such comparison) it is also clear
that in almost every grade of deviation represented by a
material frequency the results in excess of the attempted
number are more numerous than these in defect.
With results indicating in so striking a manner the exist-
ence of a pronounced personal equation on the part of the
observers, the calculation of any probable error seems super-
fluous, especially since probable errors are given for a subse-
quent test.
A measure of the magnitude of personal equation as well as
a demonstration of its existence and direction must be sought.
This is most simply expressed in terms of the mean deviation
of the estimates from the desired value.
Since in our experiments the attempt was made to lay out
a sample of a given size, an excess of plus deviations either
TABLE I
PERSONAL EQUATION FOR INDIVIDUAL EXPERIMENTS
Experi-
ment
Trials
Number
Sought
Observer B
*.
Observer C
««
Observer D
*,
I.
700
50
+ .I7ldr.I07
+ 1 .60
+ .926dr.!O6
+ 8.74
+2.I70dr.I33
+16.32
II.
III.
600
600
SO
SO
+ .5IOdr.098
+.5iodr.o8o
+ 5-20
+ 6.38
+ I.O23dr.I04
+ .s6odr.o89
+ 9-84
+ 6.29
+ .592dr.I2I
— .368dr.ii4
+ 4.89
- 3.23
H.+
III.
1,200
50
+.5iodr.o63
+ 8.10
+ .792 dr. 069
+ 11.48
+ .II2dr.o84
+ 1.33
IV.
600
100
+.78sdr.i88
+ 4.18
+ 1.043 dr. 197
+ 5-28
— .2I2dr.2IO
— 1 .01
V.
600
100
+ .28odr.I5I
+ 1.85
+ 5-72
+ .890dr.I72
+ 5.17
IV.+
V.
1,200
IOO
+.532dr.I2I
+ 440
+ .968dr.I26
+ 7-68
+ -339dr.i36
+ 2.49
VI.
200
200
+.8osdr.057
+ 14.12
+ I-96odr.058
+3378
+3. 05 5 dr. 077
+39-68
VII.
450
2C
+ .28odr.o62
+ 4-52
+ -75l~^~ °9^
+ 766
VIII.
600
+ .205+ 064
+ 3 20
IX.
2C
+ .805 dr. 067
+ I2.OI
— .o6odr.O58
— 1.03
X.
45O
+ -4"? 3 +-064
+ 6.77
•••-'j
+ 3.04
IX.+
J
X.
1,000
2C
_|_ 6^8-f- 047
+ 17. C7
+ .042 dr. 040
+ i. 06
XI.
600
SO
+.l80rh.090
+ 2.00
+ .22ldr.o87
+ 2.54
+ .342dr.I2I
+ 2.83
1 This is also conspicuously the case in the short series of trials by Observer A,
whose estimates are not to be discussed in detail.
498
/. A. HARRIS
in number or magnitude, or both, is indicated by a mean error
with the positive sign. This will indicate a tendency to under-
estimate any given quantity, since an actually larger number
than that desired was laid out.
Table I. gives the results. Of the 28 means for the indi-
vidual experiments 25 have the positive sign, i. e., in 25 out of
28 experiments the observers had a tendency to lay out too
many seeds.
The results are shown graphically in Diagram 3. Here the
signs, frequencies and amounts of the personal equations are
shown by the direction and length of the lines. The dotted
ac
. EXPERIMENTER D
EXPERIMENTER C
EXPERIMENTER B
2.5
-
20
-
I.S
-
1.0
1
C.S
c.o
-
MEAN PERSONAL
EQUATION
MEAN F
ERSON,
L EQUATION
MEAN PERSONAL EQUATION
I
i ,1
L i ll
i!
ZERO BAR.
ZERO BAR i
ZERO BAR
EXPERIMENT
J II l|l IV V VI VII VIII XI
EXPERIMENT
1 II III IV V VI V,ll IX X XI
EXPERIMENT
, ii in ,y v v, ix x xi
DIAGRAM 3. Sign and Magnitude of (Absolute) Personal Equation in All the
Experiments. The mean deviation from zero for the several experiments is indicated
by the lengths of the lines. The dotted lines indicate the mean of all the experiments
by the individual observers.
transverse lines show the mean values of all the personal
equations for the individual observers.
ERRORS OF JUDGMENT 499
The magnitudes of these means are, however, very low.
For Observer B there is not a single case in which it amounts to
one seed, although in every case it is positive in sign. For
Observers C and D the values are somewhat higher. The
largest single value is that of Observer Z), experiment VI.,
where the estimates were on an average three seeds off. In
this case however there were only 200 trials and the attempt
was being made to estimate in lots of 200 seeds.
To the question of the actual magnitude of these deviations
from zero personal equation I shall return presently. For
the moment, the feature of these results which impresses me
is the fact that so slight a personal bias should be so persistent
among the three observers throughout the two years during
which these observations were carried on.
With regard to the significance of the deviation of these
means from o little need be said. The probable errors of the
mean have been calculated from the usual.
E A-. 67449 -=,
Where <7 is the standard deviation of the series of errors (to
be discussed shortly) and N the number of observations. In
Diagram 3 the amount of the probable error is indicated in
each case by a cross on the ordinate indicating the amount of
personal equation. The ratio of the deviation of the mean
from o to its probable error has been tabled in each case. Of
the 36 constants (including these in which two consecutive
experiments have been combined) 28 are over 2.5 times as
large as their probable errors. Of these, 27 are positive and
one negative in sign. There can be no reasonable question,
therefore, of the statistical trustworthiness of these individual
constants.
One may test most critically the existence of personal
equation by splitting these masses of observations up into
sub-classes, say into the groups of 50 estimates known as
periods. Constants must then be determined for these in
the same manner as for the " general population" as the sta-
tisticians call it.
The detailed analyses of one series of data in this way is
sufficient, since the others will be treated in a slightly different
manner, giving nearly comparable end results, in a subsequent
500
/. A. HARRIS
paper. Table II. for the first experiment furnishes data for
determining whether the personal equation demonstrated in
the massed estimates is persistent throughout the course of
the experiment in the individual periods. In Diagram 4 the
solid line in each panel, the zero bar, shows the o average of
errors of observation which would be secured if there were no
TABLE II
Period
Observer B
Observer C
Observer D
Personal
Equation
Steadiness of
Judgment
Personal
Equation
Steadiness of
Judgment
Personal
Equation
Steadiness of
Judgment
I
+ 1.40
6.24
+ -44
5-34
+ 1.92
6.85
2
+ .82
5.46
+ 1.84
4-93
+2.30
7-03
3
- .80
4.10
+ 1.02
4.13
+3-88
4.89
4
~ .56
3-95
+ I.O2
3-83
+3.02
6.22
5
- -44
3-69
+ 1-72
4-30
+3.78
5.20
6
+ .10
3-95
+ 2.04
3-86
+ 1.36
4.96
7
+ -32
4-23
+ 1.24
4-OS
+ 2.58
4-75
8
- .72
3.87
+ -SO
3-75
+ 1.22
3-95
9
+ .90
3-79
+ 1.40
4.22
+ I-SO
4-34
10
+2.04
3-88
+ .18
4-50
+ 1-74
4-13
ii
+ .20
3 -75
+ -SO
3-35
+ 1.04
4.07
12
+ .32
4.11
~ .36
3.66
+ 1.68
3-95
13
- .44
346
+ 1.06
3-70
+ .80
542
H
~ -58
3-23
+ .36
2.96
+3.56
445
personal equation — i. e., no bias towards too high or too low
estimates. The circles show the actual means for the 14 in-
dividual periods of 50 estimates each. The light line shows
the mean deviation for the whole 14 periods, the amount of
which is forcibly brought out by the shaded area. The sloping
lines show the rate of change. These will be discussed later.
For the whole series of observations, Observer B had within
the limits of the probable error, no personal equation, i. <?.,
her mean deviation from the true value was only + .171 ± .107.
Here it appears that in 8 of the periods her means fall above
and in 6 of the periods below the o bar. For Observer C,
who appeared from the massed observations to have a distinct
personal equation of about one seed, the diagram shows that
in 13 out of 14 cases the period means fall on the positive side
of the line. Finally, for Observer Z), who in this series has
the greatest bias of the three towards underestimating the
number of seeds in a sample — that is, towards laying out too
many seeds in the attempt to get 50 — all 14 period means are
positive.
ERRORS OF JUDGMENT
501
EXPERIMENTER B
EXPERIMENTER C
EXPERIMENTER D
MEAN R E.
ZERO BAR
/ 2 3 4 € 6 7 8 <? /O II 12 /3
PERIOD IN THE EXPERIMENT
DIAGRAM 4. Personal Equation in the Fourteen Individual Periods, Each of
Fifty Trials, of Experiment I. The circles indicate the personal equation for the
individual periods. The shaded area shows the amount of personal equation for the
whole experiment.
502
/. A. HARRIS
With regard to the actual magnitude of the personal
equation, it seems reasonable to assume that if there be a
tendency to err from the true value in any definite direction,
the actual mean deviation observed will be to some degree
dependent upon the number of objects with which the ex-
perimenter is dealing.
The simplest assumption is that the actual amount of the
personal equation in any given case should be approximately
proportional to the number which the experimenter is at-
tempting to estimate. On such an assumption (which on
more extensive investigation may or may not be found to be
borne out by the experimental facts) one may take the ratio
of the actually observed mean deviation to the ideal number.
Concretely, one divides the measures of personal equation
given in Table I. by 25, 50, 100 or 200 as the case may be.
Expressed in this way we have an 'error per object,' or a
TABLE III
RELATIVE PERSONAL EQUATION AND DIFFERENCES IN RELATIVE PERSONAL EQUATION
FOR THREE OBSERVERS IN THE INDIVIDUAL EXPERIMENT
'Series
Observer B
Observer C
Observer D
D— C
D — B
C-B
I.
+ -0034
+ .OI8S
+ .0434
+.0248
+ •0399
+ .0151
II.
+.OIO2
+ .0204
+ .0118
-.0086
+ .OOI6
+ .OI02
III.
+.OIO2
+ .0112
-.0073
— .0185
-.0175
+ .OOIO
II.+III.
+.OIO2
+.OIS8
+ .OO22
— .0136
-.0079
+ .0056
IV.
+.0078
+.0104
— .0021
— .0125
-.0099
+ .0025
V.
+ .0028
+ .0089
+ .0089
— .OOOO
+ .006I
+ .006l
IV. +V.
+.0053
+.0096
+ .0033
-.0062
— .0019
+ .0043
VI.
+.OO40
+ .0098
+ .0152
+.0254
+ .0112
+ .0057
VII
+.OII2
+ .O3OO
+ .0188
VIII
+ .OO82
IX
4- O322
— .OO24
— .O34.6
x
4- OI73
+.0066
•w^w
— .OIOO
IX +X
_l- O2CC
-j-.ooi6
— 0238
XL
+ .0036
+.0044
+ .0068
+.0024
+.OO32
+ .0008
'relative personal equation.' The resulting values are given
in Table III.
The individual entries and the averages show that the
relative personal equation is low. The observers tend to lay
out about i per cent, too many seeds.
Comparisons between the different workers are possible on
the basis of these relative values which may be averaged.
From the differences for the individual experiments
ERRORS OF JUDGMENT
5<>3
(leaving the combined series out of account) the following facts
are to be noted.
Observer B has a higher personal equation than Observer C
in 2 cases and a lower personal equation in 7 cases. The
average difference between them in terms of relative personal
equation is only .0004. Observer B has a higher personal
equation than D in 2 cases and a lower deviation in 5 cases.
The average differences, regarding signs as before, is only
.0050. A comparison of the records of Observers C and D
shows that in 4 experiments Observer D has a greater personal
equation than C whereas in the other 4 experiments precisely
the opposite conditions prevailed. The average difference is
only .0040.
From these experimental data taken as a whole one cannot
conclude that there is any demonstrated difference between
the personal equation of the three observers. All have a bias
in the direction of laying out more than the intended number
of seeds, but that one is worse than another cannot be
asserted.
If now one considers these differences between the personal
equations of the three observers in their relation to their
probable errors, as shown for the differences in the absolute
values given in Table IV., it appears that in a high proportion
of the cases they are statistically significant. This is true in
TABLE IV
DIFFERENCES IN PERSONAL EQUATION FOR INDIVIDUAL OBSERVERS
Experi-
Diff.
Diff.
Diff.
ment
D—C
£Diff.
D— B
•^Diff.
C— B
•^Diff.
I.
-fi.244zb.i70
+ 7-32
+ 1-999 ±-171
+ 11.69
+ .755±.i5i
+ 5-00
II.
— 43i=b.i6o
- 2.69
+ .082±.IS6
+ -53
+ -5i3±.i43
+ 3-59
III.
- .928±.i45
— 6.40
- .878^.139
- 6.32
+ .050±.I2O
+ .42
II.+III.
— .68o±.io9
- 6.24
— .3981*1.105
- 3-79
+ .282±.093
+ 3-03
IV.
— 1.255±.288
- 4.36
- .997^.282
- 3-54
+ .258^.272
+ -95
V.
— .003^.232
— .01
+ .6lO±.229
+ 2.66
+ .6i3±.2i7
+ 2.82
IV.+V.
— .629^.185
- 340
— .I93±.i82
- 1.06
+ .436±.i75
+ 2.49
VI.
+ 1.095 ±.096
+ 11-41
+2.25o±.096
+23-44
+ I.i55±.o8i
+ 14.26
VII
+ A7T4- Tl6
+ A.06
VIII.
IX.
— .865±.o88
— 9.8i
X.
— .266±.o84
— -i. 17
IX.+X.
— .<;Q6±.o62
* V
— Q.OI
XL
+ .121 ±.149
+ .81
+ .162^.151
+ 1.07
+ .041^.125
+ -33
I
504 /. A. HARRIS
cases in which (for example) B has a greater personal equation
than C as well as in these in which she has a smaller personal
equation. The reader may compare the entries in Table IV.
for details.
The statement that there are statistically significant dif-
ferences between two observers for individual experiments,
and that these differences are sometimes positive and some-
times negative may be taken at once by that still considerable
body of students who are hostile to the newer statistical tools
of research to discredit entirely the methods employed and
to cast doubt upon the conclusions just drawn. Such an
attitude seems to me quite unjustified.
The true interpretation of the results seems to me to be
rather that the observers vary somewhat in their personal
equation from experiment to experiment, just as they vary
from time to time in general health, physiological tone, and
mental vigor, alertness, or whatever one may care to call it.
As a result of this variation from time to time one observer
may show an abnormally high personal equation in a partic-
ular experiment in which a second observer shows an un-
usually low one. On an other occasion the condition may be
exactly reversed.
Thus in an individual experiment one observer may seem
to be decidedly better than another. In the long run there is
no fully demonstrated difference between them.
2. Steadiness of Judgment
Steadiness of judgment will best be measured by some
expression showing the scatter of estimates around their
mean. The best constant for this is the standard deviation, a.
o -TV _ /Sum of (deviations from mean)2,
\ Total estimates
which here is most easily calculated from the formula1
S.D. =
where S is the conventional summation sign, N is the number
of estimates and d indicates the deviation of the estimate from
1 Sheppard's modification has not been applied to the second moment.
ERRORS OF JUDGMENT
5<>5
the true number of objects, i. e., the actual number laid out
less the required number.
The constants with their probable errors are given in the
first three constant columns of Table V.
TABLE V
Experiment
Standard Deviation
Coefficient of Variation
Observer B
Observer C
Observer D
Observer B
Observer C
Observer D
I.
4.180^.075
4.I39db.075
5.2Ild=.094
8-331
8.128
9.989
II.
3.552^.069
3-774db.o74
4.407 ±. 086
7.032
* 7-397
8.7II
III.
2.899^.056
3. 243 ±. 063
4.i37±.o8i
5-739
6.414
8-335
II.+IH.
3. 242 dr. 045
3.527dz.049
4.301^.059
6.418
6.943
8.583
IV.
6.839^.133
7.i6o±.i39
7.635^.149
6.785
7.086
7.65I
V.
5.498^.107
5.682±.in
6.248d=.I22
S483
5.632
6.193
IV.+V.
6.2io±.o86
6.464^.089
6.998±.096
6.177
6.402
6.974
VI.
1 1. 928 ±.040
1 2. 089 dz. 04 1
i6.i72d=.055
S-940
5.986
7.964
VII.
I.937d=.O44
/?.O72±.o6o
7.663
II.Q2Q
VIII.
2.3I4±.O4t;
o.1 70
IX.
2 •j'j-j-l- oj.8
2 OT3 + O^T
O O4.2
8 072
X.
2 OTO-f- 045
I 7IA-4- O3O
7 OO4.
6.813
IX.-f-X.
2 2O2-{- 033
i 887± 028
8.587
y.C'ic
XL
3.255zfc.063
3.i47dz.o6i
4.399±.o86
6.486
6.166
8.738
For steadiness of judgment, we have no absolute standard
comparable to a mean deviation of o in the personal equation
test. The accuracy of an observer must be estimated by
comparison with others. A relative measure of steadiness of
judgment permitting comparison between different kinds of
experiments is desirable. Since it is reasonable to suppose
a priori that errors of estimation will be larger when the ob-
server is attempting to lay out samples of a large number of
seeds than when she is dealing with a small number,1 this
relative measure is best furnished by the biometrican's coef-
ficient of variation, which is obtained in this case by dividing
the standard deviation multiplied by 100 by the ideal number
plus or minus the observed personal equation, as the sign of
the latter may indicate.
These relative measures of steadiness of judgment are
given in the last three columns of Table V.
The differences in standard deviation measuring steadiness
1 Should the number be made very small indeed there would be practically no
error of estimation after a little experience, since the observer could all but invariably
lay out the correct number.
J. A. HARRIS
of judgment between the three observers are set forth with
their probable errors and their ratios to their probable errors
in Table VI.
TABLE VI
DIFFERENCE IN STANDARD DEVIATION FOR INDIVIDUAL OBSERVERS, THAT is, DIFFER-
ENCE IN STEADINESS OF JUDGMENT
Experi-
ment
D-C
Diff.
^Diff.
D — B
Diff.
^Diflf.
C— B
Diff.
^Diff.
I.
+ I.O72zfc.I2O
+ 8.93
+ I.O3I±.I20
+ 8.59
— .04i±.io6
- 0.39
II.
+ .633±.ii3
+ S-6o
+ .8ssdb.no
+ 7-77
+ .222±.IOI
+ 2. 2O
III.
+ .894±.io2
+ 8.76
+ 1.238^.098
+ 12.63
+ .344±.084
+ 4-10
II.+III.
+ .774=11.076
+ 10.18
+ 1.05 9 ±.074
+ H-3I
+.28s±.o66
+ 4.32
IV.
+ .475 ±.204
+ 2.33
+ .796^.200
+ 3-98
+ .32Izb.I92
+ 1.67
V.
+ .566^.165
+ 3-43
+ .75Odz.l62
+ 4-63
+.i84db.iS4
+ 1.19
IV. +V.
+ .534^.131
+ 4.08
+ .788±.I29
+ 2.05
VI.
+4.083 ±.069
+59-17
+4.244dz.o68
+62.41
+.i6i±.os7
+ 2.82
VII
-4-1 i^ci 082
+ 13.84
VIII
•*
IX
— .3 20+ .063
— «; 08
x
— .2Q6+-OCO
— 5.O2
IX +X.
— .-j ic d=.O43
— 7-33
XI.
+ I.2S2±.I05
+ 11.92
+ I.i44±.io7
+ 10.69
— .io8±.o87
- 1.24
The differences for the standard deviations of the three
observers are more consistent than those for personal equation.
Observer B has more erratic judgment than C in 4 cases, less
erratic in 5 cases. In 5 out of the 9 cases the difference may be
considered to be significant with regard to its probable errors.
In 2 of the experiments it is Observer B who is significantly
more variable in estimation, while in 3 cases it is Observer C
who has the most irregular estimates. These significant
differences which differ in sign from experiment to experiment
are probably to be explained in the same way as those in
personal equation discussed above. The relative steadiness
of judgment as measured by the coefficient of variation shows
a mean of 6.971 in the case of Observer B as compared with
6.946 in the case of Observer C, a difference of only 0.025!
Thus in the long run there is no discernible difference in the
steadiness of judgment of B and C, although in the case of
individual experiments now one, now the other, may be higher.
For the comparison between both B and D and C and D
the case is quite different. In every individual experiment
Observer D has a higher standard deviation, or in other words
ERRORS OF JUDGMENT
5<>7
less steady judgment, than either B or C. In practically every
instance the differences may be considered significant in
comparison with their probable errors. The average relative
scatter of estimates of Observer D as measured by the coef-
ficient of variation is 8.743, a value about 1.80 higher than
that of either of the other observers.
III. RECAPITULATION
The purpose of this paper, and of another on the influence
of previous experience upon errors of judgment which is to
follow, is the presentation in terms as succinct as possible of
the results of a series of experiments on errors of judgment in
the estimation of moderately large numbers of objects.
DATA TABLE A
Amount
of Error
I. A
I. B
I. C
l.D
II. B
II. C
II. D
III. B
III. C
III. D
0
4
61
79
4i
62
64
61
76
80
61
+
-
+
-
+
-
+
-
+
-
+
-
+
-
-r
-
-r
-
-r
—
I
7
60
73
64
18
10
42
77
62
73
63
60
18
76
69
77
64
11
2
8
9
61
13
63
61
63
43
16
17
61
34
11
18
71
14
13
ij6
44
61
3
8
9
65
55
58
46
56
42
38
44
49
44
53
43
70
38
56
43
43
47
4
6
3
S^
32
44
36
49
22
38
41
1i
1C
27
34
10
21
38
28
30
Ii
5
7
9
33
36
42
22
46
18
31
20
39
16
27
22
18
22
31
16
21
32
6
8
I
19
18
32
14
48
8
18
9
22
12
20
14
8
6
17
ii
13
18
7
7
6
20
17
31
II
14
6
14
4
19
9
8
4
10
4
13
8
4
3
6
9
6
33
10
9
4
12
3
10
6
3
i
4
2
6
16
9
5
i
9
7
7
2
19
6
6
7
2
5
5
i
—
i
—
5
5
10
5
3
4
3
7
2
12
3
2
—
2
2
5
i
—
—
3
—
3
3
ii
7
—
2
i
3
I
13
2
—
—
I
2
3
i
—
—
2
—
i
—
12
i
—
I
2
2
I
7
3
2
—
I
I
—
—
—
—
—
—
2
2
13
3
i
3
—
4
—
5
—
—
—
I
—
i
—
—
—
—
—
—
—
IS
3
—
—
—
i
—
4
i
i
16
3
—
i
7
4
iS
—
—
—
I
3
19
22
—
—
23
32
I
—
91
46
330
309
361
260
447
212
295
243
338
I98
288
251
309
215
296
224
243
296
The experiments consisted in attempts to lay out samples
of a definite number of small objects. The number which
the observer was attempting to obtain in each sample (25, 50,
5°8
/. A. HARRIS
100, or 200) was constant for considerable periods. The error
of each estimate was at once determined and recorded by the
experimenter, who on the basis of these known errors made a
continuous effort to improve in accuracy of estimating.
Two characteristics of the series of errors of estimation
made by the three observers are here considered — personal
equation and steadiness of judgment. By personal equation
DATA TABLE B
Amount
of Error
IV
B
IV
.C
IV
D
V.
B
V.
C
V.
D
VI]
. C
VII
.D
0
3
i
3
7
2
7
4
7
3
0
4
7
9
5
6
6
+
-
+
-
+
-
+
-
+
-
-f
-
+
-
+
-
I
37
34
32
28
31
35
49
42
47
43
35
32
86
69
63
57
2
34
34
32
31
27
30
44
42
46
40
47
36
63
52
52
3
34
31
44
27
18
42
36
29
34
37
38
29
28
2O
36
25
4
33
31
34
25
23
35
36
28
29
24
33
29
21
9
28
22
5
25
25
20
22
22
31
27
28
33
30
34
27
4
i
16
4
6
20
29
30
2O
31
35
23
30
19
29
16
i
i
8
i
7
20
17
14
24
24
17
16
22
23
21
20
16
—
—
6
—
8
12
22
19
10
13
19
10
12
2<;
16
20
15
—
—
4
—
9
29
10
17
14
9
18
ii
18
10
12
12
—
—
3
—
10
IO
7
21
8
IO
16
ii
2
7
5
9
6
—
—
i
—
ii
9
8
14
12
8
9
10
6
4
2
5
4
—
—
i
—
12
6
5
II
6
9
6
3
4
7
3
13
5
—
—
i
—
13
ii
5
8
4
7
5
3
i
IO
i
3
2
—
—
—
—
14
3
3
6
i
9
2
2
i
2
i
6
2
—
—
—
—
IS
6
3
7
6
4
4
3
—
—
I
i
3
—
—
i
—
16
4
i
3
i
2
3
—
I
—
i
2
—
—
—
—
17
4
i
3
—
4
—
i
—
I
—
6
I
—
—
—
—
18
i
—
2
i
i
2
—
—
I
—
i
—
—
—
—
19
4
—
I
i
2
I
—
—
—
—
—
I
—
—
—
—
20
—
—
—
i
2
—
—
—
—
—
i
—
—
i
—
21
—
—
—
—
3
I
—
—
I
—
i
—
—
—
—
22
—
—
i
—
—
—
—
—
I
—
—
—
—
—
—
23
24
25
—
—
i
—
—
3
—
—
—
—
—
• —
—
—
—
—
27
32
—
—
—
—
i
—
—
—
—
—
—
—
—
—
—
303
266
320
243
259
314
288
265
317
253
315
238
203
152
223
161
we understand a bias in a given direction — a tendency to
estimate too high or too low. By steadiness of judgment we
mean consistency in estimation as measured by the closeness
with which the errors of estimation clusters around their
mean value.
Personal equation is measured by the mean (regarding
signs) of the deviations of the samples from their ideal value.
ERRORS OF JUDGMENT
5°9
Steadiness of judgment is expressed in the absolute terms of
the standard deviation of the errors of estimation about their
mean, or in the relative terms of the coefficient of variation.
In the case of all three observers there is a slight but sig-
nificant personal equation, which, notwithstanding the
DATA TABLE C
Amount of
Error
VI. B
VI. C
VI. D
0
6
8
i
+
-
+
-
+
-
I
7
5
3
7
8
3
2
8
5
4
8
6
S
3
3
7
i
4
6
4
4
—
4
10
S
7
2
5
5
8
5
3
6
4
6
8
2
10
3
7
4
7
8
I
S
8
5
4
8
3
3
5
—
S
2
9
8
4
10
4
4
2
10
9
4
9
2
7
5
ii
6
S
8
6
S
12
5
3
2
2
5
S
13
i
S
2
2
9
6
H
4
2
6
4
4
3
15
3
2
4
3
i
—
16
6
4
3
i
2
3
17
i
i
S
2
2
2
18
2
—
2
4
4
I
19
3
i
3
2
4
3
20
2
2
—
—
—
2
21
2
I
—
I
3
3
22
2
2
2
—
3
2
23
2
I
I
2
I
24
2
S
2
I
—
2
25
I
I
I
3
2
26
3
—
2
—
I
27
i
2
2
2
2
I
28
—
3
I
2
—
29
—
—
—
2
2
I
30
—
2
I
—
3
—
31
—
I
—
—
—
32
—
I
—
—
—
33
i
—
I
I
—
34
—
I
—
—
—
35
—
I
I
I
I
36
—
—
—
—
37
i
—
I
I
—
38
—
I
—
I
39
—
'
—
—
—
40
—
—
—
I
—
42
—
—
—
I
I
43
i
—
—
—
7i
—
—
—
I
—
107
87
Ill
81
123
76
/. A. HARRIS
constant effort to improve, persisted throughout the two
years during which the experiments were intermittently made.
In only three out of the twenty-eight experiments did the
observer lay out samples of too small average size. In a
large number of the individual experiments the personal
equation is certainly statistically significant (trustworthy) in
comparison with its probable error.
It is impossible on the basis of the present series of experi-
ments, extensive though it is, to assert that the personal
DATA TABLE D
Amount of
VIII. D
IX. B
IX. C
X. B
X. C
Error
0
103
83
116
82
109
+
-
+
-
+
-
+
-
+
-
I
97
97
91
70
91
104
74
72
99
82
2
79
52
50
68
79
45
48
48
3
35
56
24
34
38
34
23
28
19
4
28
20
35
13
22
16
23
9
9
4
5
8
7
13
2
5
3
7
i
2
—
6
8
i
ii
I
I
2
i
—
—
i
7
i
2
3
—
—
—
I
—
8
i
i
—
—
—
—
—
—
9
i
—
—
—
—
—
—
—
10
—
i
—
—
~~
—
~~
—
264
233
305
162
203
231
218
ISO
I87
154
equation of any one of the three observers is on the whole
higher than that of the others, although the figures do suggest
that the bias of observer D may be slightly greater than that of
either of the others. In the case of individual experiments
there may be significant differences between two observers.
In one experiment x may have a decidedly lower personal
equation than y, while in another period of observation
exactly the reverse condition may be found. This is taken
to indicate a variation in the magnitude of the personal equa-
tion of an observer from experiment to experiment.
For steadiness of judgment there is no absolute standard
comparable with the zero mean deviation of the personal
equation. The data show a coefficient of variation about 6.9
per cent, in the case of Observers B and C, and of 8.7 per cent,
the case of Observer Z), who has a decidedly greater scatter
in her estimates — that is a far less steady judgment — than
ERRORS OF JUDGMENT
either of the other observers. Indeed, in every individual
experiment her standard deviation is higher than that of
either of the two other experimenters.
Thus while there is no certain differentiation among the
experimenters in personal equation, they differ distinctly in
steadiness of judgment.
For a more detailed consideration of these two character-
istics the reader must see the subsequent paper.
Finally, I must emphasize again the fact that these data
DATA TABLE E
Amount of Error
XL B
XI. C
XI. D
0
7i
78
S*
+
-
+
-
+
-
I
68
72
66
86
55
5°
2
57
57
64
59
50
55
3
58
49
48
50
5i
40
4
40
28
33
23
40
36
5
23
26
25
23
24
27
6
12
12
13
7
20
22
7
II
6
10
6
17
II
8
4
i
2
2
ii
10
9
i
2
2
I
7
4
10
—
I
—
5
i
ii
—
—
I
—
3
i
12
i
—
—
—
3
—
13
—
—
—
—
i
I
14
—
—
—
—
—
—
15
—
—
I
—
—
—
16
—
—
—
—
2
—
19
—
—
—
—
—
i
275
254
265
257
289
259
are presented purely for their intrinsic value. They were
secured quite incidentally in the carrying out of large plant
breeding experiments. The chief value of the observations
perhaps lies in the fact that they represent far larger experi-
ments than the average professional psychologist is able to
make. Comprising as they do 28 experiments due to three ob-
servers all of whom carried on the work at considerably sepa-
rated intervals over a period of two years, during which they
made over 15,000 estimates, the constants have a reliability
which cannot possibly be attributed to short series. Purely
psychological discussions, even the review of literature with
some of which the writer is quite familiar, is left to specialists.
ORIGIN OF HIGHER ORDERS OF COMBINATION
TONES1
BY JOSEPH PETERSON
University of Minnesota
It is well-known that Helmholtz gave three or more
different explanations of the origin of combination tones.
According to his own statements combination tones may be
generated (i) from the clicking action between the hammer
and the anvil of the ear, when the primaries are powerful;2
(2) from the asymmetry in vibration of the tympanum; and
(3) from disturbed superposition of vibrations due to some
objective connection between the primary periodicities, such
as a common windchest found in the polyphonic siren or the
harmonium, making the air puffs for each tone weaken
periodically the puffs for the other tone. The tones generated
by conditions (i) and (2) were called subjective; those by
condition (3), objective. Helmholtz did not seem to regard
case (3) as a condition of disturbed superposition as demanded
by his mathematical explanation based upon vibrations 'so
large that the square of the displacements has a sensible influence
on the motions';3 for he states explicitly of this case that he
will 'draw attention to a third case, where combinational
tones may also arise from infinitely small vibrations.^ This
is of course an error. Helmholtz admitted that in conditions
favoring objective combination tones conditions (i) and (2)
were also operative, thus making all audible objective com-
bination tones also largely l subjective.'5 I have suggested6
that conditions (2) and (3) are practically identical physically
1 Read before the Utah Academy of Sciences, April 3, 1915.
2 'Sensations of Tone,' p. 158.
3 Ibid., Appendix XII., 412.
4 Ibid., 419. Italics mine.
5 Ibid., 157.
6 'Combination Tones,' etc., PSYCHOLOGICAL REVIEW MONOGRAPH, No. 39, 1908,
17 ff., 103 ff.
512
COMBINATION TONES 5*3
both being dependent upon the principle of disturbed super-
position of vibrations; that such superposition is in all
probability in case (2) most pronounced in the liquids of the
inner ear. This view has subsequently been supported by
Clemens Schaefer.1
The view of Helmholtz, expressed under case (i) above,
has not been substantiated by recent research. My own
experiments lend support to the conclusion of Helmholtz
respecting objective combination tones, that they are in
large part 'subjective,' i. e., that they are in a considerable
measure due to disturbed superposition of vibrations of the
primaries within the ear itself.
As to the nature of higher orders of combination tones two
contradictory views were expressed by Helmholtz. One was
that they are higher order difference tones in the sense
suggested by Hallstrom, that they originate from a first
difference tone and a primary tone or from two difference
tones.2 His other view was that all higher order combination
tones take origin directly from the primaries. This is in
agreement with a theory developed mathematically in 1881
by R. H. M. Bosanquet,3 and more recently by Clemens
Schaefer in the article referred to in a preceding paragraph.
This view that so-called higher order combination tones take
their origin directly from the primaries is also supported by
the experimental results of a number of recent investigators.
From the method of his statement of the laws of the occur-
rence of difference tones, Krueger has been interpreted by
some writers to favor the Hallstrom view of 'higher order'
combination tones. This interpretation, which Krueger
assured me personally in a conversation4 is wrong, is used by
R. M. Ogden as evidence against Krueger's theory of con-
sonance. Ogden writes that "Stumpf's investigations in-
dicate that combination tones are always directly derived
from the objective tones, and not from beats, nor, except in
the highest ranges of the scale, from one another as Krueger
* Annalen der Physik, Bd. 33, 1910, 1216-1226.
2 'Sensations of Tone,' p. 154.
8 Phil. Mag., 5th Series, XL, 1881.
4 At Cleveland, Ohio, December, 1912.
5H JOSEPH PETERSON
maintains."1 This is an unfortunate mistake as to Krueger's
view, and, as I hope to show in another place, the actual facts
really support rather than refute his theory of consonance.
As to the facts, I am willing to go farther than either Stumpf
or Ogden and assert that there is no satisfactory evidence in
existence to support the view quoted above, that 'in the
highest ranges of the scale' combination tones may originate
from other combination tones. This is a matter of importance
in relation to theories of consonance.
Though Stumpf quotes approvingly my own view as to
the so-called higher order combination tones, and seems to
accept and to use my experimental evidence, he registers his
unwillingness to dispute so generally as I have done the
possibility of difference tones being derived from overtones
of the primaries.2 No evidence is offered for such difference
tones except that Stumpf has perceived, as other investigators
have perceived, difference tones lying in pitch between the
primaries. Such tones are, of course, easily accounted for
on the theory of direct origin from the primary tones on the
basis of disturbed superposition; but the fact of their existence
is contradictory to theories like that of Max Meyer. Riicker
and Edser3 established the existence of objective intermediate
difference tones, and similar 'subjective' tones are to be
expected on the view that disturbed superposition is the cause
of all combination tones.
While Helmholtz held to the position maintained by
Stumpf as to the derivation of certain difference tones from
upper partials, experimental evidence indicates that few
difference tones, if any at all, have such origin. Though the
possibility of such tones is not disputed by theory — for these
tones would follow the same laws as combination tones from
the primaries — there is, so far as I know, no evidence at all
to indicate that audible difference tones derived from upper
partials exist. This is not incomprehensible when one takes
into consideration the fact that the overtones of the primaries
1 In a review of Stumpf in Psychol. Bui., 9, 1912, 117.
2 Stumpf, C., ' Beobachtungen iiber Kombinationstone,' Zeit. f. Psych., 55, 1910,
iff.
8 'Objective Reality of Combination Tones,' Phil. Mag., 5th Series, 39, 1895.
COMBINATION TONES 5J5
are themselves to a considerable extent ' subjective,' or de-
pendent upon periodicities arising in the liquids of the inner
ear.
In the case of the second difference tone, D2, which is
easily perceptible with a number of intervals, is there any
evidence that it can be generated by the higher primary tone
and the second partial of the lower primary, as suggested by
the formula D2 = 2,1 — &? In the psychological laboratory
of the University of Chicago the resonated forks Ut4 : Mi4
(4 : 5) give an exceptionally prominent second difference
tone, 3, on very gentle sounding, i. e., at an intensity of the
primaries which leaves the first difference tone, DI, entirely
inaudible.1 D2 can evidently not arise from the first difference
tone and the lower primary. On substituting Ut5 for Ut4 —
sounding gently Uts with Mi4 — one should make the tone D2
even more prominent, if it depends upon the tone correspond-
ing to Ut6, by the formula 4-2 — 5 = 3. But with such
substitution a much greater intensity of the primaries is
required to make D2 (3), now DI, audible at all. The first
upper partial of piano tones in the middle of the register is
easily perceptible even to the unaided ear — as are several
others of the higher partials. In the case of the major third
(4 : 5) with the following tones c2 : e2, b2: e[?2, and d2 : ^ the
respective second difference tones at approximately the pitch
of g1, gl?1 and a1 were easily audible, though the several first
difference tones were inaudible. But when the octave of
the lower tone in each pair was substituted for this tone (i. e.,
c3 for c2, b3 for b2, d3 for d2) the difference tone in question in each
case became inaudible even though the intensity of the
primaries remained the same as before the substitution, or
was slightly increased. The tone in question was therefore
evidently not dependent upon the upper partials.
In the case of forks Ut4 : Mi4 (4 : 5) the first difference
1 It is well known to those who have experimented with tones that such things as
tuning forks have individualities almost as marked in certain respects as those of persons.
Certain pairs of forks of the few sets with which I have become well acquainted give
clear difference tones while others of precisely the same pitch, and resonated similarly,
give very weak ones or none at all. The absolute pitch of the primaries is also an
important factor in determining the presence or absence of combination tones.
5 1 6 JOSEPH PETERSON
tone, I, is easily audible when the forks are loudly sounded.
Is the second difference tone dependent upon this one? The
fork Ut2 gives a tone coincident in pitch with this first differ-
ence tone and, on gentle sounding, almost indistinguishable
from it in timbre. Even though this is the case, when Ut2
is substituted for the primary Mi4, that is when the interval
is represented by the forks Ut2 : Ut4, the second difference
tone, 3, now made the first, disappears altogether. This is
true whether Ut2 is sounded gently — to represent the dif-
ference tone for which it is substituted — or loudly. It is
evident, therefore, that the second difference tone, 3, of the
major third in question does not arise from the first with the
lower primary, by the formula / — (h — /) = / — DI = Z)2
or 4 — (5 — 4) = 4 — i = 3. If this is true of the second
difference tone it is also true of summation tones if they are
really difference tones of higher orders.
But there is yet a better test available. In the case of
summation tones theoretically explained according to the
formulae
h + I = n(h - /) and h + I = nh - ml
a conclusive quantitative test is applicable, one, however, that
was first worked out by the writer only a few years ago. In
the case of the fifth (2:3) the difference tone of the fifth
partials (10, 15) coincides with the summation tone (5). The
summation tone is not easy to hear unless its origin is objective
to the ear. However with the use of resonated tuning forks
I found that it was perceptible to the practiced ear when the
forks Ut3 : Sol3, or Re^s : La(?3 were sounded a little above
medium intensity. As an objective check a number of
students trained in experimental psychology were asked to
select from among a number of high pitched forks the tones
that were audible in addition to the primaries. With the
intervals given were also Fa3 : La3 (4:5) and Ut3:Las
(3 : 5). The fork representing the summation tone was in
each case selected, as well as forks representing certain first
upper partials that were easily perceptible to a trained ear.
Occasionally I suggested to the subject a wrong fork, but it
COMBINATION TONES 517
was in every case rejected finally. One of the subjects readily
sang the note representing the summation tone. There could
be no doubt as to the existence of the tone in question.
By tuning an auxiliary tone to very nearly the pitch of
these summation tones well-marked beats were obtained.
The auxiliary tone is generated by an unresonated fork
sounded gently and held close to the ear. After the auxiliary
tone is exactly tuned to the pitch of the summation tone, any
change in pitch of the primaries is at once noted by a beating
between the auxiliary and the summation tone. On this
principle it was found possible to determine the true nature
of the latter. Take, for instance, the fifth. The primary
tones and their respective upper partials are represented by
the following numbers:
2, 4, 6, 8, 10, etc.
3, 6, 9, 12, 15,
Now it is evident that if the tone in question is a true sum-
mation tone, lowering one of the primaries one beat per second
will also lower it one beat. But if it is really a difference tone
it will in this case take origin from the fifth partials of the
primaries, i. e., from 10 and 15. In such a case lowering one
of the primaries one beat will lower or raise the tone in
question five beats per second. But in the practical working
out of this test a difficulty was encountered. Lowering the
pitch of a primary tone makes the interval imperfect, and the
beats of interference prevent a careful study of the summation
tone with the auxiliary. To get rid of these irrelevant beats
the following procedure was adopted: After depressing one
of the primaries one vibration per second, the auxiliary
fork was attuned to the summation tone, a change that could
be accomplished with certainty. The primary tone was then
brought back to its true pitch. The only beats then re-
maining were those due to the interference of the auxiliary
with the summation tone. The demonstration was convinc-
ing; there was absolutely no trace of rapid beats, but slow
beats with a rate of one per second were heard. If these
beats were due to interference between a difference tone of
5 1 8 JOSEPH PETERSON
the auxiliary and a primary with the other primary, as might
be argued, one should hear also rapid beats of the auxiliary
with this high tone, if the latter originated from the upper
partials. It will be remembered that only such summation
tones as were audible were used in the test. Besides, rapid
beats of five per second are more easily perceptible than slow
ones of one per second. No such beats, however, were audible,
and the auxiliary was so weak that no difference tones were
probably generated; certainly none was perceptible. The
summation tones tested were consequently not explicable as
difference tones of upper partials, but were real summation
tones in the Helmholtzian sense. These experiments, then,
all contradict the view that combination tones may originate
from upper partials. While they do not show the absolute
impossibility of such origin, we must remember that no case
to prove the derivation or combination tones from upper
partials exists, and that all the audible overtones of musical
intervals are likely to a considerable extent intra-aural in
origin.
XXIII. PRACTICE IN GRADING AND IDENTIFYING
SHADES OF GRAY
BY WARNER BROWN
The following report consists of two parts. The first
part deals with the effects of practice in grading during ten
sittings in which an attempt was made to arrange fifty gray
cards in the order of their apparent brightness. No cor-
rections were made. The result shows that there was no
improvement in the accuracy of the arrangement. The
second part, on the other hand, shows that in the course of
the ten sittings there was a substantial improvement in the
ability to remember and recognize four of the fifty cards.
i. PRACTICE IN GRADING
The experiment was performed twice a week for five
weeks. Each worker was provided with two sets of fifty
cards 2.5 cm. wide and 6.5 cm. high. The cards were covered
with gray paper ranging from nearly white to nearly black.
After being shuffled the cards of one set were laid out by the
learner in a row according to the following directions:
The fifty gray cards are numbered on the backs according to an arbitrary system.
You are to arrange these cards according to their apparent brightness, with the lightest
on the left, the darkest on the right, and the others graded in between. Do not take
more than 15 minutes for this part of the experiment. After you have arranged the
cards write down, in the table below, the numbers on the backs of the cards in order.
These directions were printed on a record-blank which
also contained the table with fifty spaces for entering the
numbers found on the cards. A fresh copy of the blank was
supplied to the worker at each practice sitting. Emphasis
was laid on the fact that the numbers on the backs of the
cards were unreliable and that there was no such thing as a
standard arrangement of the cards; but at the same time it
was confidently asserted that a good record in the later part
of the experiment (to be described shortly), depended upon
a scrupulously careful arrangement of the cards in order of
520 WARNER BROWN
brightness, for the amount of error in the memory and recog-
nition test was to be measured in terms of the worker's own
scale of brightness, and any irregularities in the scale would
tend to increase the score of his errors.
In order to determine whether any improvement in the
accuracy of the arrangement resulted from practice it was
necessary to discover the correct arrangement of the cards.
As the cards differ considerably in color and in character of
surface it was thought that an empirical determination of
the correct order would be preferable to a physical measure-
ment of brightness in which these disturbing factors would
escape consideration.1
The empirical determination of the order of brightness of
the cards was made from the records of 37 persons. As
each worker had made 10 arrangements there were 370 judg-
ments upon which to base the average position of each card.
The data from this computation are presented in Table I.
This table shows not only the average position assigned to each
card (with the mean variation) but also, under the heading
"most probable position," the position which each card would
occupy in a series of fifty if they were arranged as nearly as
possible in order in a single series according to the average
judgment. This last arrangement takes account of the fact
that while some of the cards have almost the same average
position they can not occupy the same position in an actual
arrangement.
When the correct position for any particular shade has
been determined it is possible to measure the accuracy of the
arrangement for any of the ten practice sittings by the amount
of the average deviation of the different persons from that
1 The cards used in this experiment were prepared from a set of 'Hering' gray
papers. The numbers given in the first column of Table I. are those found stamped on
the original rolls, but it is evident from the data of Table I. that many of these numbers
are without meaning, either because they fail to designate the approximate position in
the series (compare No. 20, No. 25, or No. 31), or because different numbers are
assigned to papers of almost the same brightness (compare Nos. 20 and 28, or Nos. 25
and 32). Enquiry elicited the information from the manufacturers that they made
up different shades at different times, according to the demand, and that the different
batches were not alike, so that when they sold an entire set the purchaser received
portions of different and dissimilar sets.
IDENTIFYING SHADES OF GRAY
TABLE I
521
Arbitrary Number to
Designate the Shade1
Average Position
Mean Variation
Most Probable Position
I
1. 00
O.OO
I
2
2.00
O.OO
2
3
3.00
0.00
3
4
4.06
0.12
4
5
5-21
0.37
5
6
5.76
0.39
6
7
6.97
0.08
7
8
8.00
O.02
8
9
9-39
0.48
9
ii
10.27
0.78
10
10
n-45
1.04
ii
12
11.71
0.95
12
13
13.10
0.85
13
14
13.16
0.83
M
15
I5-56
0.87
15
17
17.49
i-4S
16
16
17.64
i-53
17
18
17.84
1.42
18
19
18.39
1.56
19
23
26.82
20
21
21-35
1.66
21
24
21.47
i-73
22
22
21.86
23
20
25.24
i.'86
24
28
25-36
1.63
25
29
26.26
1.77
26
27
26.67
1.51
27
30
27.29
1.70
28
33
29.46
1.52
29
26
30.04
1.72
30
32
30.42
1-37
31
25
30.60
1.52
32
35
34-35
1-43
33
37
34-59
1.53
34
34
35-64
i. 60
35
39
35.96
1.52
36
40
36.40
i-75
37
36
37-66
1.41
38
31
38.23
1.74
39
38
38.70
1. 21
40
45
42.09
I.2I
41
43
42.86
0.96
42
44
42.90
1. 12
43
41
43.04
1.18
44
42
44-75
1.04
45
48
46.02
0.80
46
47
46.56
o.SS
47
46
47-79
o.39
48
49
49.00
O.IO
49
50
49-94
O.II
So
position. A larger deviation means a larger number of cards
wrongly placed. Table II. shows the average deviation for
each of the ten days of practice. Fewer cards were wrongly
1 See foot-note on preceding page.
522 WARNER BROWN
placed the first day than any other day. There is not the
slightest tendency to reduce the number of misplaced cards
as practice advances. These conclusions are the same whether
the deviations are calculated from the average positions or
from the 'most probable positions' as defined above. Prac-
tice does not increase the accuracy of the work of grading the
shades of gray.
TABLE II
THE AMOUNT OF INACCURACY IN GRADING THE 50 GRAY CARDS ON EACH OF 10 TRIALS
The average amount of displacement from the true position has been calculated for
37 persons for each of the 50 cards. The 50 figures so obtained have been averaged
to give the figures in the table.
Av. Deviation from the
Av. Position of the Av. Deviation from the
Card; All Cards Most Probable Position
Trial Combined All Cards Combined
1 0.95 0.93
2 1.04 I.O3
3 1.08 .10
4 1-13 -13
5 i. ii .11
6 1.04 .04
7 i. ii .12
8 1.06 1.07
9 i. 08 i. 08
10 i. 06 1.07
2. MEMORY TRAINING WITH SHADES OF GRAY
Most experiments intended to demonstrate an improve-
ment of memory with practice require so much time and such
an unusual amount of patience that they can not be made use
of in connection with an ordinary course of laboratory ex-
periments. The present experiment shows a very sub-
stantial improvement of memory, or at least a form of memory,
in the course of no more than five hours of actual work and
with very little of the effort usually involved in learning.
The portion of the printed directions dealing with the test of
memory is given herewith; the conditions of work will be made
more clear in what follows.
According to the program for each day's work, given below, pick out four cards
from the series as you have arranged it, and observe them carefully, without looking
at the numbers on the backs of the cards. Put the first set of cards away. Spread out
the other set, face up, on the table. Pick out the same four shades, as nearly as you
can, from this second set. Mark in the table the numbers found on the backs of these
four cards.
IDENTIFYING SHADES OF GRAY 523
PROGRAM
First day. Pick out the loth, 2Oth, 3Oth, and 4<Dth.
Second day. " nth, 2ist, 3ist, and 4ist.
Third day. " 9th, iQth, 29th, and 39th.
Fourth day. " I2th, 22d, 32d, and 42d.
Fifth day. " 8th, i8th, 28th, and 38th.
Sixth day. " I3th, 23d, 33d, and 43d.
Seventh day. " 7th, lyth, 27th, and 37th.
Eighth day. " I4th, 24th, 34th, and 44th.
Ninth day. " 6th, i6th, 26th, and 36th.
Tenth day. " loth, 2Oth, 3Oth, and 4Oth.
The record of the day's work, when filled out, shows:
(1) the actual series of grays as arranged by the student;
(2) the four shades selected by him in the memory test;
(3) the amount of his error for each shade. The amount of
the error depends upon the worker's own arrangement of
the cards. Suppose he has sought for the loth, 2Oth, 3Oth, and
4<Dth cards, and that the shades which he has picked out in
the memory test, as they lay scattered in irregular order on
the table, are the nth, 2Oth, 33d and 39th, according to his
own arrangement as shown in his record and regardless of the
true brightness or arbitrary designations of the cards; the
errors in this case would be plus one, zero, plus three and minus
one, or disregarding the direction of error, a total of five.
The worker was informed at once of the actual numerical
amount of his errors, but he was not allowed to look again
at the cards themselves. It will have been observed that
different shades were used for the tests on different days so
that the practice did not consist of learning to recognize
certain shades but was confined to the more general process
of learning to perform the mental operation of recognizing
this kind of thing.
The actual amount of improvement in the memory test
was large. Table III. shows the work of three seasons. In
1911 the recognition test was made with the same set of cards
that was used in the memorizing by simply shuffling them up.
This involved the difficulty that if there were any spots or
other marks on the cards studied they might help in the work
of picking out. As the cards suffer considerably from wear
5^4
WARNER BROWN
TABLE III
THE DAILY AMOUNT OF ERROR IN RECOGNIZING 4 SHADES OF GRAY OUT OF A SET
OF 50
The figure is the average per person of the difference between the position-numbers
of the 4 shades which were selected and the position-numbers of the 4 shades which
were sought and should have been selected. The amount of error decreases with
practice.
Trial
1911,
33 Persons
'913.
37 Persons
i9I3,
Errors on
Probable' Scale
1914.
22 Persons
I
IS-5
I4.8
14.9
14.2
2
134
I6.4
I6.3
"•3
3
12.7
n.8
12.8
9.9
4
8.3
12.3
12.2
9.6
I
9.1
8.2
9.6
n.6
IO.O
II. I
7-5
9.9
I
8.2
9.5
10.5
8.6
II.4
9.0
6.4
9.8
9
5-9
8.2
8.2
8.3
10
5.8
9.0
9.9
IO.O
during the practice such spots multiply, and better records
can be made with their help at the end of the period of
practice than at the beginning. None of the students of the
1911 group realized that he was making much use of these
marks but the result shows that this group made a larger ap-
parent improvement than the latter group who were deprived
of this help by being required to select the cards from another
set. In 1913 this source of error was eliminated. The subject
scattered the cards of a second set and selected from among
them according to the directions which have been reproduced
above.
The experiment of 1914 was designed to eliminate anothe
possible source of error. It was felt by many of the worker
that the familiarity with the cards which was acquired in the
course of arranging them in order of brightness might be
responsible for a large amount of the apparent improvement
of memory. Accordingly the experiment was so arranged as
to reduce the amount of this factor to a minimum. On the
first day the procedure was the same as in the preceding season
in order that the first test of memory might be made under
exactly the same conditions, but after the first day there was
no arranging of the cards in order by degrees of brightness.
IDENTIFYING SHADES OF GRAY 535
Instead, the cards were arranged by the subject according to
number with their faces down (they were numbered according
to the 'most probable order' described above), then the whole
set was turned over and the four cards for the memory test
were selected. The worker then preceded to scatter another
set and pick out the cards which seemed to be the same, just
as in the experiment of the year before. The experiment was
identical with the one of the year before in every respect
except that the workers were not familiarized with the material
through practice in arranging it. Precisely the same op-
portunities were afforded in 1914 as in 1913 for the employ-
ment of special devices, illegitimate as well as legitimate, as
helps in choosing the shades.1
The scoring in 1914 was on the basis of the average of
previous determinations as shown in Table I. under the head-
ing 'most probable position.' In order to make a comparison
possible between these results and those of 1913 the latter
were rescored on the same basis; that is to say, the error was
stated in both cases as the difference in position between the
card sought and the card found if they were all arranged in
the 'most probable' order. The results obtained for the 1913
experiment according to this method of scoring are not es-
sentially different from those obtained when the scoring was
done on the basis of the worker's own arrangement for the
day, and are only a little more regular.
The improvement of recognition through practice is no
less rapid in 1914, without the additional familiarity which
comes from arranging the cards in order, than in 1913 when
this factor was fully operative. With this assurance that the
factor of familiarity is of no great importance, and with the
evidence that the results can be scored satisfactorily on the
basis of the learner's own daily arrangements of the cards
without undertaking the laborious process of measuring the
actual brightness of the different ones, this experiment gives
a simple and easily workable demonstration of the improve-
ment of a form of memory through practice.
1 The most important of these devices is the estimation of the distance from the
end of the scale to the lightest or darkest of the four cards to be memorized. It is
526 WARNER BROWN
comparatively easy to find these cards by simply counting up mentally, as the cards
lie scattered at random on the table, from the white or black extreme to the eighth or
whatever one is desifed. The other two cards are frequently found by grading in
between the two extremes. Only half as many errors were made on the lightest shades
as on the middle ones and fewer errors occurred on the darkest shades than on the
middle ones. There was no general tendency to under- or over-estimate the shades of
gray in the memory test, but too light a shade was selected for the lightest and too dark
a shade for the darkest, and the same tendency appeared with the other two shades, as
may be seen from the following tabulation of the average number of errors per card.
'Too Light Too Dark
Lightest card 1.08 0.77
Second card 1.90 1.57
Third card 1.72 l.8o
Darkest card 1.17 1.52
Average all four cards 147 1.42
BF
1
P7
v.22
Psychological review
PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET
UNIVERSITY OF TORONTO LIBRARY