Full text of "The "standard-alphabet" problem, or, The preliminary subject of a general phonic system : considered on the basis of some important facts in the Sechwana language of South Africa, and in reference to the views of Professors Lepsius, ..."

i I

^Mmmn..!^ \

C? E

^ ^,-OF-CALIFOfi^ ^\\EU

I |^^\| |y

r^I >^4 < * l

3 S

PS ="

I i

^\\E-UNIVER%

^" "%

I s

I 1

.1 $.

. .\Ml Ul

I fe

^ I

e= 55

1 I

s a

1 ?

// -f

THE

"STANDARD-ALPHABET"
PROBLEM:

OR THB PRKL1MINABY SUBJECT OF A

GENERAL PHONIC SYSTEM,

CONSIDERED ON THB

BASIS OP SOME IMPORTANT FACTS IN THE SECHWANA LANGUAGE OF SOUTH AFRICA,

AND IN REFERENCE TO THE VIEWS OF PROFESSORS LEPSIUS,

MAX MULLER, AND OTHERS.

A CONTRIBUTION TO PHONETIC PHILOLOGY,

ROBERT MOFFAT, JUN.,

SURVEYOR,
Fellow of the Soyal Geographical Society.

" The loss of the living traditional pro- " But the linguistic scholar will prefer to

nunciation implies a loss of much more than follow the written system fixed by literature,

what we generally call pronunciation." and to neglect the varying deviations and

Sunen. shades of modern pronunciation." Lepsiui.

fttfalisjrers.

LONDON TRUBNEB & Co., PATERNOSTER Row.

, ( J. C. JUT A, CAPE TOWN.

a AFRICA. ^ Q BROWNE> NATAL

1864.
The right of Translation is reserved.

Unavoidable delays, contingent on the work being edited in a
foreign land, have delayed its publication. The concluding sheets
have been revised by the Rev. J. FRfenoux, of Motito, S.A.

THE PRINTER.
October, 1864.

GEOHGE CNWIN, GBESHAM STEAM PBE88. BUCKLERSBURY, LONDON.

Stack
Annex

735"

TO HIS EXCELLENCY

SIR GEORGE GREY, K.C.B.,

A SCHOLAR IN AFRICAN PHILOLOGY, AND AT THE SAME TIME ITS
DISTINGUISHED PROMOTER,

THIS HUMBLE ATTEMPT TO ARRIVE AT SOME OF THE FUNDAMENTAL

PRINCIPLES OF HUMAN SPEECH,
UPON THE BASTS OF A NEW ORDER OF FACTS,

THE RESULT OF PERSONAL RESEARCHES INTO THE . LANGUAGE OF A REMOTE

PORTION OF THOSE INDIGENOUS SOUTH AFRICAN TRIBES WHICH

HAVE LATELY BENEFITED BY HIS EXCELLENCY'S RULE,

is (BY PERMISSION)

MOST RESPECTFULLY DEDICATED, BY HIS OBLIGED SERVANT,

THE AUTHOR.

CONTENTS.

INTRODUCTORY REMARKS,

CHAPTER I.

PRINCIPLES OF CONSONANTAL CLASSIFICATIONS 1

I. Of Classifications at present in use (Ancient and Modern) .... 1
II. General Principles of a Classification suggested by the Mutation

of Consonants in the Sechwana Language 14

CHAPTER II.

ANALYSIS OF SECHWANA CONSONANTS 19

I. The Guttural Series 20

Parenthetical Section [Confusion in Nomenclature and Ortho-
graphy in the usual Classification of the Gutturals] 23

II. The Lingual Series 40

III. The Labial Series 51

CHAPTER IK.
ANALYSIS OF~OTHER CONSONANTS EXAMINATION INTO THE POWERS OF

THE REMAINING LETTERS IN THE GENERAL ALPHABET OF DR.

LEPSIUS THE CLASSIFICATION OF SUCH AS ARE 'REALLY^ ELEMENTS,
AND OF OTHERS THAT MAY BE SUGGESTED BYJTHEM, UPON PRINCIPLES

RESULTING FROM ANALYSIS IN PRECEDING CHAPTER 57

I. Faucales (of Lepsius) 58

(a) Nature of the Spiritus Asper 69

(b) Are there P.inary or Quantitive forms of the Spiritus ? . . . . 75
Cc) The Spiritus influenced by the long quantity or Syllabic Accent 78

(d) Vocalization oftlte Spiritus, and of some Consonants .. .. 80

(e) The Nasals and their Vocalization 90

PAGE

II. The Palatales (of Lepsius) and the Letter q 99

The Oriental q probably the elementary form of the aspirated

consonant 'g noticed under Classification of Sechtoana Gutturals 101

ILL Cerebrates Indicffi (of Lepsius) " 106

IV. Linguales Arabicse (of Lepsius) 112

V. Dentales (of Lepsius) 121

Inquiry into the nature of the letters s and z, th (in thin), and th

(in this), and the probable existence of their elementary forms .. 121

VI. Labiales (of Lepsius) 129

CHAPTER IV.

STJMMABY OF PEECEDING ANALYSES 131

1- Compendious View of the Simple Consonants, as suggested by the

Phonology of the Sechtoana Language 131

IT. On the Application of the Nomenclature and Orthography, above

suggested, to systems of permutation in other Languages . . . . 144

CHAPTER V.

COMBINATIONS OF Simple Consonants 167

(a} Latham's " Law of Accommodation," how far correct ?.. .. 168

C b) Consonantal Diphthongs 174

principles arrived at in this work

[_No MSS.have been found for this."]

INTRODUCTORY REMARKS.

IT is in the nature of man and his prerogative to generalize,
however limited the range of his observation ; and in proportion
as this is extended, is he enabled the more or less confidently
to demonstrate the truth of his knowledge, or the degree of
credibility of what he anticipates. Therefore, I presume that
fresh statements, whether of facts or inferences, will be welcome
from any individual in a new field of research, however obscure
or diffident he may be ; especially in this golden age of inductive
science, when particulars, and instances, and data, in every de-
partment, are being scrambled after by all classes of students.

I have been prompted by such considerations as the above to
publish, for the information of others devoted to the study of
language, the results of my observations during a few years of
assiduous research in an elementary branch of the subject. This
I had long selected for my own amusement and instruction in
leisure moments of a professional vocation ; and having recently
been engaged in an active trading life on a wide frontier, among
native tribes, the destruction of whose language is as inevitable
as their speedy social dissolution, I have enjoyed unusual facilities
for the prosecution of my object.

While the quotations, which confront each other on the
title-page of this treatise, present in a concise form the opinions
of two of the first continental scholars on the same subject, they
will also serve to convey an idea of the nature of the task I have

now undertaken. Where such absolute difference of opinion
exists between men possessing stores of learning, there must
rather be a deviation in the researches of either of them, than a
deficiency in his materials. The one, in a special case, proposes
to explain an ancient rock-engraven literature by means of an
investigation into the actual relations of the material elements
of human speech, as " transcribed * * * from the lips" of
those " among whom it has been traditionally preserved ;" * the
other attempts to establish an " absolute rule" of phonetics on
the historical relations of those elements, as represented to the
eye by letters in various existing literatures. I need not proceed
to argue the question as to which is suggesting the proper path
of research for the collection of data, in order to arrive at the
laws of " a natural science," as that of phonetic philology un-
doubtedly is ; and whether the rudiments of this science are to be
conveyed to the mind by the artificial means of letters to the
eye, or rather the more legitimate one of sounds to the ear.

Dr. Lepsius, by means of an immense command of ancient and
modern graphical materials, of both dead and living languages,
and taking the Indian grammarians as a guide, has arrived at
the construction of a " universal linguistic alphabet ;" but truly
elaborate as it is, and however convenient it may be for students
whose attention is confined to the historical forms of the Indo-
European tongues, there is decidedly something of an artificial nature
about it, which must necessarily be discordant with the views of
others who, in confining their investigations to the "living
traditional pronunciation" of primitive tribes like those of South
Africa, are led to arrive at conclusions of a more* demonstrative
character. The quotation to which his name is attached sufficiently
explains the basis of his system, and the difficulty of his labours.

* Dr. Lepsius's researches (1835) into the relation between the Egyptian
and Coptic, after all, I believe, only extended to the comparison of litera-
tures viz., the hieroglyphics and the liturgy.

The following treatise is the result of a mode of research
suggested by the pointed remark of the late Baron de Bunsen, con-
tained in the other quotation. It is a survey of the elements of
articulation as they occur in the crude and simple speech of a
barbarous people, and the principles which enter into their various
mutations and combinations ; which, in an order analogous to
that usually pursued in all natural science, must precede a con-
sideration of the more complicated or syntactical stages of the
material forms of human language. It maintains at the outset,
that " we can understand the historical forms of speech only by
watching and comprehending the process of utterance as it goes
on even now in the individual speaker ;* but rather by observing
the effects of the process on the ear, and in the various permu-
tations of the elements of articulation, than the cause in the
physiology of the voice which pre-supposes the anatomy of the
organs. It, in fact, claims for the humble Sechwana language,
spoken by numerous degraded tribes on the south-eastern borders
of the South African desert (or Khalagare wilderness), and
why not for other unwritten tongues ? that perfection of
phonic purity which Professor Max Mliller concedes only to
the Vaidik Sanskrit, whose historical orthography is more
than two thousand years old ; and for the very same reasons,
viz., that it (can be) " studied by means of oral tradition only,
and .... in the absence of a written alphabet, the most minute
differences of pronunciation (have) to be watched by the ear,"
and " it (has) suffered less from the influence of phonetic corrup-
tion than any tongue from which we can derive our observa-
tions.''! While it does not deny that the phonic forms of the
ancient vernacular Sanskrit were perfect, whatever corruptions
may since have crept into their phonetic representations, it holds
that those of the Sechwana, and some other barbaric dialects, are

* The Saturday Review, Jane 29, 1861, p. 673.

f Proposals for a Missionary Alphabet, &c.,\)y Max AJ tiller, M.A., p xxii.

perfect, and still accessible to the inductive philologer, but at an
immeasurably further advanced stage of inductive science. It
therefore deals only with natural facts, discovered in situ by per-
sonal observation.

As thus treated, the subject may be found to have some new
phases when viewed by men of learning, for whose consideration
I would with becoming reserve and humility submit my views.
It is just possible that a few of the facts, and such conclusions
as have been arrived at, or to which they may come, may
account for various phenomena in the accidence of those written
tongues, in which the development of the essential form has
greatly modified the accidental structure. Though I have
availed myself of the common prerogative of generalizing, and
even speculating to the utmost of my ability, and presume to
state such a possibility, I have given every particular that has
fallen under my notice, to enable the reader to arrive at his own
conclusions, by the aid of such accessory knowledge as he
doubtless possesses beyond me.

In craving the candid attention of the reader, and lest the
title of this treatise should lead him to expect more than it con-
tains, I would beforehand state

a. That the object of the work is not to prescribe a new
system, but rather to contribute to the construction of one,* or to
illustrate the science of universal phonics by the collection and
arrangement of instances from the phonology of the Sechwana
language of South Africa. It is an attempt to treat inductively
on a subject, which, by being usually based on the physiology of
the human voice, has hitherto only been examined deductively ;
nevertheless, reserving to myself the privilege of employing the

* On reference to the tables of consonants in Chapters IV. and V., the
letters in bold type will show the reader the extent to which the classification
is based on data furnished by the language; those in italics", the mode in
which I have attempted to complete it by a train of speculation suggested
by them.

latter speculative mode of analysis, besides so much of classi-
fication as the extent of my train of facts will have admitted.
Without -venturing to assert that, by a consideration of the
elements of articulation of any one spoken language, a correct
system of phonics may be framed applicable to all, it will satisfy
me to intimate that at least an imperfect, and not-incorrect or frag-
mentary, system may thus be framed ; and that the consideration
of other languages, containing additional elements, would, by in-
creasing the number of instances, contribute to a more copious
induction. I have, therefore, proceeded upon the principle that
it is absolutely necessary, for the purposes of this elementary
branch of the science, to arrive at fixed results in one pure and
living dialect, before advancing to a comparative view of different
tongues. This will, I trust, obviate, in my case, a common
objection urged against writers whose observations are confined
to a single language.

b. However I may feel the want of some fixed mode of re-
gistering my researches, as a saving of both time and labour, a
graphic scheme is entirely secondary to my immediate object.
Anything of the kind, even though it may emanate from a
master-mind in philology, must be regarded as immature, till
suggested by a phonic system resulting from an inductive survey,
similar to, but of a far more comprehensive nature than, that I have
attempted. It is in vain to expect it so long as men seek to arrive
at the nature of vowels and consonants exclusively, either by
experiments on the action of the vocal organs, or by artificial
contrivances to imitate them, or by the comparison of existing
historical alphabets.

To the self-experimenting phono-physiologist in the one case,
the facetious advice of Professor De Morgan to the meta-
physical student is particularly applicable, viz. : " I would not
dissuade a student from * * * inquiry ; on the contrary, I would
rather endeavour to promote the desire of entering upon such
b2

subjects : but I would warn him, when he tries to look down his
own throat with a candle in his hand, to take care that he does
not set his head on fire."* In the second case, it happens that
though the ingenious contrivers of speaking machines "have
succeeded in imitating a great part of the sounds used in speech,"
they confess that " every simple and independent sound and
consonant requires a special apparatus ;" they must, consequently,
admit that the production of the unique combination and operation
of the different apparatus would be a task about as hopeless as that
of any optician who would attempt to devise a means of imitating
the peculiar structural arrangement of the eye, whereby the
automatic alteration in the curvature of the crystalline lens adjusts
it to different ranges of vision ; in fact, that what the telescope with
its sliding focal adjustments (spite of its comparative perfection)
is to the wonderful structure of the visual organ, so are " reed
tubes" and " vibrating tongues" to the complicated organism
required in the perfect enunciation and articulation of the most
simple elements of voice. In the third case, it may be urged that
the fact of the English, or any other historical alplwbet, only indi-
cating a limited number of elements, does not necessarily imply that
the language does not contain several additional elements, though
these may be represented by irregular combinations of letters. In
the transliteration of such, " where," to use the words of Professor
Mliller, " for reasons best known to the archaeologist, one sign
may represent different sounds, and one sound be expressed by
different signs, new and entirely distinct questions are involved,
and capable of solution by archaeological and philological research
alone. "-f Hitherto the advocate of historical orthography has as
little to show as either the physiological observer or the artificial
experimenter, in any attempt to establish a natural classification
of the elements of speech. The one mode of research has only

* " Formal Logic," p. 27, Note. f " Proposals," &c., p. 20.

shed a few rays of light on the other, and phonology, instead of
being already " reduced to its last analyses," as a learned American
writer remarks, is, in his own words, as echoed from Sir Robert
Taylor's Institution, Ox/ord, -verily and without equivocation,
"exactly the same that Sanskrit grammarians more than two
thousand years ago defined its elements to be in their own
primeval tongue ;"* but not what it may be, if the modern
philologer will base his inferences on facts of a proper descrip-
tion.f

In corroboration, it may be alleged that the whole system of
phonetic philology, as at present based on ancient and existing
alphabets, and physiological classifications, is but a labyrinth of
graphic schemes. To the various powers given to the Roman letters
in different European alphabets, and various letters representing
the same power, there may be added (1) both letters and
sounds introduced by travellers and navigators of different
nations, in lists of words collected irregularly and carelessly
from uncivilized tribes ; (2) the more complete alphabets of

* Bibliotheea Sacra, Oct., 1859, p. 673. (See also Proposals, &c., by
Max Miiller, p. 22.)

f- The following, from the pen of no less distinguished a philologer than
the late Rev. Richard Garnett, of the British Museum, and which has
occurred to me since the above quotation was written, approaches more to
the sober and truthful:

' It is presumed that enough has been advanced to show that the scale
of permutations in the Indo-European languages, as laid down by Grimm
and Pott, will admit of being considerably extended beyond the limits
which they have assigned; and that it is very unsafe to fix upon Sanscrit,
or any other known language, as a model to which all others are to be
referred. It is believed that there are numerous phenomena in language
of which neither Sanscrit, Greek, Teutonic, nor all in conjunction, can
furnish a satisfactory solution ; and that the real original articulations of
speech have in many cases yet to be ascertained. This can only be
attempted by a copious induction of all known varieties of cognate forms,
and all that we can rationally expect to achieve is an imperfect approxima-
tion to the truth." Philological Essays, p. 254.

missionaries and priests labouring under the same disadvantages
of a diversity of plan ; (3) the host of cumbrous alphabets of
Oriental languages, dead and living; and (4) the numerous
"transliterated" forms of these alphabets introduced by Oriental
scholars of different schools, each according to a " method of
notation peculiar to himself," not to mention the orthoepical
schemes of many authors ; so that it may truly be said,
the operations of the linguist are trammelled by his own
materials.* This complexity of his phonetic materials has been
especially increased since the efforts of Sir William Jones, in
1788, and of Count Volney, in 1795. Missionaries labouring
among conterminous tribes have often made attempts to arrive
at uniformity, and philologers have as frequently seen the
absolute necessity of a universal alphabet for the analytical
purposes of their science; but the general public, with Isaac
Pitman before them, are too prone to suspect every innovation
as only a preliminary to the practical " abrogation" of the ancient
forms of literature. Therefore, such characters as I have made
use of, or even any suggestions on modes of writing the elements
of articulation, must be regarded as arbitrary, though, as much
as possible, in keeping with the Roman graphic system ; at all
events, most of my remarks upon them are confined to the foot-
notes. It will, nevertheless, be seen that, by working on a new
basis, I have attempted to provoke a little discussion on a subject
which every student of unwritten tongues must be most anxious
to see satisfactorily settled ; in order that the constant trouble of
making myriads of alterations may be dispensed with in the col-
lection or publication of orthographical data.

c. Much less do I presume to enter into the controversy on
the subject of " Romanizing" existing ancient and cumbrous

* The numerous comparative tables, occurring in this work, of letters
intended by different authors to represent the same series of sounds, will
alone show the confusion of alphabetic systems.

alphabets, in which some of the first Oriental scholars have been
engaged.* Independently of all the arguments which it is pos-
sible to allege for or against such an innovation on Oriental
graphic systems, it must be confessed that it would amply repay
the labour of any man with the necessary ability, and possessed
with the " phonetic crotchet," as it has been called, to reduce the
" twenty different vernacular tongues" of India, having now "fifteen
various alphabets," to the same phonetic system, based upon iden-
tical phonic principles, without consul ting one volume of philosophy,
poetry, or theology. He would verily be examining them in situ,
and classifying them by a most rigid analytic formula ; while, by
giving his attention to Pali, Sanscrit, and Arabic, he would have
to resort to the " books" of the Buddist, the Hindu, and the
Mussulman which ancient literary repositories, in point of value,
bear the same relation to the living dialects, that a few drawings
of fossils in a museum would have to originals still im-
bedded in the rock. It cannot, then, be denied that a body of
men, by a division of labour, and acting upon preconcerted views
of a phonetic system, would arrive at still more comprehensive
results, which no existing literatures could ever afford. In the
same manner as the learned Bunsen showed that a knowledge
of the traditional Koptic, gathered from the priests, would be
necessary to enable the Egyptologer to decipher the illegible
groups of hieroglyphics, so it would be easy to prove that a
still surviving colloquial dialect would shed light on the most
ancient sacred literature ; for " all sacred language is * *
essentially, nothing but an earlier stage of the popular dialect,

* For: Sir William Jones, Volney, Gilchrist, Monier Williams, Sir C.
Trevelyan, H. T. Prinsep, Dr. Yates, Dr. Duff, Dr. Caldwell, Max Miiller,
Lepsius, the " Times," and others.

Against : J. Prinsep, J. Tytler, Dr. Jarrett, H. H. Wilson, Mohl, and others
See Evangelical Christendom, May, 1860, p. 237.

preserved by means of the sacred books,"* only in an imperfect
and fragmentary, however correct, form.

d. Nor do I attempt to expatiate on the necessity of intro-
ducing a uniform phonetic system for the practical purposes of
the missionary. If we are to credit the opinions of many writers
of the present day, as to the qualifications of a missionary for
example, in India then he requires the preparation of a savan.
He has not only to master the root language of the people among
whom he is labouring, but also that from which its theology has

* " Egypt's Place in Universal History." Bunsen, Vol. I., p. 258.

Indeed, if we would wish to form an idea of the objective value of a
sacred literature, in a philological point of view, and compared with the
traditional language of the people among whom a new religion has been
introduced, we have only to examine any elaborate modern translations
of the Bible into unwritten tongues, such as the Sechwana, for instance,
which will bear comparison with many. The missionary acknowledges
that, in order to maintain the tenor of Scripture, he is compelled in a
measure to mar the colloquial idiom in a few cases; ex gr., to indulge in
circumlocution, where a curt phrase would convey the identical meaning
more elegantly. It cannot be otherwise, as, in the colloquial idiom, there
is so involved a reference to the instinctive customs, habits, and modes of
thought of the barbarian, that it would often be inappropriate, and in
many cases unnecessary, to introduce it; for, excepting the book of Job,
with its host of allusions to the scenes of the outer world, the subject of
all sacred books is generally too confined to allow of the introduction of
more than & fourth part of the spoken words in a comprehensive language.

Bearing this in mind, and the fact that a sacred literature is a new subject
in the language of an ancient or barbarous people, it is not surprising that
we hear of slurs cast by scholars in India on missionary vernacular, and in
this country on school Kaffir. In South Africa, I have heard an intelligent
missionary, after one year's application to the " book," or otherwise " sacred,"
Sechwana, convey the simple truths of Scripture to the natives in the most
intelligible strain, becaiise, during the acquisition of the language, his
phraseology has been confine/l, almost exclusively, to that new subject ; while,
again, I have heard some, comparatively uneducated, of a few years'
standing, so attached to this book or sacred idiom, without at the same time
making themselves daily more acquainted with the common idiom for
colloquial purposes, that their garrulous repetition of it on all subjects was
absolutely disgusting. It is sometimes as well that a translator is as much
at home in the one as in the other.

XVII , -.

been derived, and, of course, their respective alphabets. Of the
six systems of Indian philosophy, the Nyaya, with its excess of
logic ; the Sankhya, with its excess of metaphysics ; and the
Mimansa, with its excess of theology,* all claim his versatile
attention.! To accomplish his object, he must study those tongues
thoroughly, for a smattering is often no better than, absolute
ignorance. But, in order to understand the true end of his
exertions, it is as well to compare his calling and sphere with
those of the philologer.

The vocation of a missionary, who provides an unwritten
language with an alphabet, or finds one imperfectly prepared, is
entirely independent and peculiar ; his aim is the communication
of Divine truth to the instinctive ignorance of heathens; to
substitute for the vague impressions which exist on their
minds of impersonal " rude powers," or subtle deities, the re-
vealed fact of a personal God, or supreme moral Governor for a
superstitious adherence to any human method of expiation, a
living faith in His incarnation and redemptive act for a servile
willingness to appease aught that will quiet conscience, the spirit
of penitence and reformation for the fear of death, the hope of
future existence. Whatever the medium, he seeks to address
man, " be it an A'rya or a Sudra." His teachings have to be
conducted, not so much by an " argumentative exposition" of
his doctrines, and an "elaborate confutation" of those of his
opponents, as " in the form of a testimony * * with respect
to the mode of exhibiting it, though not in the spirit of the
teacher * * dogmatic. r \ He has to declare " intrinsic

* " Indian Logic," by Max Miiller. Appended to " Laws of Thought, by
W. Thomson, D.D.," p. 363.

f " Moreover, as is well known, the peculiar philosophical notions of the
learned Hindus must be understood by those who would effectively evan-
gelise that race." Evangelical Christendom, May, 1860, p. 244. See Ibid.,
Oct., I860, p. 524.

| Works of Rev. Robt. Hall, by Olinthus Gregory, LL.D., Ac., Vol. I., p. 302.

xviu

primordial truths," without the aid of a " syllogism, or quo-
tation" * of human wisdom, and by the mere external means of a
new vernacular. Therefore, in many cases he considers it
immaterial to him what orthography he may use,f as his practical
operations differ from those of the philologer, of whose science
he applies such a knowledge as he happens to have acquired to
suit his own peculiar plans. Moreover, as to his sphere of
duty, what is true of human nature in South Africa, where
the most rigid statist would be compelled to admit it, will be
true of man in all lands viz., that it is only amongst poor, or
isolated, or dismembered communities, that the missionary has
been most successful. One would think, therefore, that as he
has more immediately to explode popular notions, and not systems
of philosophy ; to dispel the superstitious polytheism of the
" unthinking multitude," rather than the atheism and pan-
theism of philosophic sects : in fact, to deal with the vulgar
rather than with the learned he would endeavour to command
a knowledge of the popular dialect, independently of existing
ancient vernacular literatures, in which " the spelling of words is
no longer phonetic but traditional." Indeed, if the fact admitted by
the "Friend of India" and cited by Sir C. E. Trevelyan,f may be
regarded as an approximate estimate of the proportion of the
educated classes to the ignorant masses in India viz., " that only
one million out of the thirty millions of Bengal can read" their
indigenous literature, the missionary can have no doubt as to
which ought to engage his attention, and the mode of writing
speech most likely to facilitate his labours. The amount

* Vinet.

f " But as it is so immaterial how the language is written, and the only
essential point being that the Word of God may be taught in it, I entirely
waive the question as to the mode of writing, &c." Rev. H. G. Knudsen,
R,M.S., in Carres. S. A. A. B. Society, p. 5.

t Papers originally published at Calcutta in 1834 & 1836, on the application
of the Roman Letters to the Language of Asia. London, Longman, 1854.

XIX

of success which has attended the labours of missionaries in
South Africa during the last half century, in the introduction
of a native literature,, is only a proof that as much could be
accomplished among the illiterate of every nation by any who
would choose the same course ; for it must be patent to all, that
what is communicable to the vulgar must be intelligible to the
learned.

The aim of the philologer is far otherwise. While, on the
one hand, he has solved some most interesting ethnologic pro-
blems, such as the identification of the radical language of the
" rude Kelt" of the corners of Britain, with that of the "effemi-
nate Bengali'' of the Indian promontory; and, more' recently,
the stock of the Hottentot of the Southern extremity of
Africa, with that of the ancient Egyptian of the extreme
North and discerns in the future still greater triumphs ;
on the other hand, he feels assured that if " truth consists in the
conformity of the names by which the representations of the mind
are expressed to the representations themselves," and if " lan-
guage is the only external condition on which philosophy is
dependent,"* he has the whole range of mental science at his feet.
To effect these objects, the analysis of language, phonetic as well
as grammatic, is his great power ; the former, by the comparison
of the material forms of human thought, as are to be found in
the various sets of combinations of the elements of articulation,
and in the collocations of words, which distinguish different
tongues ; the latter, by an inquiry into the essential form, or
the law of the process, by which the varieties of names and
syntactical constructions in different languages are but different
expressions for similar cognitions in all.f His sphere of re-

* Sir William Hamilton's Lectures, Vol. I., p. 382.

f This will perhaps be the fittest place for me to append, in explanation
of my meaning here, the following remarks, which I have extracted (slightly
modified) from a letter addressed in February, 1857, to a distinguished
individual greatly interested in the study of aboriginal tongues :

search alone is analogous to that of the missionary's duty. As
with the missionary, the more unsophisticated and humble the
people, the more successful he is likely to be ; so with the
philologer, the more simple and primitive the language in which
he works, the more correct are his conclusions likely to be in
all cases ; for in it the various forms and combinations of either
sound or meaning* are more easily discernible, and resolvable

" It has often appeared to me that many of the so much misunderstood
elementary principles of universal grammar, which are the subject of great
diversity of opinion among scientific men in Europe, may be explained and
proved, and in some instances discovered, by a clear investigation of such
hitherto unwritten tongues, when carefully reduced to writing ; e.g.,

" Mr. Home Tooke's idea of prepositions and conjunctions is, that they
do not form distinct classes of words, but are merely abbreviations of nouns
and verbs." Eneye. Brit., Vol. X., p. 673.

Page 657. " It has been proved, by such evidence as leaves no room for
doubt, that if, though called a conjunction, is in fact a verb in the im.
perative mood, of the same import with give, so that we may substitute the
one for the other without in the smallest degree altering the sense."

The identical word holds good in the Sechwana, in the sentence ha Tti
bobola nka shwa (If I sicken, I may die) ; ha, the equivalent of our con-
junction if, is nothing else than the verb h.a (give), which is in a measure
obsolete in the language, and is usually employed in asking a gift, or at
meals, as mo ha (help, or give him), naea being the more common word
on other occasions.

AGAIN "From (the preposition) merely means beginning, and nothing

else." "As from always denotes beginning, so to and till always

denote the end. There is, however, this difference between them, that to
denotes the end of any thing ; till, only the end of time." Ibid., p. 681.

Now, in the Sechwana sentence, ki le lea ea go CWa (or go simolola ha)
Kuruman go ea Khatwe, the words go cwa and go ea, which are nothing
else than prepositions in their primitive forms, respectively mean to come
out (or to begin tvith), and to go to. Go tsamaea (to be in the act of going),
which is used for till or until, implies time. These are coincidences
showing that, however the accidental forms of language may vary, the
essential form or meaning is the same.

* In illustration I add the following from my note-book, with an example
from the primitive dialect of the Sechwana :

Nothing is more common in writings on mental science than a reference
to the connexion between thought and language; but this seldom exceeds
' half belief and feeble assertion.' The remark of Professor de Morgan,

XXI

into their elements. I have somewhere met with the remark,
that " the concerns of barbarians, unconnected and remote from
all contact with literature and civilization, and destitute of all

" I doubt whether we could have made thought itself the subject of thought
without language," (Formal Logic, p. 34) is, however, more bold and sug-
gestive ; and he surely speaks here of language in its essential forni. But
though we are told by another able author that logic is a science of " the
structural laws according to which man thinks," and by the above profound
logician that " logical truth depends upon the structure of the sentence," I
doubt whether either of them would concede that logic " can but result as a
generalization * * * from an inductive survey" of the science of
universal grammar, i.e., language in its essential form. To any inclined to
this opinion, as well as to those who maintain that language "is not
essential to thought," and go so far as to say " it must not be supposed
that an examination of the rules of language would answer every purpose of
a logical system," the undermentioned facts cannot be devoid of interest.
For it is to be expected that parties holding such opposite views must be at
variance as to whether or not the modality of a judgment belongs to the
copula in other words, whether "logic can take cognizance of the pro-
bability of any given matter;" that is, more plainly speaking, to say,
whether the expressions will be and may be, and their negatives, are not also
simple forms of the copula as much as is or is not.

The Sechivana auxiliary verb, deprived of a host of accessory particles,
may be simply classified under two tenses and three moods (proper).
Moods. Tenses.

Past. Present.

1. INDICATIVE. Jci le Jca tsamaea. Jci ea tsamaea.

I did go (=1 went.) I am going.

2. p Jci le M tla tsamaea. Jei tla tsamaea.

I (did) shall go. I shall go.

3. POTENTIAL. Jci le Jci Jca tsamaea. iiJca tsamaea.

I (did) may go. I may go.

It is evident, from the above, that there is no such thing as a "future
tense ;" but rather, besides what are popularly called the indicative and
potential moods, another mood, of which this " future tense" has all along
been the erroneous representative, and which might, for common gram-
matical purposes, be called the conjectural mood. There is very little doubt,
in my own mind, that these three moods express " modality," or degrees of
knowledge, at times past and present, viz., certainty, probability, and pos-
sibility ; but, as my attention is for the present confined to phonology, I can-
not here enter more fully into the subject. The fact will also, no doubt, be

xxii

historical records, will scarcely be thought to require any great
portion of attention from a philosophical inquirer." True
judging from the fate of such nations and tribes as have in many
countries been discovered by the traveller and navigator, and
which have dwindled before the approach of the colonist, very
much in the same manner as the herds of elephants, ostriches,
and antelopes, before the repeated sallies of relentless hunters
such a remark carries with it a phase of plausibility ; but it is a
false conclusion, based upon the assumption that, inasmuch as a
barbarian is a degraded being, everything pertaining to him
must be correspondingly liable to depreciation. Civilization has
advanced to such a consummation, through all its stages of
improvement, from writing and printing, to the electric, and
more recently the printing telegraph, that distance and time alike
have become annihilated in human intercourse, and enlightened
man has forgotten his ancient position, wherein, without the
means of constructing the symbols of his thoughts, he spoke in
native fluency and simplicity the language he now transmits,
in all its complexity, with the speed of lightning, across con-
tinents and seas. Each science and art, in the historical order
of its occurrence to the human mind, and its application to the
supply of human necessities, has developed and modified, and in
fact, destroyed the original language to such an extent and his
facilities of intercommunion have increased so amazingly by
means of the various modes of expressing thought that he is apt
to magnify the nature of the idiom he speaks, forgetting that it

interesting to those writers who have lately been speculating on the nature
of the auxiliary verbs shall and will, as they occur in English. " Whether
Aristotle's rudiments of logic have not antecedent rudiments which time
may yet bring to light is a somewhat unsettled problem in speculation."
(See Ferriers "Institutes of Metaphysic," p. 14.) Where, but in the
principles of universal grammar, are such rudiments to be found? and
whether more likely in the primitive dialect of the barbarian, or the complex
idioms of the civilized ?

XX111

is only the measure of its power that has been enhanced by the
means of its communication ; that its development and complexity
are the result of the development of mind, and the accession of
new kinds of knowledge ; that, whether man avail himself of
electricity or printing, or his organs of speech, he cannot surpass
the rapidity of his thoughts, and in this respect the degraded
Bushman can vie with the philosopher. When, therefore, we
are told that Home Tooke conceived an "original thought"
regarding the significance of the particles in his native tongue,
and, though " ignorant of the characters even of the Anglo-Saxon
and Gothic languages," acquired their crude and barbarous forms,
" to ascertain whether he had made a discovery," and this most
successfully ; and again, that Jacob Grimm was indebted for his
" law" to his researches into the ancient forms of the German
language, it seems surprising that students in philology have not,
ere this, inferred from such facts the probability that, in the
simplicity of the barbarous and unwritten tongues of newly
discovered regions, are to be found most of the data necessary
for the solution of some of the difficult problems in their science.*
It seems not unreasonable for us to expect that, just as men are
indebted for the greater part of what they know of metaphysical

* I am enabled to fortify these remarks by the following quotations from
able authorities : " The language of tribes who roam wild in a condition
of savage life, is necessarily simple and primitive. So long as they continue
separate and distinct from a civilized race, it is marked by the genuine
impress of nature ; but as soon as they mix with nations more refined than
themselves, in proportion as they gain morally or mentally by the inter-
course, it is observable that, in the same degree, the parent language
becomes vitiated or changed. Modifications and inflexions, unsanctioned
additions, tralatitions, and neologisms, like parasitical plants adhering to an
ancient and venerable stock, are then first observed disfiguring the natural
root; and, as the genius of modern literature has become disdainful of
indigenous compounds, a kind of hybridous vocabulary takes the place of
the old tongue." Edinburgh Review, April, 1844, p. 455.

" It is in.the ruder languages that the important phenomena of develop-
ment and growth the laws of language are best studied." R. O. L.,
Encyc. Britann., Sth Ed., Vol. XI11. p. 195.

science to so much of what the ancients knew as was preserved
in the monasteries of the middle ages, so they will have to look
to the perspicuous structure of these primitive languages, which
have hitherto remained concealed in the dark recesses and shades
of history, for any principles elucidative of the laws of speech
and thought.

But the condition of harbarous man is ephemeral ! Guided by
the knowledge of the past, we can only rest our hopes or fears of
him on sheer probabilities. That his normal constitution will be
annulled, and his social system dismembered, and that this will
result in his gradually disappearing from the dismal scenes of his
degradation, the whole tenor of modern history only convinces
us. This ephemeral nature renders him the more interesting to
both the missionary and the philologer. The former redoubles
his efforts in scattering the seeds of knowledge, and raising the
bright cloud in the path of the benighted to futurity ; the latter,
in grasping at each new appearance of truth, is rendered impatient
by the conviction that " every day destroys a fact, a relation, or
an inference." It is not to be wondered, then, that the missionary
occasionally anticipates the duties of the philologer, and that the
philologer sometimes borrows his materials from the missionary.
It appears impossible that all that would be " acceptable to the
scholar" in a uniform system of notation could be " convenient
to" the evangelist ; for the one requires a perfect and elaborate
system of phonetic symbols as a means of etymological analysis,
the other " a commoner alphabet, more suited to a work-day
age" of missions ; nevertheless, if the two can accommodate their
plans to each other, it is to be expected that mutual facilitation
must result.

In conclusion it cannot be denied the subject which I have
chosen is an unusually dry one, inasmuch as it is an attempt to
arrive at the laws of a process which the great Macaulay
would have said " is not likely to be better performed merely

because men know how they perform it ;" and cannot, therefor e
be expected to command the attention of more than a few of the
most zealous students of phonetic science, much less of those
who are indifferent to the gratification of "finding out laws
from facts, causes from effects, necessary truth from fleeting
occurrences of the day." At the same time, it must be admitted
that a writer on such a subject could never hope to realise any-
thing remunerative, beyond the gratification of having embodied
opinions, based on his own researches, and of anticipating dis-
cussion in imparting them to others. But if even the materials
contributed are turned to account, without regard to my own
opinions, I shall consider myself fully compensated for my labours.
Throughout the work, which will be issued in three separate
parts, it will be my endeavour to trace the facts and phenomena
of the language to first principles. In the following pages on
the Consonants, I have worked upon a few materials which have
been known to missionaries for the last forty years viz., the
permutations of Initials ; especially as they have occurred to my
own observation during the last four years. I have dealt only
with simple consonants and their mutual combinations, as well
as other elements affecting them. It may be thought that, in
confining my generalization of particulars to initial consonants,
I have made an arbitrary selection, and avoided any reference
to numerous permutations which are perceptible in the
comparison of different tribal idioms of the Sechwana such, for
instance, as may be seen on a considerable scale in the comparison
of the " three members of the Bantu family of languages," which
Dr. Bleek considers " can be brought under certain laws,
similar, to some extent, to those detected by Jacob Grimm, as
affecting the relations between the different Teutonic tongues,
and other members of the Indo-European family of languages,
I have only to allege, in defence, that in the one class of instances
which I have chosen for analysis, the changes are constant, and

limited to cognate pairs of consonants no one initial element
being changeable to a third ;* whereas, this cannot be said of the
other class alluded to by Dr. Bleek. It must not be supposed
that, while thoroughly analysing the constant examples, I am
neglecting the dialectical variations, to a proper discrimination
of which, however, a far more critical knowledge of the South
African languages than has hitherto been published for sacred
purposes is absolutely requisite.

I should have preferred to treat of the Vowels first, but have
found it necessary to reserve the consideration of them for the
Second Part. This portion of the work will be based entirely on
a series of new facts, the fruits of my own researches into the
phonology of the language, on the subject of a peculiar order of
mutation existing among the vowels. The Third Part will be
confined to " sounds" formed by the combinations of simple or
compound consonants, with simple or compound vowels, and
which answer to the "Palatals" of some writers, the " Un-
stable combinations" of Dr. Latham, and the "Specific Modi-
fications" of Professor Max Miiller. My peculiarly unsettled
circumstances will render the issue of these remaining parts
rather uncertain.

As it is the first public essay of one who, during the last
twelve years, has seen little of current literature, and still less of
society, and the greater part of it has been penned in the native
village or at the wild encampment, it will be needless for him
to attempt to disarm the censures of critics, by defending him-
self on the score of style. Having, amidst the harassing
vicissitudes of Colonial border-life, long neglected the art of
composition, his work might doubtless have been rendered
" more readable" had the phraseology been corrected throughout
by some literary friend; but it has not been his lot to find
convenient access to any one.

* Except in one instance of h to A:/ for which I have attempted to account.

After a practical knowledge of this language during the few
years of childhood, an alternating use of the low Dutch patois
and English in the Cape Colony during youth, followed by an
education in England, and subsequently by a return to a rude
life among South African Boers, and more recently intimate
intercourse with the same tribes among whom I was born, it
cannot be said that any of these tongues is properly vernacular
to me ; or that I am open to charges of either national prejudice
or organical habit, which would be unfavourable to a proper
comprehension and discrimination of new elements. At all events,
the desultory nature of my experience will help to explain and
palliate any want of literary ability in the execution of a laborious
task. The work having been printed in England, during my
residence about 600 miles within the South African coast, some
allowance will no doubt be made for any lapses and mistakes
which I should otherwise have been enabled to rectify.

It was not till the Second and Third Chapters were in the
press that I succeeded in procuring a copy of "The English
Language," by Dr. Latham, nor till the first proof of the Third
had been returned, and the Fourth was in the press, that I met
with the " Missionary Alphabet" of Professor Max Miiller, and
the works of Professor Monier Williams, the Rev. Richard
Garnett, and Dr. J. Miiller, together with an exposition of
" Grimm's Law," in the 8th Edition of the " Encyclopedia Bri-
tannica." I was, nevertheless, enabled, in correcting the later
proofs, to avail myself of any notes or quotations from these
works, by interpolations, wherever they appeared to place my
own views in relief.

In reperusing it, I have reason to regret that a controversial
tenor is perceptible throughout ; but it is difficult to see how this
can be avoided in a work containing innovations on prevailing
notions. However I may have presumed to differ from Dr.
Lepsius, I need only say, that its publication was suggested by

his able and interesting pamphlet, without the thorough perusal
of which I could not have placed the results of my researches in
the form in which they now appear.

R. M., JUN.
Natal, South Africa,

January 28th, 1862.

These pages must be offered to the public in an unfinished
state. The lamented author did not live to complete what was
to him a labour of love. He died in his 36th year, at Mangeeri,
near Kuruman, South Africa, after only a few days of acute
suffering. In him the aborigines of interior Southern Africa,
whose language was his favourite study, have lost a disinterested
and enlightened friend.

The attempt is now made, in accordance with the desire of
his widow, to present in as complete a form as possible the result
of his investigations. As far as the 64th page, the proof-sheets
had been fully revised and corrected by himself. Beyond this,
it has been deemed best that nothing should be added, but the
typographical corrections absolutely required.

Those who may read this attempt to contribute something to
the general store of philological facts and inductions, will pass
lightly over such imperfections as must necessarily be found
under the circumstances.

J. S. M.

& Urban, Natal,

1st December, 1862.

CHAPTER I,

PRINCIPLES OF CONSONANTAL CLASSIFICATIONS.

I. OF CLASSIFICATIONS AT PRESENT IN USE.
IT is not my purpose to discuss at any length the propriety of
the distinction usually made between vowels and consonants, as it
would suffice for me to assume that they are distinct, in accord-
ance with the opinions of the majority of writers on the subject.
The following remarks, in a quotation from an antiquated work,
will perhaps be sufficiently distinctive to prepare the reader for
the sequel :

" WHAT thefe Vocal Organs precifely are, is not in all refpefts agreed by
Philofophers and Anatomilts. Be this as it will, it is certain that the mere
primary andjimple Voice is completely formed, before e*ver it reach the Mouth, and
can therefore (as well as Breathing) find a Paflage through the Nofe, when
the Mouth is fo far ftopt, as to prevent the leaft utterance.

" Now pure and fimple VOICE, being thus produced, is (as before was
obferved) tranfmitted to the Mouth. HERE, then, by means of certain different
Organs, which do not change its primary qualities, but only fuperadd others, it
receives the form or Character of ARTICULATION. For ARTICULATION is in
raft nothing elfe, than that Form or Character, acquired to Jimple Voice, by
means of the Mouth and its federal Organs, the Teeth, the Tongue, the Lips, &c.
The Voice is not by Articulation made more grave or acute, more loud or foft
(which are its primary Qualities), but it acquires to thefe Charafters certain
others additional, which are perfeftly adapted to exift along with them.

"THE fimplefl of thefe new Chara&ers are thofe acquired through the mere
Openings of the Mouth, as thefe Openings differ in giving the Voice a Paflage.
It is the Variety of Configurations in thefe Openings only, which gives birth
and origin to the feveral VOWELS ; and hence it is they derive their Name, by
being thus eminently Vocal, and eafy to be founded of themfelves alone.

"THERE are other articulate Forms, which the Mouth makes, not by mere
Openings, but by different ContaEls of its different parts ; fuch for inftance, as it
makes by the Junftion of the two Lips, of the Tongue with the Teeth, of the
Tongue with the Palate, and the like.

" Now as all thefe feveral Contacts, unlefs fome Opening of the Mouth
either immediately precede, or immediately follow, would rather occafiort
Silence than to produce a Voice, hence it is, that with some fuch Opening,
either previous or fubfequent, they are always connected. Hence alfo it is,
that the Articulations fo produced are called CONSONANTS, becaufe they found
not of rhemfelves, and from their own powers, but at all times in company with
fome auxiliary Vowel.

" THERE are other fubordinate Diftinftions of thefe primary Articulations.
******

" IT is enough to obferve, that they are all denoted by the common Name
of ELEMENT, inafmuch as every Articulation is from them derived, and into
them refolved." Hermes, by lames Harris, Efq., p. 318. (1771 .)

The above general description of the fundamental distinction
between the two main classes of the Elements of Articulation,
though written nearly a century ago, and founded upon the
opinion of one of the ancients,* is perhaps as clear as, if not clearer
than, anything I have met with.

According to it, the VOWELS are pure sounds, the variations
in which are caused by different configurations of the aperture of
the mouth ; and the CONSONANTS are elements resulting from
interruptions of the breath by the contact of different organs ;
and though the vowels can be pronounced without them, they
cannot be completely uttered without accompanying vowels.
It is, therefore, incorrect to speak of consonants as sounds, inas-
much as they require the apposition of these to render them
audible. In speaking of them I therefore retain their common
name, or otherwise call them " explodents?\ but only in reference
to the necessary separation of the organs after contact in every
case ; and in the same sense may be understood the word " sound"
when it occurs applicably to consonants in all my quotations,
without distorting the meaning of their authors.

True, some consonants have been called semi-vowels,

* Ammonius.

fin this restricted sense T place the word in inverted commas throughout
the analyses. The word dividual, used synonymously by Dr. Lepsius, would
have done just as well, but occurred to me too late.

liquids, &c., because though such arc formed by a contact, and
are therefore " explodent" there is only a partial interruption of the
breath, part of which escapes, giving to the otherwise mute
element a liquid or continuous nature. There is, however, so
much uncertainty attached to the subject of these peculiar
consonants, and the most able writers differ so in classifying
them, that it is not well to anticipate a proper analysis by any
conjectures; and I prefer to base my conclusions respecting them
upon a thorough generalization of facts which will come under
consideration in the order of this Essay.

Though M. Majendie writes,* " Grammarians distinguish
letters into Vowels and Consonants, but this distinction cannot
suit physiologists," his division of the elements of sound into
" those which are truly modifications of the voice, and those which
(as he thinks) may be formed independently of the voice," does
not differ materially from that of Harris. Nor does the system
of Girard and Beauz(3e,f who " confine the term 'articulation' to
the Consonants, and designate the Yowels by that of 'sons,'
(sounds)," differ in anything but terminology from the distinction
made above. For the Vowels, the term Sonants, in contradis-
tinction from Consonants, would be as suitable as any, were it
not also applicable to certain elements which may, perhaps, be
proved to be vocalized Consonants. More satisfactory terms than
those employed by Harris, and still in common use, could
scarcely be found.

Among the ancient grammarians, as well as some moderns
who have taken it upon them to interpret Grecian and Roman
authors, the consonants appear to have been classified upon one
or two principles, which have since been introduced into other
highly developed living languages, and remained long in vogue
among the learned. For a suitable example, it is only necessary

* El. Sum. Physiol., vol. i., p. 154. Cited by Sir J. Stoddart.
t- Gram. Geu., vol. i., p. 5. Cited by Stoddart.
B 2

to refer to a grammar of the Greek. Its fourteen consonants
were arranged according to three organs e.g.,

Mutes.

Tenues. Mediae. Aspirate.

Labials p b ph

Linguals t d th*

Semivowels.
Liquids.
771

n I r 8

Palatals k g kh

Of these consonants, nine (viz., those of each organ) were
found capable of another ternary arrangement, according to their
properties or gradations of breathing, into tenues, medics, and
aspirates ; and a certain relation was thus established as existing
among them. These were all denominated mutes, in contra-
distinction from the vowels.

Again, " the ancients found, in the humming and hissing of
the letters I, m, n, r, ," which did not fall under this second
ternary arrangement, " a transition to the vowels, and therefore
called them semi-vowels ; and the first four were named liquids,
en account of their mobility, and easily combining with other
letters."! The sibilant s appears not to have admitted of any
specific description.

This classification still obtains among some men of learning,
who maintain the distinction between "tenues (p, k, t} i.e.,
slender, weak consonants ; medics (b, g, d,} or consonants re-
quiring a medial quantity of air for their articulation; and
aspirates (p7i, kh, th,) or strong consonants." J The partiality of
many to it has of course been strengthened by their attachment
to the classic tongues, as well as by long prescription. Among

* The nature of the articulation of the ancient letter 9, has not yet been
satisfactorily decided; it may have been either t, combined with the spiritus
asper, or equivalent to th, in the English thin. I have preferred to assume
the former, in accordance with the opinion of Dr. Lepsius. See " Standard
Alphabet," p. 37, note 2.

f Buttmann's " Larger Greek Grammar," p. 11.

* "Egypt's Place in Universal History." Bunsen, p. 278.

many modern systems which have sought to supplant it, that
which retains the same division of the organs, but divides the
gradations of breathing into explosive and continuous, is beginning
to obtain among some eminent linguists, and probably owes its
introduction to a certain peculiar law of correlation existing
between several consonants ; but it is difficult to say whether
the discovery of this law is to be traced to any deductions from the
physiology of the human voice, or to the frequent interchange
and approximation of those consonants, which must have been
apparent to any observer.

At all events, I must for the present content myself by placing
before the reader such explanations of it as have been given by
writers upon a physiological basis.

"All the oral consonantal sounds, except I and? 1 ," says Sir John Stoddart,
" are produced in pairs, each pair having the same position of the organs,

but with a certain difference of effect The difference of effect

in each pair is produced in the same manner throughout the whole."
Glossology, p. 127.

This principle is elsewhere explained by Volney, thus: "Each contact
(or near approximation) of two organs forms two consonants, which differ
only by the degree of intensity of that contact, and which, under the names
of strong and weak (or the like), are absolutely of the same family."
Alfab. Europ., p. 71, cited by Stoddart.

Again, Dr. Richardson, in assuring his readers that Hofne Tooke
was guided by some general views of the " interchange of letters," writes
as follows :

" The perpetual change of t into d is familiar to all, and there is an
organical cause for these and other changes of B into P ; V into F ; G into
K ; Z into S ; J into SH ; and the Anglo-Saxon -D, that is, TH, as pro-
nounced in that, into their Q, that is, TH, as pronounced in thing. The first
of each pair ^including D into T) differs from its partner by no variation
whatever of articulation, but simply by a certain unnoticed and almost im-
perceptible motion or compression of or near the larynx, which causes what
Wilkins calls some kind of murmur." TJie Study of Language ; an Expo-
sition of the Diversions of Purley, p. 31. He then describes how Home
Tooke, by illustrating " the whole series of these organic changes," viz.,
Of w&GdBJZZZ
Into / p C t 9 sh s ss

in two parallel (but not very elegant) lines, as they would be repeated by
an Englishman and Welshman respectively, shows that a Welshman, by

" failing in the compression," changes seven of the English consonants ;
" to which compression," he adds, " we owe seven additional letters."
Ibid. p. 32.*

Having never seen an exposition of (> Grimm's Law,"f of which Dr.
Kichardson seems to think the above " general remarks" of Home Tooke
" evidently lay the foundation," I am, of course, unable to judge of the light
which it may have shed on this subject.

This correspondency, or correlation^: of certain consonants, has
given rise to a variety of terms, which differ according to the
views which the writers seemed to entertain of its nature. The
following are a few which I have encountered :

(Latham.^
'[SirJ. HerscM.

soft '

lenis

vocal Walker.

spoken ... Pitman.

soft } T _

. Volney.%

weak J

'Article, " Stammer" Penny

Sharp

Hard

Fortis

Breathing

Whispered

Firm and dry

Strong

Voiceless

flat.

" \LepsiuSy and others.
us j

Hard
Atonic
Mute
Surd

voice

weak

subtonic ...
semi-mute,
sonant .

Cyclopedia.
Adclung.

J3ishop,

Sanscrit Grammarians.^

* The fact of such crude remarks on this important subject making their
appearance in an able work, so recently as 1854, leads to the suspicion that
the laws of phonics must still be based on imperfect conjecture.

f A writer in an able Review thus speaks of " Grimm's Law :"
" It consists in a permutation, or, if you will, a play of letters, whereby

almost any word may be made gerrnain to any other Not,

however, that the said law, with its machinery of ' nine equations,' is
without a real foundation in the history of language. We only mean that
it is exaggerated, and exactly in the manner of all hypothesis, all analysis,
by being run out into a vicious circle." North British Review, Feb., 1859.

1 This is called, by Sir John Herschel, " a constant relationship or
parallelism to each other." Richardson's Dictionary, 8vo., p. 19.

These are cited by Stoddart.

The two series of consonants, which are generally supposea
to have this correlation, may thus be represented in parallel
lines :

k p t
g b d

ill (thin) s sh ch (-M) / ch (-fofc)
t/i (this) z zh gh v j

By means of a vertical line, I have separated from the rest the
first three pairs, each answering to the tenues and mediae of the
ancient mutes, to which the furies and lenea explodents of the
modern system are identical. Among the remaining instances
we find, of the ancient alphabet 1, the consonant s and its cor-
relative z, which latter was considered to be a double consonant ;
2, the hard guttural aspirate ch (-M). A corresponding soft form,
gh (pronounced like ch in loch by a Scotchman, or in buch by a
German), of this aspirate, and four other pairs, were thus left to
be included in the second division of a more comprehensive
classification. This appears to have been suggested as I
have before hinted, and as may be gathered from the tenor of
the preceding remarks by the above explained principle of
binary quantities.

I leave the intelligent reader to compare with the above two
parallel series of interchanges, obvious to the most superficial
observer, the following tableau of the " Simple Consonants in
the European Alphabets ;"* and have no doubt he will concur as
to the probability of this classification being based as much on a
vulgar view of the correspondency of sounds, as on any deduc-

* The same, on the graphic system of Dr. Lepsius :

Explosivae or dividuce.

Fricatives or continues.

Ancipites.

fortis. lenis. nasalis.

fnrlis. lenis. seniiruc.

Gutturalet

k g n

x'(x) h x'(r) y

Dentalcs

t d n

r I

(0'(G) 0' (S)

La^i nl' -a

p b m

f v w

Standard Alphabet, p. 38

tions of the phono-physiologist with the exception of, perhaps,
the two gutturals above noticed.

Ancipites.

Gutturalea ...
Dentales
lAibiales ...

Explosives or dividuce
fort, lenis. nasal.

A Ger. g Ger. ng

t d n
p b m

Fricatives or continues.

fortis. lenis. semivoc.

Ger. ch h Danish g Ger. j
!Fr. ch Fr. j
Sharp s Fr. z
En. th (-in) En. th (-ine)
/ Fr. v Eng. w

gutt. r

r I

Under the explosives, the only element not in the ancient phonetic
tableau is ng ;* thus completing the set of nasals m and n, which
were semi-vowels of the old arrangement. The remainder of
these viz , r and I, are excluded, as they anciently were ; and
left doubtful as to whether they belong to the first or second
division. The letter s, formerly considered peculiar, and more
recently called by a distinguished authority! " the last vowel
and the first consonant," but to this day still rather inexplicable,
sits the basis (phonically) of a formidable array of sibilants. The
elements w and y, sometimes called consonants, at other times
vowels, are introduced as fricative or continuous consonants,
guttural and labial respectively.!

Again, the unfortunate aspirates, which, Volney would say, re-
quire efforts of the lungs compatible with the vehement passions
and strong desires of the savage or rustic, are excluded from
this general tableau ; and only preserved from nonentity by the
fact that they occur in such highly important languages as the
Sanscrit, Bengali, and Chinese, &c., in the phonetic tables of
which Dr. Lepsius has inserted them. One of these, before

* Its omission was a mere oversight, as it occurred in the living pro-
nunciation in the first syllable of such words as ay-icos, ty-xos.

f Sir John Herschel, in Richardson's Dictionary, 8vo., p. xix. This
quotation does not occur in the synoptical table of sounds, in a later volume
of " Essays, &c.," by Sir John Herschel.

J These, of course, occurred in the ancient living-pronunciation, in the
diphthongs vi and ta of such words as veK.vi-a and Ov-ias, though not in the
phonetic tableau of the grammarian.

referred to viz., a soft form of cA(-M), viz.,^7t, the most common
instance of a rough guttural in the European alphabet, has by
some mode of analysis been divided into two forms of a gentler
and a harsher degree, and, under the letters ^ an( i 7> mar-
shalled in common with the sibilants, under the head of
fricatives, as if because swallowed up by the majority, or " de-
mocratic test of number" in which the latter exceed them!
Nevertheless, we find h, the letter equivalent to the spiritus asper,
huddled up between them, as if tacitly to imply that they are
aspirates. If we seek an explanation of this apparent anomaly,
we find the latest writer on this system saying

" The essential distinction of the three fricative formations, s, s, and 0,
together with the corresponding soft sounds z, z and Q', from the guttural
. . . x consists in the friction of the breath being formed and heard
at the teeth." Standard Alphabet, p. 45.

But, to add to the confusion of nomenclature, at p. 33 of the
same work, h, which I have described as being huddled up between
the two gutturals in the above general tableau, is called a " frica-
tive basis."

The principle of this classification is thus explained by Dr.
Lepsius, its recent and most able expounder, before referred to.
After disposing of the subject of the common and generally
admitted organic divisions, he adds

" There is another essential difference in the pronunciation, in as far as
either the mouth at the above-mentioned places* is completely closed and
re-opened, or the passage of the breath is only narrowed, without its stream
being entirely interrupted by closing the organs. The consonants formed
by the first process we call explosive or divisible (dividuce), because the
moment of contact divides the sound into two parts ; the others fricative,
from their sound being -determined by friction, or continuous (continue),
because this friction is not interrupted by any closing of the organs."
S tmid ml Alphabet, p. 30.

Before placing before the reader some of the elements of
another classification founded on incontrovertible natural facts, it

* " In the throat, at the teeth, or with the lips." Ibid.

may perhaps not be regarded as presumptuous if I endeavour
to show, on physiological grounds, where the system just de-
scribed is probably at fault.

As the consonants k, t, p, and g, d, b, respectively, fortes and
lenes of the division of explodents, differ in no respect except
terminology from the tenues and mediae of the more ancient
arrangement of mutes, I shall make no reference to them ; for
whether viewed in the above binary order, or the ternary one,
according to the organs, they are, so far as they go, intact
and indisputable. I have mainly to do with the group of " pairs ;"
which, while they bear an organical relation to the three series
above, are all supposed to differ from them in the gradations of
breathing, and have been denominated continues or fricativa.

NOTE. As the gutturals which are included in these terms have only
apparently a different phonical basis from the rest, and it is chiefly in relation
to them that the terms fricative and aspirate have been confounded, they
will require another line of argument ; I, therefore, reserve a further con-
sideration of them for the section on gutturals in the following chapter.

By the former term is meant the non -interruption of a sound;
by the latter, that the sound is determined by friction. From the
fact of both terms serving to describe the same instances, it is
evident that partial interruption of the breath is implied in all.
More definiteness ought to be expected in the treatment of
philological subjects ; and, till we succeed in arriving at that, we
may as usual grope in the mazes of sophistication, and be tan-
talized by a play of words. One would think that, of all the
natural sciences, none ought to demand more accuracy in its
rudimental nomenclature than that of language, whether phonetic
or grammatic.

The word continues includes several elements not classed under
it by those who make use of the term. " Vowels," which I have
assumed to be entirely distinct in their nature from consonants,
" are," in the strictest sense of the word, t{ continued sounds, pro-
duced when the passage of the air through the fauces is unin-

terrupted, the fauces being only more or less narrowed"* Not
only the four " pairs" of sibilants i.e., s, z ; /*(-in), th(ls) ; sh,
zh ; and /, u; the principal examples classed under this term
(in the letters s, z ; 6, 2; s, z; f, v) but also I and r, which
have been placed under the head of ancipites, because they
appear to betray a little of the nature of explodents and even
the nasals m, n, and n (-ng) in certain modified forms can all
be proved to be continuous consonants. There can be little
doubt that it is the " indistinct vowel" element (whether by the
term be understood action of breath alone or voice j, easily assumed
by them all which renders them so, and has led to their being
frequently called semi-vowels ; at all events, it is very probable
that it is nothing else than their peculiar nature in this respect
which renders them liquid with all other consonants.

The term fricatives conveys the idea of the breath being par-
tially confined, as well as, that it may be almost wholly inter-
rupted ; but, like the above, it includes elements not admitted by
those who apply it exclusively to some consonants. " Each vowel
requires a different elevation of the tongue, or contraction of
the lips/'f by which different degrees of frication, however appa-
rently imperceptible, are produced. And it follows that all the
above consonants, which easily assume the "indistinct vowel"
element, must, where so modified, also be fricatives ; with (but
only apparently) the sole exception of the nasals, in the enun-
ciation of which the breath is withheld, and there is no faucal
passage to cause a frication.

The two terms are thus strictly compatible in respect to the
instances I have shown can be included under them both. The
nasals are apparently the only exceptions, being continuous and
fricative in their modified forms, but not faucally.

* Encyc. Britann., 7th Ed., article Physiology, p. 683. The italics arc
my own.

f Ibid.

I shall now endeavour to illustrate this position by confronting
the two following quotations, which will show where contradiction
and confusion exist in the application of these terms. Dr.
Lepsius, treating of the "indistinct vowel-sound" attached to
some consonants, writes as follows :

" This vowel is inherent in all soft* fricative consonants, as well as in

the first part of the nasal explosive sounds It assumes the

strongest resonance, as may be easily explained on physiological grounds,
in comhination with r and I, which, as is well known, appear in Sanscrit as
r and I, with all the qualities of the other vowels." Standard Alphabet, p. 27.

Again, as follows :

" It is a decided mistake to reckon m and n among the consonantes con-
tiniKB ; for in m and n it is only tVie VOwel element* inherent in the
first half, which may be continued at pleasure, whilst in all the continuous
consonants it is the consonantal element (the friction*) which must he
continued, as in/, v, s, z." Standard Alphabet, p. 30, note.

In what precedes, I have implied that the vowel-element and some
amount of friction are inseparable in all articulations partaking
of the former, inasmuch as the breath is the medium of the voice.
The learned Professor seems to have forgotten, (1) that the
consonants v and z in the latter paragraph were some of the
soft fricatives of his classification, alluded to in the former as
partaking of the "indistinct vowel sound;" (2) that it is as
much the "indistinct vowel" element in v, and z, which is con-
tinued, as in any forms of m and n ; (3) a fact amounting to a
postulate in the science of phonics, that the consonantal element
does not consist in the friction but in the contact, whether partial
or complete, of two organs. In short, the important principle
appears to have been lost of, that if we are to regard the teeth
and the palate as organs indispensable to the action of the lips
or the tongue, in forming articulations by interrupting the emission
of the breath by their contact, it follows the elements produced
by them are also " explodent," because the moment of contact (of
the tongue with either of them, or the lips with one of them)

* The bold letters are substituted by me, as italics already occur.

divides the sound into two parts : that they differ from the other
mutes or " explodents," in that the contact is partial, thus causing
a gentle frication between them, and prolonged if necessary,
whether in the form of breath or voice, therefore continuous;
but these are not reasons sufficient to warrant their exclusion
from the division of " explodents." Dr. Richardson, treating
of the consonants b, p,f, v ; c (k), g (y), d, t; I, m, n, r, x ; s, z ;
very concisely says

"Each and every of them requires, however, for its complete utterance,
a breathing (precedent), a closure or collision of some of the organs of
speech, and an apertion or separation of them, with a breathing (subse-
quent)." Richardson's Dictionary, 8vo., p. xiv.

Moreover, granting, what is implied in the preceding quo-
tation, that " the complete consonant is best perceived when
placed between two vowels,"* and that "the full pronunciation
of an explosive letter requires the closing and opening of
the organ,"f which are both very plain statements on the
part of those who sanction this classification, and tantamount
to saying that the formative process of all consonants is the same,
it follows that those which easily assume the "vowel-element"
are also " explodents ;" this accessory " element" having its
origin in the pliability and mobility of the tongue an organ
indispensable even in the formation of the pure vowels.

The above train of reasoning suggests that there is something
very unsatisfactory in this classification of elements ; and, con-
sequently, that any graphic system founded upon it must
also be liable to objection. In referring to the two terms
continuous and fricative, I have endeavoured to show, (1) that
they include a large number of instances (vowels) having no
essential or generic resemblance to those within the scope of
induction (consonants) ; (2) that though consonants may possess
a secondary attribute (semivocal), this ought not to exclude
them from classification under a more general attribute (" explo-

* Standard Alphabet, p. 30, note. f Ibid.

dent"), definitive of the nature of the generalizations arrived at
by a survey of all the points in which they differ from or re-
semble each other.

In conclusion, granting, what is generally admitted, that both
the vowels and the pure " explodent" consonants (k, t, p ; g, d, b } )
may be aspirated, it follows that the other " explodent" con-
sonants, which easily assume the " vowel-element," may also be
aspirated i.e., those usually called fricativce or continues.

I thus show that the aspirate "explodents" include some
fricatives, the rest of which, therefore, fall under the simple
" explodents." A very important question then arises, as to
whether the correspondency existing between s and z, th(m)
and f/((is), sh and zh, and / and v, (s, z ; 6, $ ; I, z ; and /, v,
of Lepsius,) is analogous to that between the mute ex-
plodents k and g, t and d, p and b. If I have succeeded in
raising a doubt in the mind of the reader, I only leave him in
a prepared state for the impartial consideration of a few simple
facts, the analysis of which, in the following chapter, will perhaps
result in a synthetic view of a far more satisfactory nature.

II. GENERAL PRINCIPLES OF A CLASSIFICATION SUGGESTED BY THE
MUTATION OF CONSONANTS IN THE SECHWANA* LANGUAGE.

" The transformation of sounds," says Dr. Bleek, " is the main

* There has been a good deal of speculation at work on the origin of the
name of the people speaking this interesting language. The common
opinion among Missionaries is, that the name is derived from tshudna (be
like each other). Another holds that it is derived from tshudna, "a little
white, or inclining to white, light coloured i.e., not black (probably in
opposition to the more dark-coloured tribes of the north), a diminutive form,
from tshueu, white." Sir O. Grey's Library, S. A. Lang., p. 184. Mr.
Fredoux, of the Paris Missionary Society, writes : " For our part, we are
inclined to give it another origin. In the idiom of the people of whom we
are speaking, we find the word Mochuana (plural, Bachuana) employed as a
kind of diminutive of Monchu, black, and signifying blackish, or inclining

characteristic of the Setshuana,"* and it is probable that in no
other existing language is this principle carried to such an ex-
tent. It appears surprising that, though so long before the
public, it should not have suggested to the attention of philologers
the probable existence of some fundamental phonical laws.

This peculiar commutationf of consonants occurs chiefly in
certain instances viz., the formation of verbal nouns, e.g., seeing,
&c.,and those cases in which the verb is immediately preceded by the
" object, -particles" self and me, as in the following examples:

Verb.

Verbal Noun.

Verb, with t:
\=self.

^ "^n'nT^m!

I 3 '" Mutation.

1. Bona

Ponot

Ipona

Mpona

b top

2. Cola

Cold

Icola

Ncola

c immutable

3. Ndaea

Nteo

dtot

4. Gorisa

Khorisho

Ikhorisa

Nkhorisa

g-kh

5. Heta

Pheto

Ipheta

Mpheta

h ph

6. Kana

Kano

Ikana

Nkana

k immutable

7. Khatla

Khatlo

Ikhatla

Nkhatla

to black." Bulletin de la Societe de Geographic, de Paris, 4 Serie, t. xiv.,
p. 371.

The probability is, that Mr. F. has hit upon the correct derivation, upon
the following grounds. A black cow is called chwana, from nchu, black ;
in the same .manner as a red or black cow, with white back, is called
khwana, from nkliwe. Any little black thing would be called nchunyana ; a
little black person, monchunyana ; but the word being also used as a
diminutive of colour, a person with a dark brown complexion is invariably
called mochivana. Had the name been derived from chweu, or shweu,
white, it would have become bashdicdna, in the same manner as chdwdna,
a white cow, from chweu ; or kbchwdna, a grey cow, from Jcu-ebu, grey, in
which examples we becomes o. (In the above examples, tsk and ch are
homophones.)

Sir G. Grey's Library, S. A. L., p. 116.

f The language also contains other instances of interchanging consonants
which may appear to be irregular forms, but allow of classification under
euphonical or dialectical laws affecting the relations between the different
languages of the Bantu family. These will be noticed in the sequel.

I Used by the Missionaries as equivalent to ng.

The diacritical mark attached to o as o, will in every instance, for the
purposes of this work, indicate that vowel in the Italian perb, or the English
a in all.

C of the Missionaries=to&, or ch in Charles.

Verb.

Verbal Noun.

Loma

Tomo

Metsa

Meed

10.

Nama

Numb

LI.

$apa

Napd

19.

Pitla

Pitlb

]'!.

Phala

Phalo

14.

Eisa

Tisho

15.

Eata

Thato

16.

Sila

Tsilo

17.

Shbka

Cbkb

18.

Tena

Tend

1'J.

Thiba

Thibo

20.

Tlotla

Tlottt

21.

Tlhaba

Tlhabo

22.

Tsenya

Tsenyo

23.

lla

Kilo

"Object-particles," Mutation.

m, n, n = ?n.

Ntoma

1 to t

Mmetaa

m immutable

Nnama

Nnapa

Mpitla

Mphala

Ntisa

r to t

Nthata

r to th

Ntsila

s to ts

Ncbka

sh to c

Ntena

t immutable

Nthiba

NtloOa

Ntlhaba

tlh

Ntsenya

Nkila

tok

Verb, with the

Itoma

Imetsa

Inama

Tnapa

Ipitla

Iphala

Itisa

Ithata

Itsila

Icbka

Itena

Ithiba

Itlotla

Itllutba

Itsenya

IkUa

The above are all the consonants found in the phonetic systems of the
Missionaries of different Societies. Additions could be made ; but I prefer
to reserve such for the analysis in the sequel, as they are based on my own
researches.

In the above list of twenty-three initials, ten are mutable, and
thirteen immutable. If there are excluded from it, for the pre-
sent, those words with the nasal initials m, n, n, and those
with compound letters, all of which are in italics, it will be ob-
served that the remaining instances* may be classified upon a few
apparent principles:

1. That the immutable elements t and p are those which have
generally been denominated fortes, and the mutable elements b
and d, lenes ; and that the language so far confirms the fact of the
existence of each in binary quantities, to which I have before
alluded. The reverse mutation in the above examples, from
fortis to lenis t never takes place. This holds good in all the
remaining instances; therefore the presumption upon which
I start, in proceeding to the analysis, is that the commutable
consonants are respectively fortes and lenes in all these other
cases not commonly admitted.

* All these are indicated in bold type.

2. It may be distinctly observed that some are aspvvite, and
the rest simple " explodents ;" that the aspirates have binary
quantities corresponding with those of the simple " explodents,"
and that instances of " fricativae or condnuse" fall under both
divisions. This decides my presumptive arrangement into
simple and aspirate " explodents," instead of into explosicce and
continues. It must be evident that we have thus suggested to us
the practicability of arriving at some of the fundamental laws of
a general phonic system, without having recourse exclusively to
either the physiology of the human voice, or to any " written
system fixed by literature," however elaborate and ancient.

The whole of these elements, though gathered from one bar-
barous language by strict attention to the " living traditional
pronunciation," will be found in the following analysis to admit
of a simple classification, corroborating partially, but in a striking
manner, the views now obtaining, and in other respects entirely
upsetting them. In the Sechwana language we have two series
of consonants, both possessing binary quantities,* each of which
is only in certain circumstances changeable to the other; but the
reverse mutation never takes place. Such normal facts are
of great importance, and where constancy is thus attributed to
their laws, these are rendered the more worthy of being regarded
as a basis upon which to plant further investigations.

* The reader, if dissatisfied with the above term, is at liberty to substitute
the word values, or characters, or natures, provided the same be used
throughout

CHAPTER II.

ANALYSIS OF SECHWANA CONSONANTS.

IN accordance with the views stated in the preceding chapter,
I have, in the following table, removed from the " European
Consonantal System" of Dr. Lepsius, the consonants to which I,
for the present, take exception, as not all belonging to the set
corresponding with the Simple " Explodents /' but have included
the few under the head of Ancipites.

TABLE OF COINCIDENT CONSONANTS.

1st Division.

2nd Division.

fortis. lenis.

Nasals.

- 9(-9V

u(-ng)

t d

p b

Ancipites.

I. Gutturals .

II. Linguals...

III. Labials ...

Thus are presented the points of coincidence between a phonic
system based upon the physiology of the human voice, and
another arrived at by induction from particulars in the languag
of a barbarous people, " in a manner wholly concerning the ear."
It is now my purpose to explain and develops this natural system,
by treating of each horizontal (or organic) series separately, filling
it up according to the principles assumed, and thus completing
the vertical sets.

NOTE. By doing so, I am apparently admitting a division according to
the organs ; but my use of it is arbitrary, in order to make myself better
understood, by adhering as much as possible to the prevailing nomenclature
where no dispute exists. So far as regards the classification of the elements
of articulation, the views started in this treatise are discordant with those
C 2

generally held, principally upon tlie subject of the vertical divisions into
simple and aspirate " explodents" instead of mlo explodents and. fricatives ;
and the distinction of each of these into binary quantities, fortis and Ictiis,
in certain instances not usually admitted. On this account, I speak of 1st
and 2nd divisions at the commencement of the section in each series, to
avoid confusion in the above terms. At the close of each section, I give
the series as developed according to the Sechwana, in the most simple
characters that can be suggested.

I. THE GUTTURAL SERIES.

1st Division.

fortis. lenis.

2nd Ziitision.

fortis. lenis.

Nasals.

n(-ng)

The above Sechwana consonants, so far, happen to coincide,
as I have said, with the organic order generally admitted.

Under the First Division of the GUTTURAL SERIES, it will
be seen that the only exception is g, which is usually considered
the lenis form of k } and improperly called in English <l hard g."
This consonant does not exist in Sechwana, which thus affords
another instance of a language possessing an element of articu-
lation in one quantity, without its correlative form.

NOTE, But an apparent anomaly presents itself at the outsat, in the case
of Sechwana verbs with vowel initials, e.y. : The verb ila (hate), in the noun
expressing the action, becomes kilo; and with the object-particles, ikila,
nkila. This does not, however, contradict the piinciple which I wish to
maintain regarding the fixed correlation of the instances in every pair ; for
the above is merely an example of the substitution of a positive consonant
where none exists. It is, perhaps, " a mere contrivance for euphony's sake;"
something like that in Greek, in the case of the particle ov (not, no,) which
takes final e before vowels, as in the example OVK iveariv* A reason why
A is naturally preferred to either t or p in such a case, may perhaps be
alleged, upon a physiological basis, that in the articulation of it the mouth
is in a position best suited for the pronunciation of a vowel.

In Sechwana, all verbs with initial k, as kaiia (fix together),
retain this consonant in the initial of the noun expressing the action
kano, and in the verb itself with the object-particles prefixed
* Buttmann.

ikana, iikana. It thus remains immutable, and its quantity
is, therefore, upon the principle I have shown to exist in the
phonic system of the language fortis. It is probable that had
the Sechwana contained the simple " explodent" g (lenis), as
an initial, this would, in the inflection, have become k. Though
I have not the opportunity of thus testing, by this language, the
correlation which is universally admitted to exist between &and g,
a collateral, and perhaps sufficiently satisfactory confirmation
of it may be found in the two aspirate forms &h and g (-$/)
corresponding with them in the series, so far as regards quantity,
and which I shall presently notice. (Vide Section ii., Ch. iv.)

Like the rest of consonants, in the confusion of so many ortho-
graphic schemes, the instances in this pair, k and g, have been
separately the subject of " nice distinctions," especially among
authors reasoning upon a physiological basis, e.g. :

NOTE. "Adelung declares that tlie German AT has a double sound;"
that in kaum (scarcely), and stick (a bag), it is different from k in klein (little) ,
and ha ken (a hook) ; " and on the letter g lie makes somewhat similar obser-
vations." Glossology: Sir J. Stoddart, p. 132. Again, Mr. Bell observes
" that in k, before the close lingual vowel ee, the tongue strikes the palate
much further forward than before ah or aw, and that the same will apply to
g."lbUJ.

I shall not presume to question the utility of such phonical
divarications, believing that consonants, as well as vowels, are
apparently modified, but only slightly influenced, according to
their different combinations, their position in the collocation of
syllables,* and the capacity of the voice: that is, either ihefortia
or lenis form of any consonant may be enunciated with different
degrees of distinctness,! but its quantity, relatively to that of
the other instance of the pair, remains constant. The importance
of adhering to this principle, in order to arrive at a classification

* Or, more correctly: their position with respect to-the syllabic accent.

f Or, in the words of Sir J. Stoddart : " It is not to be understood that
either the one or the other articulation in each pair does not admit of nice
shades and discriminatory touches as it wre, perceptible to some ears, and
not to others." Glossology, p. 132.

of the normal elements of articulation, and so as to account for
the nature of all variable instances, cannot be estimated. To
the philologer, the desideratum of the day is a phonic standard,
with which to compare undefined sounds; a phonetic formula, or
analytic means of explaining apparent exceptions, and bringing
them within a synthetic view, as a further means of evolving the
principles of the science of etymology.

Under the Second Division of the GUTTURAL SERIES, the Sech-
wana presents probably the normal instances of what are generally
considered the most questionable elements in universal phonics ;
that is, upon which the opinions of linguists have been most divided.

The data are as follows :

1. All verbs with initial g* (-gh), as gata (tread), commute
this consonant into kh (or aspirate k) in the inflected noun ex-
pressing the action, khato ; also in the verbs with the object-
particles ikhata, iikhata.

2. Again : All verbs with initial kh,f as khwetsa (drive),
retain this consonant in the inflected forms khweco,
ikhwetsa, nkhwetsa.

In the former instance, the articulation of g (-gh) is changed
to kh ; in the latter, kh remains immutable. Both instances
are constant i.e., without exceptions; therefore g is a lenis
form, and kh the fortis of the same pair, and consequently
they are both aspirates corresponding to the simple " explodents "
k and g. Such laconic inferences may lay me open to a charge
of presumption ; but in the sequel I shall have recourse to col-
lateral evidence, and the testimony of able authorities.

* Called by Dr. Bleek " a soft kind of guttural .... denoted by
g, as in. Dmtch." Sir G. Grey's Library, South African Languages, p. 113.

It ! the g in Dutch dag = eh in German buch.

f This digraph = fch in milk-h^use, with mil elided, the accent on khouac,
and the h distinctly enunciated.

Confusion in Nomenclature and Orthography, in the u*ual
Classification of the Guttural Consonants.*

[Sir John Stoddart, in speaking of a pair of gutturals,
very correctly says " They are evidently susceptible of modi-
fication by slight differences in the position or action of the
organs."f I have already had occasion to refer to them ; and
shall do so again in the details of this section, in order to show
that the circumstance of their partaking of the nature of fricatives
as well as aspirates has hitherto led to a classification which, how-
ever elaborate, is unsatisfactory to some like myself engaged in
local researches into a language till recently unwritten, and to
this day without a proper grammar or vocabulary.

Indeed, it may truly be said of all the gutturals, but especially
of their aspirate forms, that they are of a nature to deceive the
ear more than any other simple consonants, especially on account
of the difficulty of distinguishing them accurately from purely
" palatal ''} utterances, or other sounds resulting from a combination
of these with gutturals.

NOTE (). In German, Adelung distinguishes the articulation cli into

* The reader is at liberty to pass on to page 40, in order to follow the
analysis ; or to read this parenthetical section separately, as from its nature
it is treated in a desultory manner.

f Glossology, p. 130.

\ Whenever this word palatal occurs within inverted commas, it is ill-
tended to represent a more general term (to which I shall have recourse in
the sequel, Part III.), of which the above word includes only one of the
special meanings.

Gutturals are by some writers also called palatals, perhaps from this
very circumstance ; and by others are made to include h, or the sfiritus
nsper.

two degree 3 , a stronger and a weaker; and Muller reckons. the three
following modifications of the same :
1. In Ueblicli, sellg, &c.
y. In tag,suehen, aud,
8. As uttered by the Swiss, Tyrolese, aud Dutch.

NOTE (1). Volney has two pairs of sounds approximating to those under
consideration viz., (l),his twelfth class, distinguished by the French 'term
q rasseyement (thickening of utterance), which has also a strong and a weak
pronunciation :

The stronger (grasseyement dur}, which he compares to the 19th Arabic
letter "," and is " common among the Parisians and Provengals, and pre'
dominates among the Berbers." Glossology, Sir John Stoddart, p. 130.
This is the letter,' the power of which Dr. Lepsius considers equivalent
to the modern Greek y. Standard Alph tbet, p. 48.

Tlie weaker (grasseyement doux), in this "the position of the organs ia
very similar to that which produces the vowel i." Glossology, Sir John
Stoddart, p. 130. Dr. Lepsius appears again to consider the sound
meant here as equivalent to " the modern Greek 7." Standard Alphabet,
p. 13, note 3.

" A transition often takes place from the one to the other of these articu-
lations, in like manner as we find the Hellenic ylXav . . . become in
Romaic yelan."

(2) His thirteenth class, containing The stronger, as in bitch ; the
weaker, as in Metternich. Glossology, Sir J. Stoddart, p. 130. Otherwise
expressed by Dr. Lepsius thus: " The German ch in ich, lie (Volney)
places as a soft sound by the hard sound in buck." Stand. Alph., p. 13, note 3.

NOTE (c) Dr. Lepsius (at page 43, " Standard Alphabet"), in describing
the difference between guttural k and a palatal &', adds: "In most lan-
guages A- aud g before the vowels e, i, a, ,* approach the palatal pro-
nunciation, whilst before a, o, u, they remain more guttural, owing to the
formation of these vowels In the- Sanscrit, the gtittural and palatal pro-
nunciation were distinguished before all vowels."

Again : (p. 48.) We find the same two sounds alluded to by Volney
(1 3th class) thus classified :

" x g- Such, ach ; pol. chata.

X' Sansc. TSfj Ger. ich, recht."

The former as a guttural, the latter as a palatal-guttural, if the diacritical
mark has any value.

Again: (Note I, p. 43.) The Germans " pronounce, for instance, the ch
in all diminutives, even after a, o, and u, not guttural as in Aachen, rauchen,

* Of his graphic system.

f This Sanscrit (?) and German sound is again called a palatal sibilant.
Stand. Alph., p. 42, note 1.

&uoA#A,bat palatal, as in Mamaeken, fr'rauuhen, Uhuchen, from Mantn, Frau,
Uhu. The guttural e/t is pronounced after all vowels in the most southern
parts of Germany."

The indistinctness of the above remarks arises from this learned philologer
not being sufficiently specific in describing what he means by the " palatal
pronunciation," or an ' approach to the palatal pronunciation." It may
mean either, that of the " pure and simple palatal" y and its aspirate form
7), or that of h or y, " easily assuming a shade of ?/."* If the former, it must
fall without the guttural series ; if the latter, it becomes a palatal-guttural.-|-
But if the pointed letter A 1 is intended for the sound of ch in choice,
as Dr. Lepsius seems to intimate (p. 41), it is then quite as likely to
indicate the "palatal" form of some other compound consonant. It
is first necessary to arrive at the nature of elementary sounds, in order
to comprehend the manner in which they are transformed or modified by
usage.

The matter is rendered still more indistinct by the following (Note 1,
p. 37) : " The modern Greek y passes at least before e, i, v, into the cor-
responding fricative sound." See above example of yeXav. The terms
palatal and fricative, intended by the learned writer to differ in meaning,
are thus confounded in reference to the gutturals. The writer elsewhere
speaks of a i: palatal sibilant," and it is easy to comprehend the nature of a
palatal-lingual ; but when (p. 43) he speaks of cases in which " the friction
connected with the palatals" is "so inherent, that in the organic structure
of the language it may be considered as still forming a simple sound"\ it is
difficult to conceive of what is meant. I think much confusion would be
avoided by a distinction being first made between simple or compound
consonants, and the modifications in either by the superaddition of the pure
" palatals" a plan I intend to pursue by the assistance of the Sechwana.
In the sequel I shall attempt to prove that an element of articulation,
exactly equivalent to ch in the German ich (above alluded to by all these
distinguished authors), and wbich exists in Sechwana, is simply the pure
"palatal" y, with the spiritus aspet: A careful perusal of the above
quotations and remarks will no doubt suggest the probability that the
other few sounds referred to are either different gradations of the soft
consonant ch in Inchen, or in some instances combinations of the pure
"palatals" with them.

Dr. Lepsius, in his elaborate work, under the series of gut-

* Or, as Dr. Bleek more happily expresses it, " being influenced by a
superadded y."

f Vide " Standard Alphabet," p. 41, in which both are confounded in
" the series of pure palatal sounds," which cannot surely include compounds
of simple consonants with pure palatals. See Part III. of this work.

t The italics are mv own.

2t>

turals, gives, besides his e.rplodents, two distinct sets of pairs,
Avhich he denominates fricatives and aspirates ;* but between
these divisions T must confess myself utterly at a loss to com-
prehend the difference in reference to the gutturals. I have, in
the preceding chapter, stated wherein it appears the classification
which he sanctions is at fault. It becomes me to bear out the
truth of my remarks on the details of each organic (or horizontal)
series. As this learned philologer has given a synoptical and
comparative view of the phonetic systems of several languages,
as reduced to his own graphic system, and also as represented
by the authors from whom he gathered his particulars, I cannot
do better than subjoin a compendious abstract of the gutturals
under those heads of fricative and aspirate, in order to enable
the reader to follow my inferences.

AFRICAN LANGUAGES.

1. Hottentot:
Wuras . ..

Fricat
fortis.

Ml
c
ch

r, r
kh
h
c
h

ive.
lenis.

r
r

f j

Aspirate.
fortis. lenis.

Knudsen

2. Kafir:
Appleyard

3. Zulu:
Grout

4. Tsuana :
AppleyarA

5. Kua:
Prof. W. Peters

6. Swahili:
Krapf .. .

7. Herero :
C. Hugo Hahn ...

8. Mpougwe :
Am. Board on the Gabun.

* The latter, it must be remembered, are not included in his " general
tableau," but only in phonetic tables of the few languages in which they
are supposed to occur.

L'7

Fricative.
fortis. lenis.

Aspirate.
fortis. lenis.

0. Fernando Po :

John Clurkf h

10. Yorfiba:

S. Crowllicr h

11. Ofi:

H.N.Kiis

1:2. Susu:

/. W. G., Am. Board Mi

13. Maude :

Mucbruir ' h

14. Vei:

S. W. Koelle h r

15. O'lof:

Roger kh hr h

1C. Housa:

Sohon h y

17. Kanuri :

S. W. Koelle h

18. Nubisch:

Lepsius h

19. Koiigara:

Lepsius (h)

20. Galla:

Ch.Tutscheli

21. Hieroglyphic :

Lepsius x

22. Koptic :

Lepsius x

23. Bega:

Lepsiua

24. Abyssinian, Ge;ex .

Ludolph i k

25. Abyssinian, Amhara :

Isenberg ..' h ch

NOTE. By the above list it appears that of twenty-five African lan-
guages, nine are represented as having both quantities of fricatives, i.e.,
fortis and lenis; thirteen contain fricatives of only one quantity, and the
remaining three, neither fricatives nor aspirates ; but, among the whole
number, only three are said to contain aspirates, viz., the two Abyssinian
dialects, and the Koptic.

From the above, a rather sweeping inference may be drawn,

viz., that about seven-eighths of the African dialects have no
aspirated gutturals.

Fortunately for my purpose, the list contains, at the outset,
three or four South African languages, with the nature of which
I happen to be more or less acquainted viz., the Hottentot or
Naman, the Kafir and Zulu, and the Sechwana ; the phonic
system of the latter being the subject of my particular attention,
I have made it the basis of these investigations. To test the
propriety of the distinction between the terms fricative and
aspirate in the case of the gutturals, as well as to show the
confusion arising from a mere comparison of alphabets, 1 shall
consider each of these languages in succession.

(a) SECHWANA.

The pair of consonants in this language, which I have already
indicated by the letters kh and g (in use among the mission-
aries), and proved, upon a certain principle, to be aspirate-
" explodent" gutturals, are indicated in the list by the same
letters,* but without a knowledge of their nature, both pla'ced
under the head of fricatives by Dr. Lepsius, and, moreover, re-
presented by the ancient Greek letters ^ an( i 7> m his corres-
ponding graphic system. To say that the consonant thus
intended to be indicated by the Greek x and called a fricative,
is anything but a slight modification by a gentler aspiration
that is, differing only in an insignifioant degree, if at all, from
lenis aspirate g (-#/<) is, I think, carrying the habit of phonical
"hair-splitting" to an excess. The appropriateness of the letter
X, in any new graphic system, would consist in its being intro-
duced to indicate an element of articulation which bears the
closest approximation, if not absolute identity, to that which it
anciently indicated. In lexicons, x is represented as having been
" a strong guttural aspirate." Its former identity to k aspirate

* Mi-. Apployard. the authority cited, probably procured Ids informaliou
on the subject at second-hand, from Mr. Arch-bell.

seems to be generally admitted ;* at all events, I may avail
myself of an independent proof of this (especially as it will serve
to illustrate the Sechwana.), in the fact that the same Greek
particle ov, previously noticed as taking final K before initial
vowels, took x before the same vowels with the spiritus asper,
e.g., ov% vtrtfjTir, which could not of course mean the doubling
of an aspirate, but the coalescence of final K with the spiritus
asper of the following vowel. This is more distinctly shown in
the case of ^t/ca, combined with tiptpa, forming lf\i]p-^og. Whereas,
X is intended by Dr. Lepsius to indicate a continuous consonant
viz., ch in the German lachen, differing only in degree, if at all,
from that I am about to notice (g, -gh}-

This reference to the orthography of a dead language enables
me to explain exactly the nature of the consonant in Sechwana,
which I, for the present, call an aspirate " explodent." In this
language it is at times gently enunciated ; at others, forcibly ;
but the fact of an aspiration accompanying the simple " explodent"
k, is unmistakeable.

The other consonant g (-#A), which the genius of the Sechwana
proves to be a corresponding lenis form of the fortis kh, and
therefore also an aspirate, is more difficult of illustration. An
evidently very close approximation to it is that intended to be
indicated in Dr. Lepsius's system by the above letter ^ > and
classed as a fortis fricative. That the Greek letter y, adopted
by him, anciently indicated the same lenis consonant as that now
under consideration, is not at all borne out by Greek lexico-
graphers. In fact, the supposed identity of y, in the traditional
Greek pronunciation, with the Arabic , satisfies me that it
would only be applicable to a vocalised form of the German ch
in the above example, or to the Cape-Dutch g] in dagen (days).

* See " Standard Alphabet" note 2, p. 37.

f An articulation approximating to what is improperly called the
" guttural r" but only more gentle.

It must now be evident to the reader that the two letters
X and x' (y)j which the able linguist has adopted to represent
the elements ch in Germ, lachen, and the Arabic respectively,
are by him erroneously applied in the Sechwana to other two
elements, which, in his graphic system, ought to be written k' (or
M), and ^, respectively ; so that only one of the two is correctly
regarded as an equivalent, but merely misplaced as to quantity
viz., the latter.

These two consonants in the Sechwana are thus described by Dr.
Livingstone, in his " Analysis of the Language of the Bechuanas,"
(Section 1.)

" kh is the k strongly aspirated, as in khakala (far) ; <j guttural as ch, in
locli (Scottice), dag (Dutch)."

At the end of the section, he calls kh,ph, &c., " aspirated explosive sounds,"
which agrees with the nomenclature I have preferred for explanatory pur-
poses, and, indeed, with the definition of Dr. Lepsius.*

Whereas, Dr. Lepsius's guttural letters ^ and ^' (y) are well
illustrated in the Cape- Dutch examples dag 1 , dag'en, in which
g l and g* are to each other as s to z, and as th (in thin) to tli (in
thine}.

(I) HOTTENTOT (OR NAMAN).
In this language there are two consonants thus indicated by

Mr. Appleyard, or rather \Vuras M ch

Mr. Knudsenf c

Mr. Wahlmann .. ch

* "Aspirates are those explosive sounds which are pronounced with a
simple hut audible hreath." Standard Alphabet, p. 49. The word forcible
in place of "audible" would convey a more correct definition. The word
" explosive'' omitted, or the words elements of articulation substituted for
" explosive sounds," would include the vowels.

f By reference to the " Correspondence of the S. A A. Bible Society,"
p. 6, containing a letter from Mr. Knudsen, this missionary, in speaking of
an orthography, writes thus : " It would also be better to use for ch the
Dutch g, or something else, and for kh another single letter." So that it
would appear he admits two forms of this consonant. As to Mr.
Wahlmanii, I believe his observations were based on data furnished by
Mr. Knudsen.

These are also both placed under the head of fricatives by Dr.
Lepsius in the above list, and thus regarded by him as identical
to kh and g of the Sechwana. From my own personal observa-
tions, in frequent intercourse with both Koranas and Namaqua
Hottentots, as well as the Bechwana, I am able confidently to
confirm the fact that this language has two aspirated gutturals
exactly equivalent to those just described under the Sechwana.
These, Mr. Wuras's* digraphs are, no doubt, intended to indicate ;
but I prefer to add the authority of an individual who, by reasno
of intercourse with natives during his youth, is perhaps not sur-
passed by any other missionary in a knowledge of the Naman
language, whether practical or critical viz., the Rev. H. Tindall,
of the Wesleyan Society. Mr. Tindall, in describing the two
gutturals of this language, which he indicates by kh and<jr/i, says :

' Qh is a soft guttural, as in the Dutch word yaan.

" Kh represents a much deeper and harsher guttural than gh. We have
no corresponding sound, either in English or Dutch, by which to illustrate
its power.

" These two letters are simple sounds, though represented by compound
(onsonants. In spelling they are pronounced at a single articulation,
without separating the combined characters ; thus ghun is spelt gh-it-n, and
not g-h-u-n. Khnp is spelt kh-a-p, not l:-h-a-p."\

The error of the Professor, detected in his Sechwana examples,
is here repeated in the application of his graphic system to the
Naman gutturals ; and not only are the terms to which I have
alluded confounded, but also the consonants misrepresented.
These two gutturals are identical in the two languages ; but in
both it is only the lenis form that coincides with one of Dr.
Lepsius's consonants, viz., his fortis ^.

* See Sir George Grey's Library, South African Languages, p. 19, note,
for some additional information by Mr. Wuras himself, in a letter to Sir
G. Grey.

f " Grammar, &c." See also Correspondence of 8. A. A. B. Society, p. 8, for
further remarks bv Mr. Tindall on these consonants.

In these dialects, two guttural consonants are arbitrarily repre-
sented in the graphic systems of the missionaries, by
Mr. Appleyard*.,. r r (Kafir)

Mr. L. Groutf ... r or r r (Zulu)

These are also denominated fricatives in the above list, and are
thus regarded by Dr. Lepsius as identical to the Sechwana
kh and g, and the Naman kh and ch, in his arrangement. I
now proceed to add the testimony of a third very able authority,
the Rev. J. L. Dohne, in his elaborate and valuable dictionary,
published two years later than the " Standard Alphabet," as to
their nature :

" The two gutturals are both harsher than the gutturals in any European
tongue. The one, called the softer, is exactly like to the compound sound
of the Dutch gr in groot , groet ; Zulu-Kafir rola, rauka, &c."

" The other is a hard, ringing, harsh sound, which it is very difficult to
describe, and still more so to utter, but by no means impossible for a
foreigner to pronounce. It is made by contracting the aperture of the
throat, and expelling, as it were, the breath forcibly, so as to produce a harsh
rustling of the epiglottis."];

(1) From this it would appear that the former, which is de-
scribed as a lenis, answers exactly to the Dutch g, with a super-
added r, and therefore to the Sechwana g, similarly compounded.
But, according to the testimony, not only of all missionaries, but
also of Mr. Charles Brownlee, a high authority in the Kafir, "no
such sound (as r ) exists in the language ;" therefore it would

* See Mr. Appleyard's later remarks, p. 107, Correspondence S. A. A. B.
Society, suggesting k and g, with points over them, to represent them.

f Now indicated in the graphic system of the " Zulu Grammar and
Dictionary Commission," as follows :

The softer, by h barred.
The harder, by k barred.

See p. 83, Correspondence S. A. A. B. Society.

I See " Zulu-Kafir Dictionary" Letter R (1), p. 306. Also Letter G,
p. 90, under which are some remarks on another soft guttural, apparently
rather difficult to describe, but referred to in the sequel.

Correspondence S. A. A. B. Society, p. 78.

perhaps be nearer the mark to liken it to a rough form of the Dutch
or Sechwana g, (Germ, ch gutt.) Mr. Grout, in his Grammar of
the Zulu, published ?ince Mr. Dohne's work, describes it thus :

A soft guttural sound, somewhat broader than the German ch, in "macht;"
and corresponding more to the guttural sound of . . . g in the Dutch
words, "God," "yoed;" thus " \ola" " %a \a'," p. 16. (In this Grammar
he has adopted Lepiuis's orthography.)

I have since had an excellent opportunity of testing several
Zulu consonants by my own ears, and have ascertained this
element (x of Lepsius, r of the missionaries) in the Zulu to be
strictly identical to the Sechwana (lenis) g of the missionaries.
I subsequently met with the abridged Zulu Grammar of the
Bishop of Natal, who thus alludes to the same consonant :

" The . . . letter r is taken to represent the guttural which sounds
like the strong German ch, as heard in aiich, noch." (P. 2.)

So that there can be little doubt as to the strict equivalence of
this Zulu consonant to the Sechwana g, and the Naman gh; but,
as in those instances, Dr. Lepsius, by making it his " lenis" 1 of ch
in lachen, has erroneously supposed it to be another example of
the vocalised form, or the Arabic element . Mr. Grout, again,
by making it the lenis of the two (meaning it to be the equiva-
lent to ch in lachen), has evidently not apprehended the nature
of the element intended by Dr. Lepsius to be the *' lenis." Such
is, consequently, the present confusion in Zulu phonography.

(2) In the pronunciation of the other instance of the pair, the
breath is said by Mr. Dohne to be "forcibly" expelled, and
would thus appear to answer to a harsher form of the Sechwana
and Naman kh ; for, even among the Bechwana, both Teh (k' ) and
g are at times very coarsely enunciated by natives possessing
stentorian voices ; just as we fancy the German guttural ch to be
pronounced more forcibly in the Scotch loch, or the Dutch daft,
though the same in quantity with them. However, the fol-
lowing is Mr. Grout's own description of the same element, which
he now indicates by x':

"The letter x represents a peculiar, hard, rough guttural sound, which
seems to be made by contracting the throat, and giving the breath a forcible
expulsion, at the same time modifying the sound with a tremulous motion
of the epiglottis, as in ultu%eza, uliu\eba, um^ezo^ P. 16, Zulu Grammar.

On my own examination of a native I was disposed to think
that this element ^' ( r r ) was not > as supposed by Dr. Lepsius,
equivalent to the Sechwana and Hottentot kh (A"), but simply the
former (lenis of Grout, Germ, ch, Sech. g), accompanied by a
decided lateral click, and that it was actually neither of the two
guttural elements of Dr. Lepsius, but a double consonant.

As in the preceding instance, I was gratified to observe that
my view of this articulation was in a measure corroborated by
Bishop Colenso's description of it.

" There is another sound occurring in a few Zulu words, which may be
pronounced either as a guttural from the bottom of the throat, or as a click
in a peculiar way. But the sound must be heard in order to be imitated.
We shall denote it by x ; and the student may get a native to sound it for
him." First Steps in Zulu Kafir, p. 2.

My own description differs from this only in being more de-
finitive. It is decidedly not a simple consonant, and Dr. Lepsius
has erred in classifying it as the equivalent of both the Sechwana
or Naman kh (k "), and the German ch, under the delusion that
the two latter digraphs indicated identical elements. Mr. Grout,
again, with the " living traditional pronunciation" of the Zulu
at his ear, has evidently never taken the trouble to ascertain the
real nature of both elements intended to be indicated by Dr.
Lepsius's letters x an ^ x' the former of which he has in fact
reversed as to quantity, and applied that with a diacritical mark
to a compound though cognate consonant.

I do not know whether the vocalised form of ch in lachen (that
is, the ^/izs-fricative-guttural ^' of Lepsius, Arabic ), which he
has confounded with the Sechwana g\ is to be found in Zulu ;
but have ascertained beyond a doubt that this language neverthe-
less contains the Sechwana or Naman kh (&}, which he has con-
founded with his x> and I have endeavoured to prove is properly

the fortis form of this ck (gutt.) At page 16 of his Grammar, Mr.
Grout says : " There is also a sound intermediate between that ofy
and k" On trying, by my own ear, so accustomed to Sechwana
aspirates, the two examples he refers to viz., ukuganda or kanda,
and utukela or utugela, I distinctly perceived that the former was
pronounced uku khanda, and the latter utukhela,* with k" in both.
I am afraid I have brought my reader into a phonetic labyrinth,
and would rather it had fallen to the lot of some other writer to
make an exposure of any imperfections in the graphic systems
of men labouring to arrive at uniformity ; but conceive it to be

* Lest, in the above statements, I should lay myself open to a charge of
obsequiousness and party-spirit, I append the following facts : In the course
of a short visit to Natal (August, 1861), I took the opportunity of riding
out about twenty miles or more from Durban, to the mission station of the
Rev. Mr. Rood (who was at the time absent). My object was to become
satisfied, before committing myself to print, of the true nature of the two
gutturals r~ and r of the American missionaries, or % an< i X f Lepsius. As
Mr. Grout's Grammar had just been published, I carefully went through his
letters of the alphabet with an intelligent native ; especially as I had my
doubts of the correctness of his descriptions of the Zulu elements ; for I had
on one occasion observed the written words isipingo and tina distinctly pro-
nounced by some gossiping natives isipingo and 'tina. The result of my
observations respecting the two gutturals is stated above. As also shown
above, I found that Mr. Grout's " sound intermediate between h and </"
was kh (k'). In addition to these facts, I observed (2) that his " sound
intermediate between the genuine p and b" was ph (p'), e.g., his word

popoza was what I should write in Sechwana pho-phoza (p'op^oza) ; (3) that
tina was really pronounced thina (t'ina), and utixo really uthixo (uftixo) ; in
fact, that not only the simple mutes k, t, &nAp, existed in the language, bujt

also their unmistakeable aspirate forms, the equivalents of those in the Sech-
wana viz., Jc' t t~ p~ I not only stated these facts to my relative, Mr.
Laurence Platt, my companion on the jaunt, but the morning after to Dr.
Mann, the able superintendent of education, while riding a short distance
with him from the Isipingo, in the direction of Durban. A day or two

corroborative quotations, of which I have already made use (at p. 2), con
tained the following paragraph :

" There is a slight aspiration heard in very many words (as in Hebrew)
after either of the letters b, g, d, k,p, t. This will account for some roots,
which in the dictionaries appear identical, having a difference of meaning,
which a native would distinguish by his enunciation, though it will require
a fine ear to detect it.

Ex. kona, it; but kona (pronounced khona), these.

kirako, its; but kwako (pronounced kunklio), thine."
D 2

my duty to divest this important subject of a confusion
worse confounded by an obsequious attachment, on the part of
those who have the living speech within hearing, to " written
systems fixed by literature," and the authority of great names.
I do not, however, flatter myself that any will wade through all
my remarks, but shall attempt to give as synoptical a view of
the subject as possible, employing the letters of Dr. Lepsius's
system, and in brackets those of the missionaries, &c. :

fortis.

X
lenis.

lenis.
X (>")

lenis.

x'(y)

of Lepsius.

of the Sechwana.
of the Nainan.

of Wr Grout. W the
hopColenso, Zuku

fortis.

*< (M)
& c (kh)

compound.

:::*:: . ...

k\kh)

Observed by Bis

and more recently by myself.

The instances which I have observed for myself are quite as strong as
the aspirated mutes in the Sechwana, though I have no doubt slighter
aspirations would be found to occur in unaccented syllables. I am not
sufficiently acquainted with the Zulu to be able to give an opinion as to
whether the lenes mutes g, d, b, occur, but I need not say I was most agreeably
surprised to find my views respecting the fortes corroborated by the autho-
rity of an able scholar.

This will perhaps be the fittest place for me to express an opinion which
I have long entertained. It is much to be regretted that, in the compilation
of his elaborate and interesting Dictionary, the Rev. Mr. Db'lme has not, by
means of diacritical marks, or other ^expedients, shown, in the case of every
word in which they occur, the variations of the following letters : e, g, m, o, r
(gutt.), hi, &c., &c., to each of which he attributes two or more powers.
The importance of such distinctive marks in the search after roots cannot
be estimated ; and is moreover enforced by a reference to precedents, in
which phonetic distinctions in such a language as the Arabic are shown
to clear up apparent difficulties in several cases of the Hebrew having one
word with entirely different meanings. Unless Zulu scholars look to their
phonography, the cognate Sechwana will doubtless, ere long, have the
advantage, like the Arabic; simply because of greater attention having been
paid to phonetic distinctions.

All the confusion consists in the instances of the 2nd, 3rd, and 4th pairs
being reckoned by Dr. Lepsius the equivalents of those of the 1st pair re.
spectively, and in &' having entirely escaped the notice of the American
missionaries ; whereas, the vertical columns show the real phonical equivalents.

There can now be very little doubt as to the strict equivalence
in quantity of certain pairs of rough guttural consonants in the
three species of languages I have just examined. I have shown
those in the Sechwana to be combinations of the simple " explodent*
gutturals k and g (gli) with the sjriritus asper ; personal inter-
course with the natives, and the authority of Mr. Tindall, enable
me to identify with them the two harsh gutturals of the Nam an
language ; and the descriptions of able linguists among the mis-
sionaries, added to my own observation, lead me to conclude that
the " Kafir species" also contains them both.* Therefore I have
considered myself justified in concluding that they are, in every
case, formed by a combination of the spiritus asper with the simple
"explodents" k and (a form of) g. But in no case are they the exact
equivalents in quantity of the pair under which Dr. Lepsius has
classified them in his tables, except that hisfortis form x answers
to their lenes forms, though differently indicated (g, gh, or r.) It
will no doubt have been observed that all the authors quoted speak
of the softer form of the two, i.e. } g (-gli) being simply a guttural, or
a soft guttural ; the proof of its being a lenis form of the aspirated
guttural, kh, and therefore also an aspirate, is alleged by me solely
on the principle existing in the Sechwana phonic system, and
which I hope to maintain throughout this treatise.

As to the remaining African languages in the list, I must of
course plead absolute ignorance, in the absence of the living
pronunciation, which would enable one to compare for himself;
but, so far as regards the absence of authentic facts, my dis-
advantages are surely not greater than were those of Dr. Lepsius,
when compelled to base his conclusions on second-hand data,
contributed, of necessity, by different and various authors, at
several intervals, and of course without a preconcerted plan.

* Since the above was in type, I have met with Dr. Van der Kemp's list
of Kaffir gutturals viz., " G, g, (like ye Dutch g in groot) ; Q, q, (like ye
English g in great); x , X. (like ye English ch in chlorosis, being ye same as
the Greek x) ; and K, k." Sir G. Greys Library, S. A. Lang., p. 47.

If out of his list of these languages I have shown him to have
fallen into egregious errors in regard to the four first, there is
surely some probability that his remaining ex am pies, based on the
researches of others, will also be erroneous. However, having
endeavoured to prove that these two consonants, distinguished by
him as fricatives in twenty-five African languages, are aspirates
in about one-sixth of the number, I am left to surmise that
they are so in all or most of the remaining instances, and that the
difference between the two terms is, in the case of the gutturals,
merely imaginary.

I have thought proper to subjoin also a compendious abstract,
similar to that preceding, of the gutturals in the Asiatic languages,
as arranged, by the same learned author, under the two heads of
fricative and aspirate, in order to place the matter in another light.

ASIATIC LANGUAGES.

I Fricative. Aspirate.

1. Hebrew: I fortis. lenis. fortis. lenis.

Without points j H

Withpoints i H 3

2. Arabic:

Ancient Literature

Smith and Robinson ..."I ^ ^

Act. pronunciation J

8. Persian:

Mirza M.Ibrahim * 9

i Sanscrit :

Oriental Literature

Bopp, 1833 k

H.H.Wilson j c h fj h

5. Bengali :

G.C.Haughton M </h

6. Zend:

Oriental Literature cV ,

Burnovf ** y h

Bopp,\mz c yh

Broclcliaus kh gh

7 . Armenian :

Ancient Pronunciation ... fr* 1_

Actual Pronunciation...] , ' t

Petermann J

8. Georgian :

Rosen ch gh

9. Albanian :
J. Q v Hah i

Fricative.
fortis. lenis.

X 7

kh gh
kh gh
kh gh

x g

kh yh
ch gh

Asplr<itf.
fortis. leiiia

10. Hindustani:

W Yales .

Oilchrist

H. H Wilson

11. Malayan:

12. Javanese:
J Crawfurd

IS. Turkish:

14. Mongolian :
j j Schmidt

15. Chinese:
Rev. J. Oough \
Rev. T. McClatchie J
Stenhen Endlicher ...

NOTE. Of fifteen Asiatic languages, one is represented as containing
. neither fricative nor aspirate gutturals; eleven as having fricatives of both
quantities, i.e., lenis and fortis; one as having a fricative of one quantity.
Again, two are represented as having aspirated gutturals of both quantities ;
three, aspirates of one quantity. The inference from this is, that two-
thirds of the Asiatic languages have no guttural aspirates.

Counting single instances, the proportion of aspirates to "fricatives" is
as seven to twenty-three In the fifteen Asiatic languages. In the twenty-five
African languages they are as three to thirty-one. The inference must then
be, that more than four-fifths of the rough guttural consonants in all
these languages are fricatives, and less than one-fifth, aspirates ! I admit
the possibility, but not the probability of its truth.

It must strike any one at all given to generalising, that, with
the exception of the Armenian, Georgian, Chinese, the two
Abyssinian dialects, and the Koptic, which are represented as
having one guttural aspirate each, besides the one or two frica-
tives, those languages which are stated as having fricatives, have
no aspirates, and vice versa; an inference which has all the appear-
ance of being based upon a spurious arrangement of instances.
It appears the more evident, from these comparative lists taken

together, that confusion exists in the use of the terms fricative
and aspirate, in regard to the gutturals. It may be gathered
from what preceded them, that the terms fricative and palatal
are also at times used synonymously when applied to that series.
For my own part, I should feel myself lost in a maze of con-
flicting terms of fricatives., aspirates, and palatals, were it not for
the light which appears to me to be thrown upon the subject by
the simple and primitive nature of the phonic system suggested
by the Sechwana.]

Under the Nasals in the guttural series, the Sechwana has
the same element as we find indicated by ng in English in the
word sing, and in German in the word enge. It is indicated in
the literature of the missionaries by n. It may be said to occur
more frequently than any other consonant in this language.

The whole series of simple gutturals, according to the arrange-
ment suggested by the Sechwana, will now be as follows :

Gutturals k

Simple"JExplodents." Aspirate "Explodents."
fortis. lenis. fortis. lenis,

n (-ng)

II. THE LINGUAL SEKIES.

1st Division.
fortis. lenis.

t d

2nd Division.

fortis. lenis.

Nasals.

Ancipit

To the extent above represented, the Sechwana gives the
same linguals, but, as will be seen, contains the elements of a
more consistent and satisfactory series. These consonants have
been variously classified. J"and d have as frequently been called

* It may be observed I have called this consonant a lenis form of k". In
order not to confound it with the mute lenis form g, and till enabled in the
sequel to explain its real nature, T shall indicate it by '</. (Sechwana 17;
German ch, gutt.)

dentals as linguals. In the words of Sir John Stoddart, " the
sound expressed by t throughout Europe is unvaried, and the
same may be said of that expressed by d." L and r have usuallv
been called linguals ; though they are included in the " Standard
Alphabet" under the series of dentals.

The following quotations, from the same distinguished author,
will show that there are few consonantal elements so little under-
stood as the linguals r and 1:

" All the oral consonantal sounds, except I and r, are produced in pairs,
each pair having the same position of the organs, but with a certain difference
of effect." Glossology, p. 127.

The lingual "letters I and r; which, however, do not form a pair, the
articulating action in the one being very different from the action in the
other." Hid. p. 138. In the same page, "That articulations so close aa
those of I and r should pass into each other is not surprising."

" Although all the European languages employ this articulation, and
generally express it by a single letter (either the Latin r or the Greek p),
they vaiy much in the smoothness or roughness of the sound, &c." Ibid.
p. 139.

" It is sometimes difficult to decide whether the sound is to be considered
as a modification of r, or a combination of it with a guttural articulation."
Page 140.

"But though the smoothness of the sound I contrasts remarkably with
the roughness of the sound r, yet the relation of both to the peculiar action
of the tongue renders the substitution of one for the other most frequent-
Hence Molly from Mary, Hal from Harry ; hence the practice called in
Germany lallen ; . . . . and hence, too, the common habit of children, when
they find a difficulty in pronouncing the harsher sound, of recurring to the
softer as a substitute." Page 141.

Ben Jonson, according to Walker, made a distinction between
" the rough and a smooth r ;" however, in English, the phonic
difference is not indicated in the graphic system e.g., r in roll,
and r in directly.*

Two are to be found in French, in the words diriger and
beurre ; in the latter word, a distinction is made by doubling the
letter. It would appear that the Armenian alphabet " has two
different characters, rra and re; the former expressing the

* We often hear children say didectly, and dilectly.

rougher, the latter the smoother sound of this articulation."*
From the tables of Dr. Lepsius we learn that one African lan-
guage, the Susu, has both r and rh ; and that among Asiatic
languages only the Albanian and Hindustani have both a rough
and a smooth r.f No doubt, numerous other instances could be
found, as, for example, the Sechwana, which I have yet to show
contains both forms of r, is represented in the same tables as
having only one. The fact of its relative position among other
consonants, in a universal classification of the elements of arti-
culation, never yet having been decided, will perhaps account for
one form being overlooked in a few languages by some students.
I have preferred to call the whole series linguals.J

NOTE. It will be observed that I have for the present excluded from.the
tecond division of this series all the consonants *, * ; 9, S ; /, v ; sh and zh ; \\
which have usually been called fricative or continuous, and always been dis-
posed into pairs of correlative quantities, fortis and lenis, on the supposition
that the affinity between the instances of each pair is analogous to that
between t and d, or k and g. It will, however, be found in the sequel, that
they are nevertheless included in the lingual series (with the exception off
and v), but by a very different classification, based upon the phonology of this
"barbarous" language. It is worthy of remark, that these several " pairs f 'f of
consonants called fricatives " from their sound being determined by
friction," or continuous " because this friction is not interrupted by any
closing of the organs;" and again, by others, sibilants have also been deno-
minated semi-vowels, " as agreeing .... in presenting only a partial
obstruction to the passage of the vocalised breath ;" but this lias never, till
recently, been tlie case with the gutturals, of tvhich tie have, in the last section,

* Glossology, Sir J. Stoddart, p. 139. In Lepsius's Tables, indicated by
r and r'

f In Hindustani there are three r's, viz., r, rh, and r; the third, from
descriptions, would appear to be a vocalised form of the second.

| 8, T, 5, v, X, p, ff, are called linguals by Buttmann.

As to 9 and 8 of Lepsius, = t h in thin and thine respectively, these elements
do not exist in Sechwana, but are referred to in Section v., Ch. iii.

|| For the positions of * and z in the series, see Chapter v. ; for those
of tli and zh, see Part III. of this work.

1[ As they are usually supposed to be.

treated.* L, r, and n, were also, as before stated, formerly included among
the semi-vowels, by those who found in their " humming and hissing" "a
transition to the vowels; and, in a quotation in the sequel, it will be seen
that the two former are regarded by Dr. Lepsius as partaking of the nature
of fricatives. From the above remarks it would appear that even the terms
fricative and linyual are, in accordance with the views stated in the preceding
chapter, but synonymes in most of the instancesin this series of consonants.

In the above table, t and d form the only pair of correlative
consonants ; their correlation is universally admitted and proved,
as will be seen by an appeal to some of the usual facts in
Sechwana. I shall now attempt to show from these facts that
an entirely different arrangement of the other members of the
lingual series than that in vogue, is required by the inferences
one feels compelled to draw from them. There may be an
appearance of presumption in this ; but when a man takes his
stand upon the firm ground of induction, he ventures less than
by expressing any bold anticipation. Moreover, I may have
something to dread in a charge of hasty generalisation. Be this
as it may, I shall have the satisfaction of having attempted to
supply some of the materials of a natural phonic system, to
which some more comprehensive and discerning mind may be
enabled, by my humble example, to add much that is new, and
to construct a better whole.

Under the First Division of the LINGUAL SERIES the data
of the Sechwana present a slight irregularity in respect to the
pair t and d, but the mutation of the latter to the former is
sufficiently distinct and in keeping with all the instances of
other consonants.

* According to a preceding quotation frorc Dr. Lepsius, the essential
distinction between these gutturals and the above three fricative pairs
" consists in the friction of the breath being formed and heard at the teeth"
in the latter. See Standard Alphabet, p. 45. The only other pair also
considered to be fricatives namely, / and v, could have been included
with the three pairs, so that the gutturals are exceptions to them all, and
their classification together, on this ground, is perfectly anomalous.

Almost the only verb with initial d or lid, as daea, com-
mutes this consonant into t or nt in the examples :

nDaea, (strike,) nteo.*t

But we find also that initial I, and a faintly tremulous r,f are
similarly commutable with the same letter t, in every similar in-
stance of inflection, but t itself remains immutable. The examples
are

Rila, (smear) tilo, itila, ntila.
Leta, (watch) teto, iteta, nteta.
*Tena, (disgust) tend, itena, ntena.

The inference from these particulars is, that the two elements
indicated by 1, r are, as well as d, lenes forms of t, and the latter
their corresponding jforft's ; therefore, both are simple " explodents."
If the mutative process in the Sechwana has any principle in
it at all, the element I, and one of the two forms of r which Dr.
Lepsius has included under the heading ancipites, are thus referred
to their proper places in a phonetic table as both " explodent"
linguals, and both lenes. There can be very little doubt that the
three lenes consonants I, r\ and d, are strictly allied. In the
earliest publications of the missionaries, we find them using d,
dl, where they now use I and r,\\ so that they appear to have

* This diacritical mark, attached to the e as e, is arbitrary, in order to
indicate the equivalent of the French e in Mere.

f The examples in which the verh is preceded by the object-particles are
irregular with, respect to ndaea viz., ititaea and ntitaea. They are, how-
ever, regular with respect to ritaea (strike), which is most common in the
Serolon.

I In the absence of a settled letter, T shall, for the purposes of this
section, indicate this consonant by r 1 .

See Sir Oeorge Greys Library, South African Languages p. 137.

|J Some missionaries, though aware of the peculiarity of the consonant, give
it, in pronouncing the word, the decidedly tremulous articulation of a rough r.

hesitated between four ways of indicating it in the new literature
of the same people.

Molimo (God, impersonal). Liyo (food).

Morimo Dliyo

Modimo

To those arguing on a physiological basis, this fact ought to
savour of a proof that the articulation of the said three consonants
is effected by the same organs of the voice.*

The distinct articulation of d, as I have implied, rarely occurs
in the language. Daea is the only instance in which it takes
any other vowel than t after it, and, in this case, it is generally
enunciated with a faint sound of n preceding it. lit the other
few instances, as well as certain pronouns, it occurs in the form
of a " palatal " ; that is, with a superadded y or i, as Dya (eat),
Dyalia (cleanse), Dyala (sow).f

The well-known consonant I is constant before any other
vowel than i and u, and frequently it is perceptible even before
these ; but that which I have for the present indicated by r 1 , as
above described, seems peculiar to these two vowels, and forms
only one of two elements which have been indicated in the
literature of some missionaries by the common letter r. This con-
sonant (r 1 ) almost defies description. To say that it is an
indistinctly tremulous r, or something between a rough r
and a d, would convey only a slight approximation to the nature
of it. The Greek p, in Kpifiavoq and rapw, which were also written
K\ifiavos and KO\W, the Latin r in meridies, for medidies, and
the Sanskrit r in roman for Ionian, indicated, no doubt, the
identical consonant. If it occurs before other vowels than i and w,

* I cannot divest myself of the idea that, in the Sechwana, the set of
lenes forms of t is incomplete ; that the elemontary forms of the consonants
th in thin (0), and also s in the words parts, par Jcs, harps, also pertain to it.
I shall revert to the subject in Section v., Ch. iii.

f All written by the missionaries without a d, but this element preceding
the y cannot be mistaken.

it will require the ear of a native well practised in observing
differences of articulation to detect it. I am not aware that any
European has hitherto been able to do so in the Sechwana, in
which it must surely exist.

Dr. Bleek, in speaking of the Sechwana " dialects," says of
this consonant

" There is frequently r found in one dialect where the other has I, and
vice versa; and, in general, one is justified to consider r in these dialects as
a sort of floating letter, and rather intermediate between I and r, than a
decided r sound." Sir George Greys Library, South African Languages,
p. 135.

Now, as to either of these consonants being peculiar to different
dialects in certain instances, I rather think this is a mistake ; fyr
both are found in most, if not all, dialects, and are convertible,
as in the particle ri (or li), which is repeatedly pronounced both
ways in the same sentence ; but in many cases they appear to
be constant, as in mocweri (a spiking of water), from cwela and
in the analogous forms.* The value of the above remark further
depends upon what Dr. Bleek considers " a decided r sound"
That r is decidedly" a floating letter" the places of the two gene-
rally acknowledged forms of which have never hitherto been
found in a phonetic scale, is unquestionable, or they would not
be included by Dr. Lepsius under the head of ancipites ! I shall
avail myself of another quotation from the latter distinguished
authority, corroborating these remarks :

" The sounds I and r t participate of both qualities (explosive and fricative),

* Some remarks on the convertible nature of r and I, in Kaffir, may be
seen in " Corr. 8. A. A. B. Society," -p. 104. Mr. Appleyard evidently writes
without any allusion to two r's, and it may be as well to state the probability
that that which the Kaffir substitutes for Z is equivalent to r 1 described in
the text. It is nevertheless possible that the Kaffirs substitute the aspirate
r 3 (described in. the sequel), for even among the Bechwana the children
educated by the missionaries are actually changing the pronunciation of r 1
in morimo to the aspirate r 2 , or its vocalised form, by imitating their teachers.

f There being no diacritical mark, the reader is in doubt which r is
meant, though it is probably intended for the smooth consonant.

being continuous, and at the same time formed 1>y a contact which is
vibrating in r, and partial in /." Standard Alphabet, p. 30.

If we take the Sechwana as a guide, this dubiousness will be
removed. These two consonants are admitted by Dr. Lepsius to be
' explodents] as well as fricatives. I have already shown, in the
preceding chapter, that some fricatives, or, as they have been
otherwise called, semi-vowels, may be simple " explodents," and
others aspirate " explodents." The Sechwana goes still further : it
points to their position in the former set, and gives us their
quantity in the pair viz., that they are both (together with d)
lenes forms of the consonant t. I hope to show, in the sequel, that
the lingual series has also a corresponding aspirate form.

I cannot here forego the following valuable remarks on the
subject of certain forms of articulation, supposed to be peculiar
to the Sanscrit; for, when resolved into their elements, one of
these coincides remarkably with another I have just attempted
to describe as occurring in this "barbarous" but extraordinary
language :

" In the Sanscrit system there are several sounds reckoned among simple
vowels which should rather, perhaps, be considered as combinations of one or
more liquid consonants with a vowel. Thus, Sir W. Jones describes r'i, the
seventh letter of the vowel series, as ' a sound peculiar to the Sanscrit lan-
guage, formed by a (jentle vibration of the tongue preceding our third vowel i,
pronounced very short,' as ' in the second syllable of merrily.' The next to
this is ' the same complex sound considerably lengthened (ree),' and then
follow two others, Iri and Iri, which he describes as ' short and long triph-
thongs, peculiar to the Sanscrit language.' " Glossology, by Sir John Stoddart,
p. 80.

If we regard these articulations properly that is, the lingual
consonants, independently of the vowels affixed to them the
former will be found to resemble very closely the element above
described, which I have indicated by r 1 . Indeed, the r "in the
second syllable of merrily 1 " is perhaps the fittest which could be
chosen to represent as exactly as possible the Sechwana consonant.
It will not be surprising if it should be proved that two such

closely allied lingual consonants as r and / have thus been
reckoned among the vowel sounds by the Sanscrit grammarians,
inasmuch as they have frequently been classed as semi-vowels
by some of our own grammarians, and by Dr. Lepsius, in a pre-
ceding quotation, as partaking of the nature of fricatives ; for, as
I have proved, this cannot be denied of any vowels, all of
which are in the strictest sense of the word also continuous.

Under the Second Division of the LINGUAL SERIES, we find
the Sechwana presents the following data :

Initial r (as indicated in the literature of the missionaries) is also
commuted into aspirate t ( or t-Ji) in the inflected forms; e.g. :

Roga (curse), thogo, ithoga, nthoga.

Again, Initial aspirate t (t\ written by the missionaries th} re-
mains immutable ; e.g. :

Thiba (prevent), thibb, ithiba, nthiba.

The inference which follows is, that the r in this case, which
is a rustling or strongly tremulous consonant* (and which I shall,
for the purposes of this section, indicate by r 2 in contradistinction
from the other), is one of the lenes forms of aspirate t or til, and
that this (th) is the correlative fords ; that, therefore, they form
the pair of aspirate " explpdents" corresponding with the simple
"explodents" just described.

Of this same element, Dr. Bleek (at page 135), in allusion to the

* Dr. Bleek, on the Sechwana, writes: "It has a sound, r, which
is of peculiar harshness, being pronounced deep in the mouth." Sir O.
Grey's Library, South African Languages, p. 1J3. This definition conveys
the idea of a guttural r, whereas that evidently meant is an r accompanied
with an aspirate by which (the physiologist would say) a greater rustling
of the tip of the tongue is produced.

differences between the Serolofi and Setlhapifi* dialects, writes : " As
regards the pronunciation of the words, the main difference between the two
dialects appears to be that, in certain words, a kind of soft r sound is peculiar
to the Serolong, instead of the h found in the Sehlapi. E.g., the Barolong
say tiro (work), for the Sehlapi tiho"SirO. Grey's Library, South African
Languages, p. 135.

It is well known that all the Sechwana dialects (if such they are to be
called) have both h and r 2 . In the example given by this able linguist, of
one tribe having tiro, and the other tihd, the former is derived from lira, or
ri/'a (to work), and the latter from riha (same meaning). In both cases the
initial consonant is pronounced as I at one time, and as r 1 at another, by
the same person in every tribe, whether indicated as I or r in the literature
of the missionaries in any tribe. The Barolofi say both rira and lira.
As to tiho and tiro, it is to be questioned whether the difference is dialectical.
I have shown that, according to the genius of the language, I and r are
both lenes, and hope yet to show that h and r 8 are also both convertible
in the set of lenes aspirates, according to a rule which obtains in the lan-
guage ; so that the difference between the latter couple is, in one respect,
precisely analogous to that of the former. It is a very common thing
for one to hear the Batlhwaro say go, re itse (we do not know), as well
as ga he tits.

The fact of the letter r representing both the smooth and the
rough "explodent,"not only in the writings of the missionaries, but
also in the English graphic system itself, has somewhat interfered
with the formation of correct notions as to its real nature. As

* Some writers treating on South African dialects are prone to a most
inveterate mistake viz., that the publications of the London Missionary
Society at Kuruman are in the Setlhapin dialect. It will perhaps suffice to
say, that this (if intentional on the part of some who, undoubtedly, have dis-
played a little rivalry towards this Society See Standard Alphabet, p. 6,
and Carres. S. A. A. B. Society, pp. 9, 118) is equivalent to a slur upon the
genuineness of the standard its missionaries have produced of a sacred lite-
rature of the Sechwana language, which they have spent some forty years
in acquiring, among the people of several tribes, far in advance of the
Batlhapin, not only in traditional seniority, but also in purity of diction
not omitting the Barolofi. Surely it must be known to such writers, that the
Batlhapui are a people whose language has been deteriorating, by reason
of their close intercourse with Koranna Hottentots, for a century or more ;
and is utterly disregarded by even such missionaries as have resided in their
territory for the last forty years. A large portion of this is occupied by
two sections of the Bahurutse, the most ancient tribe in the nation viz.,
the Batlhwaro and Bachweh.

I have said, the rule which generally obtains is that before the
vowels i and u; the consonants I and r are convertible.* Both
I and an r may be found (in the missionary literature) before
all the other vowels without any interchange that is, in situations
in which we cannot replace one by means of the other ; the only
exception really appears to be before i and M.

It is therefore a matter of importance that, though writers
upon a physiological basis have failed to specify the difference
between the two forms of r in such a way as to leave no doubt as
to their position in a phonetic table, the simple commutation of r
into ih in the Sechwana language, not only before the vowels a, e,
and o, but also before * and u (as in Rita, thito; Rima, thimd'),
should have led to the conclusion that r has a form separate
and distinguishable by the ear from that in Rila, tilo ; Riba, tibo,
and above indicated by r 1 . Practice has further enabled some per-
sons to detect the difference at first hearing.f The result is, that
we are satisfied as to the existence of r inconvertible to I before
the vowels i and u ; and this I have attempted to show is one of
the lenes forms of aspirate t (f or th). As an example among
many of the proficiency which may be arrived at, I give the fol-
lowing words with identical vowel sounds, but different meanings:

Moriri ( a worker), from lira (tiro),] containing both simple and
(Or Morihi) riha(tih&)J aspirate r.

Moriri (a smearer), rila (tilo) containing two simple r's.
Moriri (hair) ? containing two aspirate r's.

* Among others, under the head of Modifications of Initial Consonants,
Dr. Bleek has the following :

I becomes t.

r (before i and u) t.

r (before a, e, o) ih.

In which cases the two letters are intended to indicate the same consonant.
Sir George Greys Library, South African Languages, p. 1 04.

f The two words ruta (teach), and ruta (conceal), are additional examples ;
the latter, which I detected only a few months ago, is as little known to
some missionaries as the former is frequently used by them.

I trust that this beautiful example of the nice shades or differ-
ences of articulation, which I have thus shown may be discovered
by another mode of investigation than a recourse to the phy-
siology of the human voice, will be satisfactory to those interested
in the study of universal phonetics.

The fact of both r's being lenes in different divisions, will per-
haps account for the failure of physiology in helping us out of
our difficulties, as it would scarcely have been practicable to
arrive at the difference between the two forms by any experiments
on the action of the vocal organs. This could not be otherwise
in the absence of unmistakeable rudimentary principles, such as
I believe I have proved the phonic system of this language to
contain.

Under the Nasals, the n corresponding with these consonants is
all that requires to be added to this series, and I am thus enabled
to present the following classification of the linguals, as suggested
by the Sechwanai

Simple "Explodents." Aspirate "jExplodents."

Linguals

1st Division,
forth. lenis.

III. THE LABIAL SERIES.

Nasals.

2nd Division,
fortis. lenis.

In respect to the labials, the Sechwana proves the correctness
of the views generally held as to the nature of the affinity existing
between the smooth " explodents" p and b, the former being un-
changeable in the initial inflexions of the verb, and the latter
E 2

changeable to p. The following examples will suffice in illus-

tration :

Pitla, (rub) pitlo, ipitla, mpitla.
Bala, (count) paid, ipala, mpala.

Upon the principle assumed at the outset, p is thus decided to
befortis, and b its corresponding lenia form.

Under the Second Division of this Series, the facts afforded by
the language present something anomalous to that principle. Ph
(p), which remains immutable in such examples as

Phunya, (pierce) phunyo, iphunya, mphunya.
instead of following the analogy of the other aspirated consonants,
is commuted from h in those of

Hisa, (burn, T*) phisho, iphisa, mphisa.
and is therefore fortis.

But to infer from this, upon that principle, that h. is the lenis
form of aspirate p (pK), would immediately suggest the existence
of a flaw in the system, and invalidate preceding inferences.

It is, however, remarkable, and for my purpose rather oppor-
tune, that this is the only case among the simple consonants
which is not strictly constant. The changes of g* to k] of /to <)
of d, I, and r, to t, and of b to p, are exceptionless ; but this
cannot be said of that of h to p] for we occasionally find h also
commutable to k\

NOTE. It is worthy of remark, that the latter commutation occurs only
in some of those cases in which h precedes the vowel u, and not also before
t, as in the case of the lenis (smooth explodent) r.

1. Sometimes both forms are found, but with different meanings ; e.g

huma (become rich)

' \khumo (wealth)

2. Sometimes both forms with the same meanings ; e.g.

8. In the case of the object-particle preceding, we find :
humisa becoming ikhumisa.
hurisa iphurisa.

Though there are a few instances of h being commutable into Je] I am not
as yet aware of any instance of initial h being changed to .' The subject
of the commutation of the three simple aspirates may thus be explained
at one view :

Initial h changing to th (?) in no instance.

1th (Ar')in a very few instances before .

,, p h (p") in the great majority of instances.

I make no allusion to/ (even though it is an analogous instance) which
prevails in the Sesuto. This is, in a strict sense of the word, a mongrel
dialect, which owes the incongruous position it maintains as a kading
dialect of the Sechwana very much to the circumstance of its having been
reduced to writing and critically cultivated by the " accomplished French
missionaries," and perhaps also to the notoriety of the Basuto nation.
My illustrations of a phonic system, however imperfect, if only correct
so far as they go, are intended to be gathered from the pure Sechwana
dialects alone. The following quotation will serve to show that some
examples from Sesuto are likely to be decidedly foreign to the scope of
my inductions. " The Kaffir/ is generally retained in the Sesuto, and the
Kaffir p becomes / in the Sesuto ; whilst the more western dialects, in which
/ is lost,* have commuted this letter in both cases into h" (Seep. 116,
Sir Qeorge Grey's Library, African Languages.) The remarks about to be
adduced in the text will show that " the western dialects" probably never
had an / to lose, and that h has rather been altered by usage from bh,
both these forms existing in the pure dialects. This is, moreover, proved
by the existence of the initial inflexions ph , iph , mph .

If it should be shown that the change to kh is merely euphonic, inasmuch
as it precedes the vowel u, it will then be practicable to explain the diversity
in the following instances: Mahura (Sehurutse), Mabhura (Sekwena),
MaJehura (Seganano), Mafura (Sesuto. Jf

I have already shown that the lenis form of the aspirate lingual
th ( ^), viz., r] is also convertible to h, the latter being used

* The italics are my own.

f Since the above was in type, it has been suggested to me, by the Rev.
J. FREDOUX, of the Paris Missionary Society, that the euphonical modifi-
cation may be produced analogously to that of ila, kilo, by the apposition
of Jf to the spiritus asper, as in huma, Jchumo, which appears to me to be the
best mode of accounting for the exceptions. The ancient example, oi/x
vTriffTiv, is a precedent, for it would have amounted to the same if the x had
been attached to the v, instead of x to the particle 06 in the Greek graphic
system.

frequently for the former by people of the same tribe. This is
the same with g] the lenis form of kh, for we often hear hae for gae,
fyc. This tendency to pronounce the pure aspirate as if there were
no simple "explodent" lenis attached, in the cases of both /and
gl would suggest the probability that some form analogous to
these two consonants may be found corresponding to the simple
" explodent" b ; and, to conclude that this must be aspirate b would
only be natural. But inference is anticipated by a legitimate fact
which comes to my assistance, in maintaining the consistency of
the peculiar phonic principles of the language. Among some
tribes viz., the Barolon. and the Bakwena, the h is generally*
pronounced like bh in the word hobhouse, with the ho dropped ;
e.g.., sebhuba for sehuba, mabhura for mahura, bhtila for hla, which
has been mistaken by some writers for/f and u.J Whether one
native pronounce it bh&la, and another hela, and a third wela,
the corresponding noun is always pronounced phUd, and the
verb with the object particles iphcla, mphla..

I can only account for the fact of h in some dialects being
almost identified in general use with the consonant bh, in others
by the conclusion above stated, that the tendency is to use the pure
aspirate in place of the lenis form of the aspirate " explodent ;''
but the difficulty is to show why this should be the case with bh
especially, and to such an extent as nearly to lead to the inference

* I say generally, for it has only just occurred tome to set on foot an
examination as to whether natives using bh ever employ the aspirate h
alone in certain cases as a normal form.

f Mr. Archbell. See Sir George Grey's Library, S. A. Languages,
p. 137, &c.

I Mr. Pelissier. Ibid, p. 116, Note.

The apparent commutation of h to tsh (-ckj, e.g., gauhe to gauchwanyane
(or, as the missionaries write it, gaucuanyane), referred to at page 115 of the
same Work, will thus be accounted for by the fact of this adverb being
pronounced gaublie by some tribes. I have heard an individual of the
Banwaketse tribe pronounce this consonant iv* in the same word gauhe, or
gaubhe, i.e., yauitfe, in which the aspirate is retained, and the labial con-
sonant is altered to the labial " semivowel."

that h is absolutely the lenis form of ph(p). But for the exception
of A becoming kh (jfc'}, I should have been inclined to doubt
the stability of the principle assumed ; as it is, the occurrence
of b h, in the form of a legitimate instance, really seems to add
confirmation to it. Though I am not sure that a physiologist
would be able to set one right on being asked to account for the
fact, it is possible the following, suggested by the perusal of
an able author, will amount to an explanation viz., that in pro-
nouncing b, its consonantal element cannot be perceived till the
lips have been re-opened,* and that to pronounce h it is necessary
to open the lips more or less, especially in the case of its occurring
as an initial without any preceding utterance ; so that there appears
to be an organic connexion between the two.

NOTE. It may be as well to append a corollary to the above, that as the
tendency in the Sechwana is to use the pure aspirate for the lenis form of
the aspirated " explodent," viz., h for either g* r or I', a degree of uncertainty
may in some cases present itself on the student meeting roots with initials
in h, inasmuch as the normal form of the initial may be either of the above
three consonants (lenes aspirates) e.g., huma, which is commuted into khumo,
as well as phumo, may possibly be found to have an allied form in guma and
r'uma, as well as bhuma (b'uma), with the same signification.

Including the Nasal consonant m which the language contains,
the following is the classification of labials resulting from the
above analysis :

Labials

Simple "Explodents."
fortis. leni

P *

Aspirate "Explodents."\

lenis. Nasals.

I m

NOTE. So far as I am aware, b" is the only instance of a lenis aspirate-
mute in the language ; d and g, so prevalent in Oriental tongues, have not
as yet occurred to me, though the Bishop of Natal appears to think they
are in the Zulu. The occurrence of at least one form is nevertheless highly
satisfactory ; as the reasoning employed on it may be applicable to the rest.

* Glotsology. Sir John Stoddart, p. 136.

CHAPTER III,

ANALYSIS OF OTHER CONSONANTS.

EXAMINATION INTO THE POWERS OF THE REMAINING LETTERS IN THE
GENERAL ALPHABET OF DR. LEPSIUS. THE CLASSIFICATION
OF SUCH AS ARE REALLY ELEMENTS, AND OF OTHERS THAT MAY
BE SUGGESTED BY THEM UPON PRINCIPLES RESULTING FROM
ANALYSIS IN PRECEDING CHAPTER.

THE train of facts in the Sechwana, in respect of the simple
consonants, having been brought to a close, it now devolves upon
me to deduce from other sources such conclusions as may be of
assistance in arriving at the probable position, in a synthetic
view, of the few foreign to that language. This will involve an
examination into the real nature of the spiritus asper, and also of
what are sometimes called vocalized forms of consonants. It is first
necessary for me to satisfy the reader, as well as myself, of
what, among the large number of remaining letters, are really
consonantal elements or simple articulations ; and, in order to
do so the more effectually, I have, for his convenience, extracted
from Dr. Lepsius's work the Table of

CONSONANTS OF THE GENERAL ALPHABET.

Explosive or I
fortis. lenis.

KvidutB.

nasalis.

Fricati
fortis.
h'h

vce or Continues
lenis. semiyoc.

Ancipitet

k q g

Y fv)

k' g 1

n f

x y

IV. Cerebrates (Indicae)...
V. Linguales (Arabic)..

t ^
t d

t d

n
n

z
z

r 1
r 1

VII. Labiales

P D

U W

NOTE. The letter s, though in the language, is included in this section on
account of its doubtful nature; and the spiritus, also, on account of their
tailing under the " faucales" of Dr. Lepsius.

The letters in bold type are those I have been able to account
for by the analysis in the three preceding sections ; all the rest
will now be considered, in the descending order of the several
series, in the form of interpolated notes. Those in italics (including
(f and 0') are what I attempt, in this section, to prove to have
elementary forms.

I. THE " FAUCALES " OF LEPSIUS.

Explosives. I Fricatives.

fortis. lenis. fortis. lenis.

', h< h

I cannot do better than give the reader, who will, I trust, have the
patience to accompany me in what may appear to be a long digression,
a table of these so-called consonants, as variously indicated by the different
grammarians of the several languages in which they are said to occur.*

The confusion in which the subject of these elements is involved is so
great, that he will require to follow the writer with almost as much attention
as the latter has bestowed in attempting to reconcile^the conflicting opinions
of different authors. Scarcely one cf these has entered upon this intricate
subject without confounding the pure (pectoral) breathings with their con-
sonantal modifications (especially the gutturals, or " tongue-root " letters),
or, in other words, the functions of the " upper or articulating " organs with
those of the " lower organs," the careful distinction between which we owe
to Sir John Stoddart, and which it is necessary to maintain in order to
reason at all clearly upon the subject.

* These I have abstracted, as in other instances, from Dr. Lepsius's
comparative alphabets.

The members of the whole Arabic series are indicated in the " Missionary
Alphabet," of Professor Max Miiller. (See Tableaux, p. xci.) as follows:
'h, ('h); h, h' I ,

1 2 3 4 I 5

But he calls the second (not the spiritus lenis, as does Dr. Lepsius, but)
" the primitive and unmodified breathing," or simple " liquid semi-vowel." He
gives a fifth element, viz., the " primitive breathing," marked by the Hamzeh
" and makes this the equivalent to the spiritus lenis.

Moreover, he makes the 1st and 3rd (viz., and ) both guttural
breathings.

According to Dr. Lepsius, the two pairs are respectively explodent and
fricative. According to Professor Miiller, both pairs &re flatus (i.e., fricatives)
and none are explodents.

When philological Doctors are compelled to differ so much as to the powers
of archaic letters, it seems rather an unfair mode of inquiry to drag in

ASIATIC LANGUAGES.
Hebrew

Explosives.
T3 N

q g

Fricatives.

C.,
-o

h h

h h-
h h

h
b
h

h' h

h
h

h h
nh h
h h

REMARKS.

An h is placed among
the fricative gutturals.
The fortis explodent is
stated, in Lepsius's con-
fronting alphabet, to be
equivalent to his q.

Arabic
Ancient Graph. ...
Smith <& Robinson.
Actual Pronune....
Persian
M. M. Ibrahim ...

Sanscrit, anc. gr. ...
Bopp

H. H. Wilson
Bengali
G. O. Haughton ...
Zend, anc. gr
Burnouf

BrocTthaus

q ...
<1

Armenian, anc. gr ...
Petermann

Actual Pronune...,
Georgian
Rosen

Albanian
J. G. v. Hahn
Hindustani
W Yates

Gilchrist

H. H. Wilson

ai a

these at all from such tongues as Arabic, Hebrew, &c., as equivalents, in
illustration of existing articulations ; and only serves to bewilder. The
confusion which perplexes a student, in examining Professor Muller's
system, consists :

1st. In his making no distinction between the spiritus and their guttural
modifications, e.g., \ and , are called in his Tables a liquida and flatus
lenis respectively, and placed under the gutturals. Elsewhere, the former
is called simply the " liquid semi-vowel," or " an " unmodified flatus," or
" a primitive and unmodified breathing," or " a pure breathing without
even a guttural modification " (!) terms very suitable and definitive ; and
the latter, a similar element, " differing in definition, but identical in pro-
nunciation."

2nd. In his adding another term, viz., flatus, inclusive of all the instances
which are classed under both Dr. Lepsius'a fricatives and explodent-
faucals.

Malayan
J. Crawfurd
Javanese
J. Craufard
Turkish

Explosives.
(a) (a)
a

Fricatives,
(h) h

h h

Chinese
Rev. J. Oough ...)
Rev.T.M'Clatchie]
8. Endlicher
AFRICAN LANGUAGES.
Hottentot

h fc h
h h

x h

Galla
TutscheJt

Bega

Abyssinian, Ge.ez
Ludolph ....!.'

h h

Ditto, Amhara
Isenbera ...

h h

REMARKS.

Mr. Tindall tells us no-
thing about two forms.
Both he and Mr. Knudsen
have an h.

Five other African al-
phabets are represented as
having one h; in twelve
additional cases the h is
included in the guttural
.series.

The above is certainly a most formidable array of instances and autho-
rities. In the case of

8 languages, 11 authorities give 4 members of the series.

1. Explosives. But if the reader will take the trouble to cast his eye
over the various signs intended to indicate these so-called faucal elements,
he will feel bound to come to the conclusion that, at least under the ex-
plosives, they are of a very heterogeneous character. In some cases,
the members of each pair are represented by the vowels M, a; a, a; in
others, by letters usually employed for gutturals Jeh, g ; in several instances
by the mark of the spiritus lenis ; in two by the spiritus asper ; and in others
by signs approximating to a hyphen. In three cases, the fortis explodent is
represented by the letter q, which we otherwise find suspended by the
learned philologer in a rather doubtful position between the gutturals and
these faucals. Of course, without a reference to the works of the authorities
themselves, as to the nature of the elements their letters are intended to
indicate or what would be more valuable, access to the actual pronun-
ciation it is a most difficult matter to enter into a proper analysis of their

relations. At all events, the dissimilarity of the signs employed hy them
must be regarded as an index to the fact of a want of unanimity on the
subject of their real nature.

In the absence of such desiderata, I append the following descriptions of
the " explodent faucals" by Dr. Lepsius, confronted with those of some of
their equivalents by another able authority.

(1) Arabic f, Hebrew N, Greek spiritus "lenis.
~Lmis-explodent-faucal (of Lepsius.) (,)

" By closing the throat, and then
opening it, to pronounce a vowel,
we produce the slight explosive
sound which, in the Eastern lan-
guages, is marked separately, but
not in the European, except in the
Greek. We perceive it distinctly
between two vowels which, following
each other, are pronounced sepa-
rately, as in the Italian sard 'a casa,
the English go 'over, the German
see'adler ; or even after consonants,
when trying to distinguish, in Ger-
man, mein 'eid (my oath), from
meineid (perjury), or Fisch-'art (fish
species), from Fischart (a name).
We indicate this sound, when neces-
sary, by the mark , like the Greeks."
Standard Alphabet, Lepsius, p. 59.

"Among the gutturals, M is the
lightest, a scarcely audible breathing
from the lungs, the spiritus lenis of
the Greeks ; similar to H, but softer.
Even before a vowel it is almost
lost upon the ear PPW, a/to/u), like
the h in the French habit, homme (or
Eng. hour). After a vowel it is often
not heard at all, except in connexion
with the preceding vowel sound, with
which it combines its own (W3?D,
matsa)."* Gesenius's Hebrew Gram-
mar (I4tth Edition) by Rodiger, trans-
lated by Davies. 1846. p. 15.

At the end of a word * * * *
long a was represented by H, and
sometimes by . These two letters
stood also for long e and o."

* " Dr. Lee gives to the Hebrew alif the consonantal power of our un-
aspirated h, as in humble, hour, &c." Glossology, Sir J. Stoddart, p. 128.

Dr. Latham describes alef as equivalent to " a vowel or a breathing."
English Language, vol. ii., p. 88.

Dr. Duff writes: " \ Alif, when beginning a word or syllable, is reckoned
by Oriental grammarians a very slight aspirate, like h in hour. But its
chief purpose is to subserve the expression of short or long vowels." App.
of the Rom. Alph. to the Languages of India, by Monier Williams, M.A.,
p. 88.

Dr. Forbes, in reference to the law of the Arabian grammarians " that
no word or syllable can begin with a vowel," writes, " therefore to represent
what we call an initial vowel . . . they employ the letter \, Alif, as a fulcrum
for the vowel. We have already stated that they consider the I as a very
weak aspirate or spiritus lenis ; hence its presence supports the theory, at
least to the eye, if not to the ear.'' (Hindustani Grammar, p. 17.) " Alif,

Arabic , Hebrew $, Fomis-explodent-faucal (of Lepsius). (?)

"27 is nearly related to S, and is
a sound peculiar to the organs of the
Shemitish race. Its hardest sound
is that of a g, slightly rattled in the
throat, as rnb?., LXX., T6pof>pa
. . . . ; it is elsewhere, like N, a
gentle breathing, as in v?, 'HXl . . .
In the mouth of the Arabian, the
first often strikes the ear like a soft
guttural r, the second as a sort of
vowel sound like a. The best repre-
sentation we could give of it in our
letters would be gTi or rg, as
something like arbaff jt ,
mora." Ibid.*

" The soft sound just described can
be pronounced hard by a stronger
explosion at the same point of the
throat. Thus arises the sound which
the Arabs write . We find it ex-
pressed by scholars generally by
placing a diacritical sign over the
following vowels: a', a, a, a, a;
sometimes below, a. This method
would suppose, from the analogy of
all systems of writing, that the were
only an indication of a change in
the vowel. It is, however, a full
consonant, preceding the vowel. We
indicate it, therefore, with regard to
its affinity to the soft sound, by
doubling the spiritus lenis, ." Ibid.

The above quotations add confirmation to my supposition above expressed)
regarding the indefinite character of the elements intended to be repre-
sented by the signs ; of Dr. Lepsius, an( l * ne nos t of equivalents in the
form of other diverse signs and letters. The Hebrew S = to the former
of these new signs or the spiritus lenis) is generally regarded as a " hiatus

not beginning a word or syllable, forms a sound like our a in war, or au
in haul." (Ibid. p. 7.)

According to Wallin (cited by Max M tiller, " Proposals," &c., p. 29) the
Arabic grammarians look upon the \ as a liquid semi-vowel, distinct from the
\ . The latter writer adds (p. Ixviii.) " where the Arabic \ is used for this pur-
pose (the Greek spiritus lenis) it is marked by the Hamzeh 7 so that this
archaic letter, as stated by Max Miiller, really seems intended to represent
" the primitive and unmodified breathing '* of which the spiritus asper and
lenis in Greek are modifications. Again, " Arab grammarians . . . consider
that a long a consists of the short a ... the pectoral semi-vowel (\)" Ibid.
p. xlvi.

* Professor Miiller makes " the sonant representative" (p. xxviii.) of
(which we shall have to notice under the head of Dr. 'Lepsius'sfricativce ) ;
" but identical in pronunciation " to the liquid semi-vowel by which he
appears to mean either the Elif or the Elif Hamzatum (p. Ixiii.) but surely
the latter, as it is a lenis (spiritus) breathing ; for the former, according to
his own phraseology, is an " unmodified breathing." But he makes N an
element "more pectoral and less modified" than 2, thus reversing the
powers of the Hebrew letters which Dr. Lepsius makes the equivalents of
the Arabic | and respectively. Ibid.

occasioned by tbe disappearance of a consonant." Again, by the above
writer, who would nevertheless call it a consonant, this is virtually described
(if one may judge by his examples), as the articulation or element which
always accompanies the lengthening of a vowel i.e., its long quantity.*
Again, it is described as equivalent to h in some French and English
words, homme, hour, &c., in which the letter is absolutely silent, and therefore
only another mark for the above hiatus. This "hiatus" does not designate
any known sound ; but, in another ancient dialect, the Greek, " under the
name of the spiritus lenis . . signified the absence of a letter, and
became a negative sign in grammatical algebra." An able writer in the
Encyclopedia Britannicaf says further respecting this superfluous letter
" We should wonder the more that a people so intelligent as the Greeks
should have fallen into such an error, if, as far as we know, Lanzi had not
been the first to expose it. His reductio ad absurdum of the spiritus lenis
has not hitherto received the attention which its acuteness merits." Again,
it " is only a mark that the E begins another word, as in the example
KAIEPQ, which is equivalent to KAI EFQ, the sign -/ being equal to the
space between the two words" . . . . If we consider the spiritus lenis
in this point of view, the inventors of it will be exculpated from the ab-
surdity of which Lanzi sought to convict them, and it will attach to those
grammarians only who retained the mark after the practice of leaving a
space at the end of each word became prevalent. In corroboration of this
view, I append another quotation from the learned Buttmann.1 " Both
spiritus are distinct letters in other languages ; the Unis is the alef or elif of
the Orientals. ... Every vowel uttered without a consonant, and,

* To .be satisfied of this, it is only necessary for the reader to examine
all the above examples " Italian sara 'a casa, English go 'over, &c., &c."
The same remark applies to the examples blacking and black ink of Professor
Miiller, of which he says: "In blacking, the voweli is introduced by the
second half of the preceding k ; in black ink, the i is ushered in by the
spiritus lenis." (Proposals, p. xxviii.) But can it be denied that in the former
case the i is without accent, or short in the latter with it, or long ? Had
he said the long i in ink consisted " of the short i + the palatal liquid (s)"
(p. xlvi.), it would have been more consistent with hisprinciple of explaining
the power of the English vowels by those of archaic letters like the spiritus
lenis. Does the learned Professor not rather mean that the power of i in
the English ink is to be explained by an abstract whispered element (call it
accent, the long quantity, or anything you please) of which the ancient
" pectoral semi-vowel " ( ), the " palatal liquid " (s), and the " labial liquid "
(,) are only analogous modifications in respect to the vowels a, i, and u
respectively ? See same page.

f Eighth Edit, vol. ii., p. 613.

J Larger Greek Grammar, p. 14.

consequently, every vowel which is to be pronounced distinctly and sepa-
rately from the preceding letter, is actually introduced by a slight audible
aspiration, which the ancients had greater occasion to make in their writing,
a* they did not separate their words''

Dr. Forbes, in speaking of "the Hamza," as it occurs in Hindustani, in
which it is a substitute for the Elif, writes : " Practically speaking, it may
be considered as our hyphen, which serves to separate two vowels, as in the
words co-ordinate and re-iterate." (Grammar, p. 17 .)

However, it is evident, from all the preceding quotations, in both the
text and the footnotes, that most of the writers to whom I have had access,
except Professor Max Miiller and Wallin, whom he cites, have
confounded the J Elif with the \ Elif Hamzatum, in calling the former the
spiritus lenis, whereas this is the equivalent of the latter element; the
former, according to them, indicating properly the primitive and unmodified
breathing which necessarily precedes an initial vowel, and the latter being
one of the two " modifications of that initial breathing." A little light is
shed on this fact by the remark of Dr. Forbes (Ibid. p. 17), that " the
sound of the mark Hamza, according to the Arabian grammarians, differs
in some degree from the letter |, . . . but in Hindustani this dis-
tinction is overlooked." According to Wallin (cited by Miiller), " \ is a
liquid semi-vowel, distinct from the \. " This " liquid semi-vowel," says Miiller,
" is heard at the end of a long a, as y and w are heard at the end of a long
t and u." In fact, as we may judge from preceding quotations, one of its
principal objects is to subserve the expression of long vowels. By enlarging
on the subject of this " unmodified breathing," it is evident I should be
encroaching on the materials which are to form the second part of this
work on the VOWELS; but, in order to approximate to some correct con-
clusion, it seems necessary to dispel the confusion occasioned by mis-
apprehension of the powers of the above two archaic letters.

Opportunely for my purpose, Dr. J. Miiller, the able physiologist,
most minutely describes the breathing inherent in vowels. He writes
"All vowels can be expressed in a whisper, without vocal tone." These
he calls mute vowels, and adds " But the sound of the vowels, even when
mute, has its source in the glottis, though the vocal chords are not thrown into
the vibrations necessary for the production of voice ; and seems to be pro-
duced by the passage of the air between the relaxed vocal chords." Elements
of Physiology, p. 1046.

Now an important question arises. Does the whispering cease imme-
diately the vocal chords are thrown into sonorous vibrations?* Whether

* At page 1051, he seems to think so of " the aspiration h ;" but this,
we must bear in mind, is one of the modifications of the breathing.

it does or does not, does there not remain a breathing in the former case
only initial (? and terminal), in the latter both initial and continuous (? and
terminal) ? And are we by this breathing, or a lengthened form of it, in
each case to understand the Elif ?* If so, in what does its so called modi-
fication, the spiritus lenis, consist ? Professor Max Miiller gives up the
solution of this in apparent despair. He says (p. Ixviii.), " practically it
seems impossible to make a distinction between the liquid semi-vowel and
the spiritus lenis on any point of articulation anterior to the palatal."
Therefore it will not be presumptuous for me to suggest the query Is the
kamzeh not a symbol indicating the vocalisation of the " mute-vowel ?" It
is remarkable that all the lenes forms of the fricatives of Dr. Lepsius
(viz., x 2> #i 0' t>), and of the flatus of Professor Miiller (viz., z, z,
zh, v), are also considered sonants, or vocalised. Why not apply the same
analogy to the spiritus lenis, or Elif hamzatum, and call it a vocalised
modification of the Elif? Moreover, is called by Max Miiller a
sonant, and Dr. Forbes considers the hamseh " somewhat akin to ,
which its shape (- ) would seem to warrant "f (p. 17); therefore it is also
probably a sonant. If, then, \ is the " unmodified breathing" lengthened,
and is a sonant, is not If = \ + ; viz., the same breathing + voice.
It is remarkable that these two letters were only applicable to vowels. The
probability then is, that both the \ and the were elements accessory to a
" mute vowel," one implying long quantity, the other voice. Apparently, a
ludicrous solution ; but that at which one is compelled to arrive by the
assistance of mere archaeological data. For me to say the Elif hamzatum
7' is equal to j + is only more unsophistical, but surely not more
paradoxical than the following indefinite conclusions of Professor Max
Miiller :-

(1). It " may be true in theory, but is of no practical importance," to
say that J is distinct from f . (P. xxix.)

(2). "The delicate sound of the guttural liquid semi-vowel [J] is in reality
the same as the guttural flatus lenis [~, and both categories may therefore
be represented by one sign." (P. Ixviii.) Again, " the flatus lenis cannot

* In either case we may also ask the question as to the terminal breath.

f Dr. Forbes says that , like the J, is " a weak aspirate, but the place of
utterance of is in the lower muscles of the throat ;" but are we to under-
stand that these are relaxed, or so extended as to allow of vibration ? "It
is," says Shakespear, cited by Dr. Duff, "one of the guttural letters, being
formed in the lower part of the throat. Its sound has been compared to
the voice of a calf for its mother, or to that of a person making some painful
exertion." Original Papers, do,, by Monier Williams, p. 83.

be distinguished in pronunciation from the guttural liquid,* and there can
be no objection to marking both by the same sign." (P. xxix.)

If the " flatus lenis," and the " liquid semi-vowel," " require different
representative types" in these dental and labial modifications, which Pro-
fessor Milller admits (p. Ixviii.), why not also in their guttural modifica-
tions, and especially in their " unmodified," (or elementary) forms ? Why, in
the latter cases, should they be regarded as sounds " differing in definition,
but identical in pronunciation" (p. Ixviii.), and in the former different in
pronunciation ? for in all cases the modifications are consonantal, and two
different elements are modified.

Since archaic letters are inevitably dragged in by devotees to "historical
orthography " to illustrate supposed equivalents in living tongues, it seems
necessary that men should arrive at some understanding as to their specific
powers before venturing to classify them. Throughout this treatise I have
to do, fundamentally, with elements objectively true ; and to reduce these to
subjective principles, which are of course theoretical truths, but only
axiomatical in proportion to the copiousness of the inductions by which
they are evolved. The same process cannot be followed in the matter of
these archaeological instances, till they are known to be equivalents of
standard objective examples.

The preceding remarks all tend to show that the element intended to
be indicated by the latter of the above two signs , (or }), amounts to no more
than an ordinary breathing inseparable from the pronunciation of any initial
vowel (or it may indicate the lengthening of a " mute vowel"), and cannot
therefore be called a consonant. That which is called its fortis form (),
in the second couple of above confronting quotations, is scarcely dis-
tinguishable from a soft aspirate, and is probably only the elementary form, of
which the Arabic sonant is a guttural (consonantal) modification ; that
is. an approximation to the vocalised form of the pure aspirate (spiritut
asper) without the consonantal element of a guttural attached, and which
I think occurs in the Sechwana, in hae for gae, hona for gona, &c., examples
not unlike To^oppa and AjuaXec ; but it is considered by Dr. Lepsius a
harder form of the above " breathing," caused " by a stronger explosion
at the same point of the throat," and of this his digraph (>) is intended to
be the exponent with what propriety will be shown in the sequel.

2. Fricatives. It will be seen, in the preceding table, that, under the
so-called fricatives (\vhich I have previously shown, in respect of the
gatturals, to be aspirates), these members of the faucal series are more
uniformly indicated by the usual marks of the spiritus asper, or this letter
with a diacritical mark viz., hh, h] or h, h, h, &c. As in the preceding

* By this ia meant the " liquid semi- vowel."

case, it will be as well to confront Dr. Lepsius's description of these
sounds with the opinions of other able writers.

Arabic ~, 8 ; Hebrew n; ! h, (of Lepsius.)

The Hebrewletter j-j is described by Gesenius to be, before a vowel,"exactly
equal to our h (spiritus asper)." The Arabic 8 , which the same able scholar
(and Professor Muller since) make the equivalent of this Hebrew letter, is
considered by Dr. Lepsius the same as the lenis h of his system, his de-
scription of which is anything but clear, but which we have a right to conclude
must also be intended for the spiritus asper,or " the common h,"* inasmuch as
" the effect of the Latin orthography upon this letter was to fix it as the sign
of the so-called aspirate.f Since, by all Dr. Lepsius's other lenes fricatives
(viz., x', ~z, z, z, 9\ and v), may be understood sonants, or vocalised con-
sonants, and the Arabic y (his h) is a lenis fricative, it must with him
consequently also be a vocalised element; but he elsewhere calls it an
" unvocalised strong fricative," an inconsistency arising from his con-
founding (like most writers) the " breathings" with " unmodified consonants."

The other fricative member of the faucal series viz., the fortis, or h of
Dr. Lepsius, is stated by him to be equivalent to the Arabic consonant ",
which, again, Gesenius makes equivalent to the softer of " the two grades
of sound" of n, while the Hebrew was a living language. This Hebrew
letter the latter author considers to have been " the hardest of the guttural
sounds." " It is," he adds, " a guttural ch, as uttered by the Swiss, re-
sembling the Spanish x and j." But the h* of Dr. Lepsius is distinctly
described by himself thus : " Not th.e common h, but a stronger aspiration,
which requires a greater contraction of the faucal point, and is distinguished
by the Arabs from the simple Ji."\ This element, which is indicated by

* It is remarkable that, in his detailed description of the faucals, any
distinctive remarks as to the position or nature of the asper appear to be
inadvertently avoided. First it is called an " unvocalised strong fricative,"
as which it must only differ in degree, and not in quantity, from ;
in the general tableau (p. 46), it is placed midway between the position of
the lenis (which is vacant), and the fortis h^ () ; in the alphabetic series
(p. 48), it is distinctly classed as a "fortis fricative ;" again, in the tables
of the different languages, it is placed in the vertical series of "lenes
fricatives."

f " The English Language" vol. ii., p. 103. 4th Edition.

t " Standard Alphabet" p. 39. Professor Max Muller calls it () a
guttural-fiatus-asper. He says, the difference between it ().aud " arises
from the higher or lower position of the point of contact by which these con-
sonants are formed in a Semitic throat [the italics are substituted] (p. Ixx.),
forgetting Sir John Stoddart's rule, that consonantal articulations are
confined to the " upper organs." Elsewhere he says, " the is formed so
low in the throat, that here a contact and explosion would be impossible"
F2

Smith and Robinson, as well as Gilchrist, by the letter h, is only
described by the latter able Orientalist as " rather a harsher aspiration"
than the other h, and " peculiar to the Arabic alphabet, but in Hindoostan
pronounced just as the simple breathing hu."* Again, it is usually laid
down as a postulate in orthoepy, that " no aspirate can be doubled." If
the h of Dr. Lepsius (lenis-fricative-faucal) is only another mark of
the spiritus asper ("), as I have shown, H in the Hebrew, and & in the
Arabic (which he himself regards as ancient equivalents of that mark),
to be, upon the authority of Gesenius and Rodiger, such a thing
as an aspirate aspirate, which If of course indicates, must be an absurdity ;
so that, if there be any peculiarity in the nature (or, I may say, quantity)
of the element he has attempted to describe, there must be wanting only
a more definitive description of it, and a more consistent letter to repre-
sent it.

In the above examination of the so-called faucals, it will be observed that
I have endeavoured to explain away the explosiva altogether, by attempting
to show that the lenls form , Cj*),f is only the well-known spiritus lenis, and
thefortis ()J probably a component part of that spiritus; in fact, that I
deny altogether to the series a division of explodents, in contradistinction
from another, whether called " fricatives" or aspirate.

I am aware that I have only proved my point by displaying to the reader

(p. xxviii.) ; therefore it cannot be a consonant. Again, " the is formed
higher in the throat (!), and occasions, it is said, a friction between the
root of the tongue and the lowest part of the palate." (Ibid.) The fact of
the matter is, the former is a " breathing" confined to the " lower organs,"
the latter to the " upper organs;" the former a strong breathing unmodified
by any consonant the latter a strong breathing modified by a guttural
consonant, forming a "liquid aspirate," or "flatus," or "fricative" con-
sonant, as different writers may choose to call the same element; so that
Dr. Lepsius's definition of it is decidedly more concise and satisfactory.

* British Indian Monitor, vol. i., p. 47.

f The following is from Professor Miiller : " This spiritus lenis is the
Hamzeh of the Arabs. . . . The Hamzeh cannot be called an explosive
letter. Its sound is produced by the opening of the larynx ; but there is
no previous effort to close the larynx, which alone could be said to give
it an explosive character." Proposals, p. xxviii.

I According to Max Miiller, it () bears the same relation to , as the
spiritus lenis T [which I have above suggested includes it], does to the
spiritus asper (*). The same writer says, " there is no tenuis corresponding
to ^ as little as to (p xxviii.) ; therefore, if and } are not tenues, which
Dr. Lepsius decidedly makes them, what are they ?

the confusion which exists on the subject among eminent scholars, who
have given their undivided attention to either ancient graphic systems
wanting in the "living traditional pronunciation,' or to the cumbrous
alphabets of Oriental literature, abounding in redundant letters ; but I hope
in the sequel to sustain this proof by an argument of a more tangible
character.

It will, moreover, be observed that I have, so far as regards the " frica-
tives, tacitly admitted that the strong aspirate which Dr. Lepsius has
attempted to describe may possibly prove to be only another quantitive
form of the spiritus asj>er, or letter h, which is classed by him as the lenis
form under hisfaucales.

Nature of the SPIRITUS.

It is evident that the whole of the above classification of the
so-called faucal series into four members, under the two general
divisions of explodent and fricative, is based upon the assumption
that h is a consonant ; for if a consonant, it must be one member
of a series of sounds which are " explodent ;" and to those who
are determined to maintain that it is one, it will be difficult
to present any plausible argument to the contrary.

I had previously shown, of all the other consonants (except
the two yet to be noticed in the sequel), that the so-called
fricatives, as w*ell as explodents, have aspirate forms ; and that,
in the case of the gutturals, the term fricative of later writers, and
the more common one, aspirate, are merely synonymes ; there-
fore there is a perfect right to assume, that if the faucal series,
with its four members, is a legitimate organical class of the
elements of articulation, it must, like all the rest, either have
a separate and additional aspirate division, or, as in the case of the
gutturals, what Dr. Lepsius calls its fricative forms are merely pure
aspirates. The above examination of the "faucales" has brought
me to the latter conclusion ; but, at the same time, I trust I
have succeeded in showing that, though Dr. Lepsius has not
alluded to the apiritua asper in his detailed description of the

several members of this series, his -arrangement of them into two
divisions is resolvable into the expression thatthere is an"explodent"
form of the spiritus asper, from which I must dissent.

This brings me to perhaps the most conclusive of all arguments
on the subject. When we find such extraordinary consonants as
the four Naman clicks* performing an important part in the dis-
tinctions between the roots of a language e. g., ca (sharp), va
(to slaughter), qa (to spread), xa (to wash),f we need not be
surprised at elements as strange being discovered in other lan-
guages ; and, moreover, when the peculiar mobility of the tongue
is taken into account, it is immediately suggested that any un-
usual variety of them is more likely to be classed under the
lingual series than any other e. g., what are called Naman
clicks, or " Arabic linguals," or " Indian cerebrals," whatever
points in common they may possess to allow of a secondary
classification, could not be otherwise disposed of in an organical
arrangement of the consonants. One has heard of the / in the
language of a Mexican tribe being "purely labial, the teeth
taking no part in it," and of the same description applying to
its sonant form v in the language of Greenland. In all such
examples of labials, linguals, or gutturals, in their mute or
liquid forms, the consonantal element, or fact of a contact
between two organs, is immediately perceptible ; but when we
are told of a series of elements, such as the "faucales," formed
" behind the guttural point, immediately at the larynx," without
the pale of what Sir John Stoddart has distinguished as " the

* In place of any remarks of my own in corroboration, I prefer to quote
the following : " The clicks ought properly to be classed among the con-
sonants, for although they are by themselves distinct articulations, yet they
cannot l>e considered complete sounds without the aid of a vowel."
Grammar and Vocabulary of the Namaqua Hottentot Language, by il.
Tindall, p. 13.

f Ibid.

upper or articulating organs,"* there is an inclination to doubt
the validity of the facts upon which it is established.

Taking for granted that no one will deny all simple con-
sonants to be formed by a contact, complete or partial, of two
organs, and, consequently, by a momentary stoppage of the
breath to the extent of that contact, it follows that the "faucal"
series, unless its members can be proved to be formed by a
contact of two organs, cannot be included among the consonants.
The " contraction of the fauces," to which the formation of these
elements is attributed, does not surely produce an effect similar
to that of a contact.

If, however, in the classification of this series, the gutturals
have been confounded with the ordinary breathings, or the
spiritus, or forcible breathings, of which the descriptions of the
" faucales" bear evident marks, it is only what might be expected
after the terms have been so often confounded.

It is now necessary to come to an understanding as to the real
nature of the spiritus asper. I have before stated, in an attempt
to improve upon the definition of Dr. Lepsius, that the " aspirates"
are those elements, either vowels or consonants, which are pro-
nounced with a simple but forcible emission of the breath ; and
the result of the preceding inquiry into the " faucal " series is,
that the spiritus asper, in apposition with these vocal or conso-
nantal elements, is not a consonant, but merely a forcible
breathing; and, moreover, that it is a distinct element by no
means " inherent in every consonant," and decidedly something
more independent in its nature than a mere "increase" of the
ordinary breath which accompanies the utterance of every

* " Sanscrit grammarians sometimes regard h as formed in the chest
(nrasya), while they distinguish the other gutturals hy the name of tougue-
root letters (^ihvamuliya)." " Proposals," <c., by Max Mutter, M.A., p.
xxx.iv. This amounts virtually to Sir John Stoddari's distinction, who,
had he made such a statement, would no doubt have omitted the word
" other."

vowel or consonant e. g., the formulas t + A, d + h, h + a,
t + h + a, represent combinations of different elements, among
which that of h can be uttered independently of either of the
others, and the others independently of it.

At present, I have only to do with this element in its appli-
cation to consonants, both mute and liquid, as, for the purposes
of analysis, included under the head of " Explodents"

NOTE. I have employed this word (" explodents") throughout the work
as synonymous with consonants, particularly in reference to all elements
formed by a contact of two organs, and a necessary withdrawal of that
contact, principally because it is applied by most authors to only some con-
sonants, and not to others, which, besides the essential character just
described, have the peculiarity of being liquid. Not only Dr. Lepsius,
but Dr. Latham also, confines the term explosive to the former, and con-
tinuous to the latter ; but the words " valvular," and " imperfectly valvular"
are decidedly more satisfactory, and suggest that the terms explosive and
imperfectly explosive would convey similar meanings, but by no means
so definitive. Of the former he says" The action is perfectly valvular
i.e., the breath is absolutely or wholly arrested as long as the parts remain"
in contact. Of the latter, " The air escapes even while the parts ....
are iu contact. The action is valvular, but only imperfectly so." (The
Enylish Language, vol. i., p. Iviii.) It is, however, difficult to say why
the old terms " mute" and "liquid" should be superseded. See the quo-
tation from Dr. Lepsius, on "the sounds r and I," usually called liquids.
This learned writer's remarks on them, in a general sense, are surely
applicable to all imperfectly valvular elements.*

I am here glad to avail myself again of the opinions of Dr. J.
Miiller, the eminent physiologist, viz., his application of "the
mute f sound of the whisper," as he terms it, not only to the
vowels as we have noticed, but also to the consonants. He
writes

*'A main error in many of the attempts at classification of the articulate

* Dr J. Miiller also, I find, calls r and I, as well as s, sJi, ch, and /,
" sounds developed by the valve-like application of different parts of the
mouth to each other." Physiology, p. 1048.

f Here the sense of the word "mute" must be somewhat restricted,
meaning suppressed, silent. Whereas, when we apply the term generally
to a consonant, it means the momentary absence of either sound or breath,
caused by a complete closure of two organs.

sounds, has been the failing to pay sufficient attention to the circumstance
of its being possible to form them without vocal tone, as in whispering;
while to recognise the essential properties of the articulate sounds, we must
first examine them as they are produced in whispering, and then investigate
which of them can also be uttered in a modified character, conjoined with
vocal tone." Physiology, p. 1045. (The italics are substituted.)

Now, the whole tenor of this work is intended to show that the
main error of classification has rather been the failing to detect the
essential property of all consonants, viz., that they are formed by a
contact, whether partial or complete, of two organs. True, some of
them (viz., k, p, t, g, &c.) are " only of momentary duration," and
others (viz., I, r, s,f, &c.) "can be prolonged ad libitum? according
to the complete or partial " occlusion of the faucal passage ;" but
the learned Doctor seems to have forgotten that the latter can
also be uttered so as to be of " momentary duration ;" in fact,
that, like the (ancient) tenues, the "sibilants," or "semi- vowels,"
or " fricativse," or " flatus,'' (as they have been variously called,)
have also tenues forms, and can therefore all be uttered in a
whisper.

I am therefore justified in concluding that it is not only the
" absolutely mute consonants with strepitus explosions," viz.,
k, g, t, d, &c., that " may, by aspiration, be completely changed
to other sounds," but also the tenues forms of the liquids r and I,
and such other elements as I hope to prove are their analogues.
It is only the ordinary breathing which accompanies articulation
in whispering ; and this is proved by the fact that it is quite
practicable to introduce or withdraw the spiritus, as a distinct
element, in whispering, just as in vocalised speech. In the
utterance of the tenues no more than the ordinary breath is
required ; but in order to aspirate them the spiritus is called
into operation. Both mutes and liquids can be uttered in a
whisper, that is, their tenues form ; but both will also allow a
distinct and forcible aspiration of any of these elements in

whispering, viz., a spiritus additive to the ordinary breathing
required in the utterance of " whispered speech," e.g. :

Simple Explodents. Aspirate Explodents.

ORDINARY BREATHING (= whispering Same |
and CONTACT (partial or complete) elements)

Now in the classification of the simple consonants, as suggested
by the phonology of the Sechwana language, we have the binary
quantities of fortis and lenis, the general but unsatisfactory dis-
tinction between which terms has been touched upon in the first
Chapter. It is only necessary for the reader to conceive of the
organs being exactly in the position required for the enunciation
of any element, to enable him to apprehend the result of a forcible
aspiration of the breath in every instance.

In the case of the fortes, in which the contact of the organs is
simply perfect or complete, and the withdrawal of the contact (or
explosion) sudden, the application of the spiritus is simultaneous
and equally short and sudden with the withdrawal.

In the case of the lenes, the contact being more than complete,
in fact, amounting to a pressure,* and therefore proportionably
(though imperceptibly) longer, the amount of interrupted air
exploded on the withdrawal of the contact is necessarily greater

* It is supposed by some that by this greater compression of the organs,
and stronger interruption of the breath, a tension of the lower organs is
caused, and an utterance approaching to vocality, whence arises the term
sonant introduced by some authors, in contradistinction from surd (for the
fortes). Dr. Latham describes it thus : In the case of b, v, d, &c., " over
and above the action of the parts within the mouth, there has been an
action of the larynx, an action by means of which the column of air that, in
the case of p, f, t, &c., was ordinary breath, was thrown into certain vibra-
tions made sonant, so to say." (The English Language, vol. i., p. Iviii.) It
should, however, not be forgotten, that in the classification of consonants
we have nothing to do with vowels or sounds, but only with contacts and
their modifications, by the addition of either the spiritus or the element of
vocalisation, which latter I have yet to notice.

and softer ;* the application of the spiritus is correspondingly less
sudden and softer. In the case of some lenes, however, this in-
terrupted breath or air has a tendency to escape, on account of
the contact, though complete at certain points, allowing of its
permeation at others; in such cases, the application of the
spiritus completely modifies the nature of the liquid, or " imper-
fectly valvular" element, and makes it an aspirated liquid, or
continuous consonant.

It is remarkable that the distinction drawn by Dr. Lepsius
between the common "aspirates' 5 and the so-called fricatives, is
exactly applicable to that above suggested as existing between
the aspirated forms of the " valvular" and " imperfectly valvular"
lenes consonants. He writes " The aspirate can follow the
explosion ; not accompany it through, as it does the friction of
the fricatives." He shortly after says, " the spiritus unites itself
more closely with the explosive letters than any other con-
sonants." f These "other consonants" are of course his "fri-
catives," or "imperfectly valvular" elements, some of which I
have already proved to have both liqiiid (or tenues) and con-
tinuous (or aspirate) forms.

Are there Binary or Quantitive Forms of the Spiritus ?
I have before remarked that the analysis of the Sechwana
consonants, in the second Chapter, afforded a legitimate proof
of all consonants being either simple or aspirate "explodents;" and
it must now be manifest to the reader, that the subsequent exa-
mination of the " faucales " contains no proof whatever that the

* _Z?.>/.,"The musclesof the tongue, aided perhaps by the co-operating action
of those of the pharynx, strike the palate more quickly, and on a narrower
point, in producing the articulation k ; but more slowly, and over a larger
space, in producing g." (Glossology, Sir J. Stoddart, p. 182.) According
to Professor Miiller, " a kind of breathing" continues " after the first con-
tact has taken place."' Proposals, p. xxv.

f Standard Alphabet, p. 4'J.

more forcible breathing, or spiritus, is capable of subdivision
under similar heads ; but the train of the foregoing remarks
would seem to imply that it will, at all events, allow of binary
or quantitive forms, as distinct elements, additive to the strictly
simple consonants, fortes or lenes. However probable I may
consider this, it would be impossible to answer the question by
means of any data from the Sechwana.

NOTE. Tt may not be out of place here to remark that, in the event of a
proof being found of the propriety of applying the idea of binary quantities
to the spiritus itself, which we must bear in mind is additive to both
consonants and vowels, there would be verified the analogy I have already
supposed to exist, in its applicability (quantitively) to both the former and
the latter.

There is decidedly something apparently stronger in the
spiritus accompanying lc t f'p*than in g]c^l>* } but ihefortis nature
of the consonantal elements in the former example, and the lenis
in the latter, are likely to mislead. Still there can be no doubt
that in both cases the spiritus are produced by the same dis-
position of the lower organs. The following quotation seems to
shed some light on the subject.

" According to the Sanscrit grammarians, if we begin to pronounce the
tenuis, but, in place of stopping it abruptly, allow it to come out with what
they call the corresponding 'wind' (flatus, wrongly called sibilans), we
produce the aspirata, as a modified tenuis, not as a double consonant.
This, however, is admissible for the tenuis aspirata only, and not for the
media aspirata. Other grammarians, therefore, maintain that all mediae
aspiratae are formed by pronouncing the mediae with a final 'k, the flatus
lenis being considered identical with the spiritus; and they insist on this
principally because the aspirated mediae could not be said to merge into,
or terminate by, a hard sibilant."* Proposals, <tc., by Max Mutter, M.A.,
p. xxxii.

Therefore, in the following page, Professor Miiller writes

" In Sanskrit no scholar could ever take M for Ji + h, because the latter
combination of sounds is grammatically! impossible."

* Or rather flatus (to be in keeping with his own nomenclature).
t Perhaps the word phonetically is here meant.

In these quotations there seems implied the fact that the
spiritus accompanying the fortes k, t, p, is different in kind from
that in the lenes g, d, b; and therefore the conclusion that it
must exist in binary quantities.

If binary forms be found to exist, judging by analogy of the
combinations in which it occurs, the lenis form will not unlikely
prove to be our common h, as in house (the Arabic 8 , or ancient
"spiritus asper")] arid the fortis is as likely to be that (the
Arabic ) which is described by Dr. Lepsius as formed by a
sudden " contraction of the faucal point," and as a sudden
emission of the breath, of which it is easy to form an idea by a
little experimenting on the aspirate consonants. At all events,
the subject is one worthy of investigation.

NOTE. In the event of quantitation being admitted in the case of the
spiritus, and the above conjectures proving correct, of course this would lead to
the abrogation of a nomenclature which has been rendered sacredly con-
ventional and classical by long prescription e.g.,

Old Nomenclature.

New

Spiritus. 1

Arabic ^
fortis (h)

Asper.
lenis O

So that what has always been called the " spiritus asper" would really be
the spiritus lenis, and the Arabic the spiritus fortis. The
utterance of such paradoxical notions makes one inclined to wince under
the mere thought of the prospective lashes of some critic, which must
inevitably follow the literary temerity of any writer. At all events, I trust
that the unequivocal expression of my speculations will not subject me to
a castigation, any more than the high-sounding term explodent and fricative
faucales, indulged in by another infinitely more able author, in reference
to the spiritus.

* This does not include the ancient " spiritus lenis" which will be again
referred to in the sequel as a third form of the spiritus.

The " Spiritus" influenced by the Position of the Long Quantity,
or Syllabic Accent.

After what has preceded on the nature of this element of
speech, the reader, who is disposed to consider the matter in an
impartial manner, will no doubt allow that it is of far more im-
portance than some able writers are willing to allow it. Dr.
Lepsius says of it

" It is of so little weight, that it does not make the preceding syllable
long." Standard Alphabet, p. 4 ( J.

Zumpt

" It is only an aspiration ; it is not considered as a vowel, and therefore,
when joined with a consonant, it does not lengthen the preceding syllable."
Latin Grammar. Trans, by Sehmitz, p. 4.

Professor Ramsay

" It exercises no influence whatever on the quantity of words, either taken
by themselves or when combined with others, in the formation of a verse."
Latin Prosody, p. 16.

Dr. Latham

" When air passes through both the mouth (or nostrils) and the larynx
simply as so much ordinary breath, the result is the sound expressed by
h. ... It is simply so much ordinary breath expired." The English
Language, vol. i., p. Ivii.

But it appears never to have entered into a part of the inquiries
of philologers as to whether the " spiritus asper" is influenced
by the position of the long quantity (or syllabic accent) ; there-
fore the important part it plays in the apparent modifications of
certain consonants, to which it is attached, has hitherto been
overlooked. I have already shown that the terms strong and
weak, in common with a host of others, have been applied by
different writers to distinguish the difference between the fortes
and lenes quantities of each element ; that such binary quantities
exist, and that of either of these there may be stronger and weaker
forms : but there is no doubt that these terms have often been
employed in the former signification, when they could really
only be explained in the latter e. g.,

" In German, Adelung distinguishes a strong aspiration (hauch) at the

beginning of a word, as in hdbe, have ; and a weaker in the middle of a
word, as in gehen, to go."*

He does not seem to have perceived that in the former case
the vowel to which the spiritus is attached is long, and in the
latter short. In the same manner we say uphold and shepherd,
or vehement and vehicle, but the spiritus is only weak in the
latter instance of either couple, because the accent is not on the
syllable to which it is attached as in the former.f In both cases
of the English examples the quantity of the consonants is the
same, but the forcible breathing, or spiritus, is modified.

I am not aware that any previous writer has taken this view
of the subject; but the Sechwana phonology, which abounds in
aspirates, and forms the basis of inquiry, bears out the truth
of my conclusions in innumerable examples, of which the
following are only a few :

Ldha (pay) Lehile (have paid)

P'ufa (gather) P*ut'ile (have gathered)

P-ap^dma (quiver) K*ok l ola (retail scandal), &c.

In all which cases the spiritus is undoubtedly stronger where
it occurs on the accented syllable.^

NOTE. It is not unlikely it will yet be shown that it is not its position
relatively to certain combinations of consonants which makes a preceding
syllable long, but rather the position of the long quantity which affects such
combinations of consonants. This proved, it will be easy to ascertain the
primitive monosyllabic nature of all tongues. To refresh the memory of
the student who is prompted by this remark to reflect a moment upon the

* Worterb, vol. i., p. 1319, cited by Sir J. Stoddart.

\ The few cases such as honest, humble, &c., do not fall under con-
sideration, because in them the h is absolutely silent.

\ At the same time, it is not requisite to indicate this distinction by
means of any orthographical expedients ; it is only as well to know them,
to enable us to guard against a habit of " hair-splitting," and discriminating
too much.

In ancient prosody, this is especially to be noticed in the case of positio
debilis, where, except in the case of the preceding vowel being naturally
long, as salubris, or the muta cum liquida, belong to different syllables, as
abluo, the succeeding syllable, is as likely to be long as the preceding one.

various ways in which, according to the modern rules of ancient prosody, a
position may be formed (as it is usual to express it), I append the rules:
1. " When a syllable ends in two or three consonants, as in ex, est, mens,
stirps. 2. When the first syllable ends in a consonant, and the second
begins with one, as in itte, arma, mentis, innova. 3. When the first syllable
of the same ivord ends in a vowel, and the one syllable following begins
with two consonants"* (excepting in the case of the positio debilis), as
aptus,factus 4. In the case of i and v, consonants (== y and w) when pre-
ceded by another consonant f alies, aries ; lingua, equus.

A proper examination of the point above suggested (as well as into
the nature of liquids and their continuous forms) will no doubt also shed
some light on " a subject of keen controversy among metrical scholars,""
viz., the quantity of a short final vowel before a word beginning with
BC (=sk), sp, st.

Vocalisation of some Consonants, viz., r and \, fyc. Therefore,
of the Spiritus. Of the guttural provisionally indicated by
gs {German guttural ch).

There is another form of the spiritus, the consideration of
which suggests the examination of a separate element that plays an
important part in the modification of some simple consonants
viz., vocalisation. This element has been variously denominated
" the obscure vowel," " the indistinct vowel sound," and " the
slight and scarcely distinct vowel sound."

NOTE. These terms are sometimes applied to certain unaccented forms
of all the vowels, as in German, lichen ; English, velvet ; Italian, ventura ;
French, t<mir, &c. ; but I shall reserve the consideration of this view of the
subject for the second part of the present work, on the Vowels.

But the element to which I particularly refer is that concisely
described by Dr. J. Muller as follows :

" A peculiar murmuring sound accompanies several consonants, which
does not resemble any of the vowels. This kind of intonating can be
produced either with the mouth open or with it closed, the nasal passage
being in the latter case open." Physiology, p. 1050.

Zumpt'i Latin Grammar, by Dr. Schmitz, p. 20. f Ibid., note, p. 8.

It is not a voioel, though invariably considered one. Till a
better term can be found, perhaps it may be called either the
(e indistinct vowel element" or the " element of vocalisation."

According to Dr. Lepsius, in a quotation of which I have
already availed myself

" This vowel is inherent in all soft fricative consonants, as well as in the
first part of the nasal explosive ; whence all these letters, as g, n, m, appear
sometimes as forming syllables. It assumes the strongest resonance, as
may be easily explained on physiological grounds, in combination with r
and I, which, as is well known, appear in Sanscrit as r and Z> with all the
qualities of the other vowels." Stand. Alph., p. 27.

Dr. Latham writes

" It is an essential condition in the formation of a vowel sound, that the
passage of the breath be parietal. In the sound of the Z'in lo (isolated from
its vowel), the sound is as continuous as it is with the in fate. Between,
however, the consonant I and the vowel a there is this difference with a,
the passage of the breath is wholly uninterrupted ; with I, the tongue is
applied to the palate, breaking, arresting, or partially interrupting the
passage of the breath." The English Language.

The generic nature of l t as a simple (unmodified) liquid, or
" pure explodent"* is, by both writers, entirely lost sight of; one
of its accessory or modified forms or specific natures, as a con-
tinuous consonant, is all that is attributed to it viz., that it is*a
vocalised consonant. It is decidedly a mistake to suppose that
I and" r, and their analogues, are essentially vocalised ; as well
might they be considered essentially aspirated consonants. It
is well to bear in mind that they are, like the other fates, g, d,
and 6, essentially simple " explodents ;" it is only from the fact
of their allowing a simultaneous and partial emission of the
breath that they are liquid, and may be continuous. Viewed
strictly as simple " explodents" (liquids), we have nothing to do

* Dr. J. Miiller says, " Kempelen classes the consonants I, m, n, and r,
among the vocalised sounds, but they are certainly not always so ; they are
heard distinctly as true mute sounds in the vocalised (intonated) speech."
Physiology, p. 1051. He uses the word mute here in the sense applicable
to k, g, &c. My word " exphdent" refers only to contact, partial as well as
complete, and includes liquids ; therefore more suitable.

with their continuous nature i,e., the continuousness of the
ordinary breath required in their enunciation, but alone with the
operation of two organs, with the contact or articulation of
which its escape is simultaneous and equally momentary. Viewed
as continuous consonants, we have to do with the prolation of
the breath alone, (i.e. ' the spiritus) or the " vocal element,"
whenever accompanying it. In the former case we have to do
with the contact so far as it is merely momentary, and from the
circumstance of the action of the organs being to a certain
extent "imperfectly valvular," also liquid; in the latter case,
with the contact so far as this is prolonged, by a continued exertion
of the breath in the form of a " fricative," or " flatus," or by
the vocal element, which implies continuity of breath. In
either case, the consonantal or '"' explodent," or unmodified liquid
nature is not perceived until the organs in contact have been
detached ; but it is only in the latter that the continuous element
is perceptible, for the passage of the escaping and continued
breath is as much as in the case of any vowel parietal i.e., " the
tongue" (in whatever position), " the cheeks, and the lips, are
the walls of the oral passage," and therefore the consonant is
capable of being vocalised.

But, in every instance, it is the breathing which may be
vocalised ; moreover, in order thereto, the breathing must be
continuous ; therefore, vocalisation would seem to imply an excess
of breath beyond the ordinary breathing required in the enun-
ciation of any simple " explodent ;" that is to say, an extra
exertion of the breath, which can be called by no other name
than aspirative.* This is tantamount to saying, that in regard
to some lenes " explodents," viz., the liquids, their continuous! and
aspirate forms are identical ; in fact, that the letters I and r, and

* In some letters this is more perceptible than in others ; e.g., in r and I,
more than in s, th, and/.

t In reference to breath alone.

their analogues yet to be noticed, for example, as pure
1 ' explodents," are not accompanied by any extra exertion of the
breath, though they may be, if necessary ; but in this case they
would become aspirute liquids.

In order to test this conjecture, the nature of the rough
breathing, or fricative element, viz., the spiritus, alone must be
considered, as assumed by every consonant, independently of
voice. When this element is attached to the mutes k, g ; t, d;
p, b; its utterance is simultaneous with the separation of the
organs in contact, as in h] g\ &c. ; when attached to the liquids
(properly so called, viz., r and I, and their analogues), its duration
is simultaneous with that of the contact of the organs, as well as
their separation ; and the longer the duration, the greater the
exertion of the breath, amounting to an aspiration, which forms
a continuous element in the consonant, the " explodent" nature
of which latter is not perceived till the organs are detached.

But the spiritus is not the only continuous element which may be
assumed by a liquid consonant; therefore the term " continuous,"
as at present used, is ambiguous. It may mean, also, the
" vowel element" above noticed ; so that, to be continuous, a
liquid consonant may be either aspirated or vocalised, and
vocalisation is probably only aspiration, modified by an accessory
element, viz., the " murmuriny sound" described by Dr. J.
Holler.

When this "vowel element" is attached to the simple "explodents,"
it merely strengthens the explosion, while it occasions a slight
hiatus,* except that in the case of the lenes g, d, b, the intensity
and comparative prolongation of the contact causes, by the same
effort of the organs, only an approach to vocality. When
attached to the liquids r, I, and their analogues, vocalisation of

* It is not unreasonable to expect the occurrence of such a consonant ;
for, in the Hottentot language, a hiatus is sometimes perceptible between a
click and the succeeding vowel.

G 2

these is the result : " the friction (of the breathing, however
strong*) ceases to be audible, and only the vowel element is
heard" in combination with the " explodent" element. The three
forms of r and I, according to which all their analogues can be
classified, are now as follows:

Lingual I

Simp. Exp. Asp. Exp. Vocalized JExp.
for Liquid.} (or Continuous.)

i r i

r r' rt

I trust that what precedes will amount to a proof that
aspiration is essential to vocalisation, or the utterance of the
" vowel element" (murmuring sound) attached to some consonants.
At all events, unless some proof be alleged, the necessity for two
scales of vocalised consonants (simple and aspirate) cannot be
obviated. The Sechwana language affords only one instance of
a proof, which may not, however, be considered satisfactory to
some of my readers. I have already thoroughly explained the
nature of the two r's in this language viz., the simple (not
continuous) " explodent" r 1 , and the aspirate (continuous) r 2 , or
otherwise r and r; but it is remarkable that it is only the latter
which becomes vocalised e.g., one sometimes hears rema (hew),
in which case continuous breath accompanies the contact of the
organs forming r, and at other times rbma, in which the breathing
is inaudible, and a buzzing sound, arising from vibration of the
tip of the tongue and the exertion of the voice, accompanies the
contact. The r is in this word aspirated, and in the verbal
noun is changed to aspirate t (t), which will perhaps, on the
ground stated above, account for the modification ; but I am not

* Part in parenthesis interpolated. The quotation is from Lepsius.

f Among European scholars, the vocalisation of a consonant is usually
indicated by means of a circular dot below the letter, thus I, r. I trust it
will not betray any desire to differ with those who have introduced it, to
suggest that it be placed at the top, as is usually the practice of pointing in
European graphic systems not formed upon Oriental models.

aware of any instance in the language in which a liquid " simple
explodent" becomes vocalised, except in that of a word like
morimo, as pronounced by native children imitating their
European teachers, who invariably vocalise the letter.

NOTE. The above cited instances in the Sechwana would seem to suggest
that both aspiration and vocalisation are accessory elements, not in the
slightest degree affecting the quantity of the consonant which assumes
them. But as I have, under a preceding head, shown that the spiritus, or
any aspirated consonant, is affected by the syllabic accent, or quantity of
any particular vowel of a word, it may possibly be found to affect the
vocalisation also*

Inasmuch as the free emission of the breath is necessary to
the enunciation of a vowel, and those "explodent" consonants
which allow of a partial escape of the breath may be vocalised,
one would think that the spiritus, which is nothing more than a
forcible emission of the breath, and a continuous element e 'uttered
with the whole oral canal open," must also allow of being
vocalised; in fact, that it is just as possible to vocalise the
aspirate as it is to aspirate a vowel.

NOTE. There must, at this point, occur to the reader some glimmerings
of a proof of what I have above suggested viz., that the spiritus has two
quantities, and this will become more evident as I proceed ; for if it is only
some lenes consonants that can be vocalized, and also the spiritus, it may
be inferred that it is the lenis form of the latter.

Since the other " soft (lenes) fricatives" of Dr. Lepsius are all
vocalised forms, and in his classification A is a soft fricative, and
he admits (p. 39) that there is a vocalised faucal, why not have
inserted this at the head of its fricative analogues? In other
words, if " this vowel (element) is inherent in all soft fricative
consonants," and h (*) is a soft fricative, it must, to use his
own mystical phraseology, also be inherent in h.

NOTE. In coincidence with Professor Max Mliller, he admits the
guttural consonant i (his guttural knis-fricative) to be the vocalised form
of the guttural consonant (his guttural /ortu-Mcative) ; but again, the
former linguist calls the simple breathing the vocalised form (" sonant

representative," p. xxviii.) of the simple but stronger breathing . Now
these are both "faucales" of Dr. Lepsius, and the is one of his
explodents ; therefore in his view it cannot be vocalised. If there is a
vocalised spiritus, and the simple semi- vowel of Professor Miiller is
nothing more" than a vocalised breathing, it is not improbable that , or
the Elif hamzatmn "^ may prove to be it ; in this case it would be the
" sonant representative," not of , but of .
The series would then stand thus :

Spiritus.

fortis. lenis ( : ). lenis ( 2 ), or vocal.

C (ft) * O "" O

The above is a conclusion rather different from that of Dr. T.
Miiller, who writes

" The only continuous consonant (?) which cannot be pronounced in
combination with a vocal sound is the aspirate h. The aspiration of the h
ceases immediately that the vocal chords are thrown into sonorous vibra-
tions." Physiology, p. 1051.

As to the former part of this quotation, I have already
endeavoured to show that h is neither a consonant nor a vowel
but a strong breathing modification of either. The latter part
is equivalent to Dr. Lepsius's words describing the " lenes fri-
catives," z, v, &c., of which he says " the breathing ceases to be
audible, and only the vowel element is heard." Dr. Miiller may
mean by h a forcible breathing, and Dr. Lepsius, by the word
breathing, the ordinary measure of breath required in the
utterance of any consonant ; so that they may refer to different
degrees of one element, to which their equivocal descriptions are
equally applicable. Dr. Miiller also classes z and v (and also
certain forms of r and f) as vocalised consonants, and it seems
difficult to conceive why he should not allow also that there is
a vocalised form of h.

NOTE. There is, for example, a striking difference between the three
opening monosyllables, o, L o, and o, in the line
0)

'O Levery one that thirsteth, &c. Itaiah Iv.
OJ

The ordinary breathing is requisite to the enunciation of every
consonant or vowel (of which in the above line would, if cor-
rect, be an example) ; were it not, we should not be able to
whisper. The operation of the " wind-chest" or trachea, is steady
and without effort,. even during the intonated* speech, when the
breathing is inaudible till any, either mute or liquid, articulations
are modified by the spiritus, when it requires effort, and the
muscles are called into play. The spiritus, whatever the dis-
position of the articulating organs, is distinctly audible when
attached to either vowels or consonants, not only in the
"whispered" but also in the intonated speech, till it itself is
vocalised, when (I admit) the spiritus is apparently suppressed.
However, it is not really suppressed ; but, doubtless, before it
reaches the oral canal, part of its force is spent in keeping the
vocal ligaments in a state of vibration, and, before escaping with
the voice from the mouth, the two together probably cause a re-
sonance in the cavity of the mouth, formed by any particular
disposition of the organs, whether guttural, lingual, or labial. To
borrow the expression of Dr. Miiller in describing certain nasals,
" the cavity of the mouth forms a blind diverticulum" in the case
of every articulation when vocalised. A proof that the spiritus
still accompanies the voice is that, whatever the disposition of
the articulating organs, when vocalised, they undergo vibrations.
So that there may be some truth in my inference that the spiritus
is necessary to vocalisation ; but it must be distinctly understood
that by the latter term I do not mean intonation.

Dr. Lepsius regards z, v, 0' (in this), and z (zh\ as vocalised
consonants, since they are the lenes forms of s,f, (in think), and
5 (s/t). As will be seen in the sequel though it is to be shown

* I use this term instead of Dr. Miiller's, which unfortunately clashes
with the sense in which the word vocalise is here used.

that they are not lenes* forms of s, &c. there is nothing in the
Sechwana to disprove that they are vocalised forms.

I shall proceed to notice the vocalised form of one of those
consonants that have already passed under analysis, which appears
to be the subject of unsettled notions. Inasmuch as the above-
mentioned vocalised elements are by Dr. Lepsius regarded as
analogues of the Danish g (=%' in his system, the softer form of
ch in lachen = x)j ^ must in his view also be the vocalised
element in the guttural series of liquids. As these kindred or
cognate elements, ch (^), and y (%), are by him regarded as the
equivalents, respectively, of those indicated by kh and gh in the
Hindustani of Gilchrist, I cannot do better than describe the
latter consonant in the words of this distinguished Orientalist :

" Kh is the rough guttural It, pronounced in the very act of hawking up
phlegm from the throat, which becomes tremulous and ruffled, while the
root of the tongue is with it forming the sound required. This letter is
familiar enough to the Scottish and other northern nations, but very
troublesome to the English, &c." British Indian Monitor, vol. i., p. 12.

There can be very little doubt that the consonant above
described () is the ch in the Scotch loch, that is, equivalent to
ch in German lachen viz., the of Lepsius.f Of the other
(), Dr. Gilchrist writes :

"Gh is * * * * the guttural Northumberland r, heard in the act
of gargling the throat with water." Ibid. p. 12.

In this description Dr. Duff coincides with him,
Elsewhere he writes of both elements : . r

" The true discriminative articulation of kh and gh depends on ruffling
the throat in a particular manner, while prolating k and g respectively."
Ibid. p. 20.

It must be evident to the reader that, in quantity, the articu^

* I.e., related to s, as t to d.

\ Dr. Forbes says " ^ has a sound like ch in the word loch, as pro-
nounced by the Scotch and Irish, or the final ch in the German words
schach and buck." Hindust. Grammar, p. 5.

lations are identical, but that the former is aspirated, and the
latter a vocalised liquid.*

Though Dr. Gilchrist has most distinctly stated that the
consonant indicated by his digraph gh is the Northumberland
burr, or r, the latter sound is separately included in Dr. Lepsius's
system as r, or the guttural r. Now Professor Max Miiller
writes :

" The English and the German r become mostly guttural, while on the
other hand the Semitic guttural flatus lenis fricatus . . (i) takes fre-
quently the sound of a guttural r. It might be advisable to distinguish
between a guttural and a lingual r ; but most organs can only pronounce
either the one or the other, and the two therefore seldom co-exist in the
same dialect." Proposals, p. xl.

But he elsewhere calls " the sonant representative" of
(p. xxviii). The majority of writers concur in making the
Oriental an element as nearly as possible equivalent to what
is often called " the guttural r" or burr of some dialects. Among
others, Silvestre de Sacy as cited by Garnett. The latter able
linguist himself says of it, " the sound meant for r has no lingual
vibration at all, but becomes a deep guttural . . . almost
exactly corresponding to the Arabic ghain"^ If so, it is not
only advisable but absolutely necessary to distinguish between
it as a guttural sonant (vocalised guttural) and any lingual sonant,

* Dr. Forbes says " e. has a sound somewhat like g in the German
word sagen. About the banks of the Tweed, the natives sound what they
fancy to be the letter r, very like the Eastern ." Hind. Gram. Professor
Miiller makes the g in German tage the sonant corresponding to the German
guttural ch in loch (which latter, however, he considers the equivalent of
the Arabic ^ already referred to as a different element from ) ; but
Germans, whom I have tried in this country, differ so in pronouncing the
g in tage, that it cannot fairly be taken as a type. If this letter, as known
to Professor Miiller, is equal to the g in sagen, and this again approximates
at all to the g in Cape Dutch dagen, which is only a very mild form of the
"guttural r," or Northumberland burr.it is very probable that these are
all examples of variations which " exist only in degree."

f Philological Essays, p. 253.

viz., between r and what I shall provisionally indicate by </" in
my orthographical nomenclature a letter bearing the same relation
to cp as r to K

I have thus attempted to show the "guttural r" to be the
vocalised element in the guttural series; at all events, it is
requisite that I should account for the nature of the Danish g>
which I propose to attempt in the sequel. The Danish g (y or ^')
is in the same system, but I think erroneously, represented as
the equivalent of Dr. Gilchrist's digraph gh (Arabic ).

The Nasals and their Vocalisation.

It is under the head of vocalisation that it becomes me to take
into consideration the subject of the nasal consonants, properly
so called. M and n have always been regarded as analogues of
I and r that is, as liquid consonants or semi-vowels. Though to
this day schoolboys are so taught, and an authority like Dr.
Latham classes them as such, they are by linguists of the Con-
tinent, and others,* classed separately as nasal consonants, and
as analogues of ng in English king, or German enge. Again,
just as m and n have by long prescription been considered liquid
consonants, the element usually indicated by ng in the same two
examples has been exclusively pronounced the nasal n. So
recently as 1855, the above able writer gives the public the
following conclusive remarks on the subject of this element,
which he considers the " English representative" of a class, but
does not acquaint his reader with a single additional instance of
other members of this class :

" The nasal sounds are vowels, so far as the actions that form them are
parietal.

" They are also vowels in some of their other properties e.g., they can
form syllables by themselves. In the Chinese, such syllables actually
exist, constituting monosyllabic words.

* Among the rest, Sir J. Stoddart. " In Hebrew, Greek, &c., it [n] is (as
I think improperly) reckoned among the liquids." Glossology, p. 142,

"They are not, however, vowels, in respect to their power of combining
with other sounds e.g., b-ng is not a syllable in the way that ba or bo is one.

" Nevertheless, the nasal sounds are essentially vowels, though whether it
may be convenient to call them so is another question. The details of their
mechanism and classification have yet to be studied, and, as they are rare
in our own language,* it is not likely that any Englishman will be the
successful investigator. The French and the Portuguese have the best
means of studying them. Neither have the muscles of the nares and soft
palate been examined, with any view towards the phonesis of what we have
called the nasal passage.

" Nevertheless, the ng in king is more of a vowel than aught else."
The English Language, 4th Ed., vol. i., p. Ivi.

The following abstract from the valuable tables of Dr. Lepsius
will show that what is properly called a class of nasal vowels has,
probably, been confounded with this imaginary plurality of nasal
sounds so called, of which the single example of ng in king is
considered by Dr. Latham the " English representative."

TABLE OF " NASALS."

Standard of Lepsius

Vowels.
a e i o u

Consonants Outt.
n

Hottentot

Wallmann

a e i 6 u

Sanscrit

an en, &c

Wilson ....

Bengali

onpf

Zend

a n

an, n

Albanian
Hahn

u n

Hindustani
Yates

aiichrist {[HI ;;;;;;

Wilson

an
n

Chinese

a.u

* This cannot, therefore, include m and n, which prevail in the English
language.

Now, under the Hottentot, to my own knowledge, there can be
no mistake as to the existence of the nasal vowels, whether
written by Mr. Knudsen a, e, &c. ; by Mr. Wallmann, or, more
recently, Mr. Tindall, as a, e, &c. ; or by Dr. Lepsius as a, V,
&c. ; but whether the guttural consonant ng exists in the lan-
guage I think there is some doubt.* Mr. Knudsen and Mr.
Wallmann, on the statement of Dr. Lepsius, both write it, but
I have not succeeded in finding that this is supported by the
authority of Mr. Tindall. It is remarkable that in such an
example as \cangJia (smoky), where the lingual n occurs before
the liquid guttural gh (German ch), it remains a lingual ; that is,
it does not become cang-gh, or, as Dr. Lepsius would write it,
can^a. In his Grammar he speaks of " a final vowel, which
appears to have the ringing sound of ing, as in ring, sing, &c. ;
but which is not sounded with sufficient distinctness to warrant
our adopting the same orthography as in English." He, how-
ever, suggests that it will require Dr. Lepsius's " w" to meet this
case not, perhaps, aware that this pointed letter is intended to
be confined to the English or German articulation, as indicated
by the digraph ng in ing, or enge.

It is not improbable that the same confusion which exists as to
the fact of the consonantal articulation ng, and the nasalized
vowels a, e, &c., both being found in the Hottentot, also prevails
among some of the linguists quoted in the above table, in
respect to the nasals to which they allude. I have placed at
the head of this table Dr. Lepsius's standard nasal vowels, and
classed under them what he considers as their equivalents in
the writings of those linguists. With the same view the instances

* Since writing the above, I have heard it in the Grikwa (impure Hottentot)
sxpression, to eibe qong (don't go just yet), in which q represents the
' cerebral" click.

| This letter here represents a click.

are shown in which, according to them, the guttural nasal
consonant is found in some of the same languages.

Of these, the Hottentot is said to have that consonant ; but this
has just been questioned in the preceding remarks. Again, the
Hindustani is represented as not having it, whereas Dr. Gilchrist*
gives it in such words as hongeen (will be), Gunge (O Ganges !).
In fact, no better example could be produced of the manner in
which correct views of the phonical nature of elements have been
sacrificed to these graphic differences. But in the former word
occurs also his " nasal h ;" this he describes as equivalent to that
in the French word bon, which, of course, an Englishman, to be in
keeping with his own graphic system, would have to write ~bong.\
Here we find a fresh source of confusion. It is evident that
Dr. Gilchrist's n (or n of 1796) is not the nasal vowel, as which
it is classed by Dr. Lepsius, but the proper nasal consonant ng in
king (n of the "Standard Alphabet"); for there is surely no
distinction between ng in king, and that in long or bong (Fr. bon).

To help myself, as well as the reader, out of this confusion, I
shall quote Dr. Gilchrist's description of the letter n in Hin-
dustani :

" n, as a nasal before j, k, g, and t, or d, requires no particular mark>
sounding exactly like our own letters nj, ng, nJt, nt, &c. in change, rung,
sunk, want, &c., but elsewhere it is the French nasal when marked n.
Preceding the labials it becomes, as in most languages, m." British Indian
Monitor, vol. i., p. 9.

* See British Indian Monitor, vol. i.

f Unless a Frenchman nasalizes the vowel, as he doubtless does in
some instances; but this would not alter the consonant. The vowel is
also modified, and requires a separate diacritical mark thus (French) sun
(to borrow Dr. Lepsius's orthography). Perhaps it is in allusion to such
an instance that Dr. J. Muller says, rather arbitrarily, " the ng of the
French language is formed still deeper in the throat !" (p. 1048.) He after-
wards (p. 1052) refers to the frequent use in French of these three con-
sonants, but especially ng, " in constant combination with vowel sounds of
nasal timbre, to the exclusion of other vowels not of nasal character."
However, I observe the n in bon (French) is given also by Max Muller as
the equivalent of ng in sing (English).

Now, in the following combinations

nj nk ng nt nd,

the n, separated from the post positive letter in each case, is in nj,
lingual; in nk and ng, guttural; in nt and nd lingual.

The example = nj in change, would be called by Max Muller, an d
many other Oriental linguists, a palatal, as in his examples inch, injure ;
but he does not deny, and I question whether any of them would, that the
tongue is the active agent in the formation of the palatals so-called, as well
as in the linguals. Besides the three common organical classes of con-
sonants, Professor Muller has two others viz., palatals and linguals
(wrongly called cerebrals, but, he says, properly cacuminals), which he calls
modifications of gutturals and dentals respectively. Merely to preserve
analogy or consistency in a tableau, or, may be, in deference to Para
Brahma himself, these must, forsooth, have their nasal exponents. Professor
Lepsius, besides these, has a third additional class viz., the linguales
(Arabic), to which he applies the same principle of a forced analogy.
See his Table, " Consonants of the General Alphabet"

To his fourth additional class, the faucales (I. in his Table) of which, as
consonants even theoretically true, I have attempted to show the absurdity,
he fortunately has no nasal exponent. A plausible law seems to have been
snatched at in every case viz., that " the peculiar character of a nasal is
determined by the consonant immediately following." (See Haughtons
Reasons for so many Indian Nasals, cited by Monier Williams, p. 81.) As
it was maintained by the Sanskrit or Arabian grammarians that there are
differences in the consonants, and inferred that there must be nasal ex-
ponents corresponding to these differences, it seems to be expected by
modern Sanskrit scholars that what was considered at least theoretically
true in Arya-avarta, or at Mecca, will probably turn out to be practically
true at Timbuktu, or in Bushmanland. However, leaving the so-called
nasals of the Sanskrit and Arabic linguals to their respective sections, I
shall here only refer to the so-called palatal-nasal. In addition to the
above law, which I admit to be true in the majority of instances, but not in
all, another seems to have been long in vogue to pervert the notions of the
linguist in respect to this imaginary element viz., that a palatal is a
simple consonant. Were it one, I admit there would be some reasonableness
in seeking a nasal exponent for the series; but emphatically deny that any
palatal is a simple consonant. By going thoroughly, i.e., inductively, into
the subject here, I should be anticipating the matter of the third integral
portion of this work, as well as a part of the following section. Reference
to a single example, which I cull from Professor Muller himself, will suffice.
He says" What we call a palatal n is generally not a simple but a com-
pound nasal, and should be written ny'" (p. Ixvii.) Why, then, include it

at all in a series of simple consonants? It must be> evident to the reader
that it is not the n that is palatal, but the combination ny, which is a
palatal analogue of Ay or ky, or any other letter or combination of letters
with a superadded y, and that my and y. y are articulations quite as palatal
as ny. In fact, combination of a peculiar kind (not of consonants alone)
is essential to palatalization. Again, as is well known, the three common
nasals m, n, and ng, exist by themselves, i.e., independently of any other
consonant; whereas Professor Miiller himself says elsewhere, in treating
of the palatals, " the nasal, again, hardly exists by itself, but only if followed
by palatals" (p. xxxvii.) ; therefore it ought at once and for ever to be
omitted from the table of simple consonants.*

In reference to some of these palatals, Dr. J. Miiller, the eminent
physiologist, says "The,;, soft g, and ch of the English language are also
compound sounds. Thej and the g being pronounced like the French j in
' jamais,' preceded by d ; the ch like t followed by sh," (p. 1052). In fact,
the initial of most palatal " sounds" can be resolved into a common lingual,
of which there is a nasal exponent ready to hand, without resorting to
either the Devanagari or Arabic for any of their superfluous letters.

Now, if guttural in nk and ng, it must, independently of the
post positive consonants k and g, be equivalent to the nasal
consonant ng in question. I have shown the French n, in some
cases, to be a consonant = ng in king, or German enge.
Assuming this digraph to indicate this nasal consonant, rung
would be correctly written, but sunk would require to be written
sung-k; therefore, in the above examples of Dr. Gilchrist,
hongeen and gunge, these would be written -hong-geeng and
gung-ge ; so that his French n and his n preceding any gut-
turals are identical elements. If his nasal n (or an) were a
vowel, and his ng = to the same digraph in the word king,
Dr. Lepsius would write the above two words hont ; but if his
ng were = to ng-g, the same linguist would write the word
hohgt.

* In the face of these facts, the learned Professor does not make the
simple guttural-nasal ng a base letter as common with m and n, but gives
this favoured place to the so-called palatal nasal ny (French signe), and
actually calls ng a " modification of the second degree" (!) See Table, p. xci.

The conclusion to which we must come is, that in the above
table we have a medley of instances, culled indiscriminately from
the works of several authors, in which the three separate
elements (1) a, the (mean) nasal vowel; (2) h, the nasal guttural
consonant; and (3) the latter combined with the kindred guttural
g, i.e., h-g, require to be distinguished;* that is, some of the
instances will, perhaps, on proper examination, be found equiva-
lent to ng, in which the kindred letter g is distinctly enunciated,
as in the word English ; some, to n (or ng in king, German enge,
French n in bo?i) alone, without the kindred element; and
some as real vowels, the pure and indisputable forms of which
are to be found in the Hottentot (a living member of the Egyptian
family of languages) and Chinese.

If it is by these dissimilar examples that we are to understand
the class of nasals of Dr. Latham, of which he considers the
English ng (n of Lepsius) as the representative, I trust that I
have proceeded far enough to show the impropriety of such a
classification, and that this element is rather a distinct member
of another set of nasal consonants, viz., the analogue (guttural)
of the lingual n and the labial m; a statement which admits of
abundant proof in what follows.

Dr. Latham writes :

" We cannot close the nostrils, as we can the lips, by the action of their
own muscles. Neither is there such an organ as the tongue in the nose.
If there were, we might form as many sounds through that organ as we do
through our mouth. As it is, however, all that we can do with a column of
air passing through the nostrils, is to narrow its line of exit by contracting
the passage. ,f The nasal and palatal muscles allow us to do so. They allow
us to bring the walls of the nasal cavity a little nearer each other, or to
separate them a little farther from each other. They do not, however,
allow us to close the passage altogether," &c. The English Language.

* I have, for the purpose of this distinction, made use of letters in the
" Standard Alphabet" of Lepsius.
| The italics are mine.

The above appears to explain exactly the nature of the nasal
action in the formation of the narisonant vowels. There is no
contact of organs in the faucal passage, this being only more or less
narrowed, as in the case of any ordinary vowels ; the nasalization
of the vowel is caused entirely by a change in the compass of the
nasal passage, whereas with the nasal consonants m, n, and j^*
there is a contact of organs in the faucal passage and a perfect
closure of the same, but no change in the nasal passage.

Just as n and m are formed, respectively, by the tongue with
the palate, and by the lips alone, so g. (ng) is formed by the
contact of the same organs required in the formation of its
kindred forms k and g ;f so far they are simple <f explodents."
They differ from the other liquid consonants, I, r, &c., in that they
are all mutes ; that is, the faucal passage is closed by the contact
of the organs J the action of these organs is "perfectly valvular"
their closure is complete. Nevertheless, they are analogous to

* In order to avoid as much as possible the use of diacritical marks in
distinguishing the timple consonants or elements, I shall in future employ
this letter in place of either the Sechwana n, the of Lepsius, or the
digraph ng. I believe it was introduced by Pitman in England, and I
have observed that Mr. Appleyard and Mr. Grout, both able linguists, have
suggested the use of it in a general South African alphabet. As the
original character of the nasal is often determined by that of the simple
consonant immediately following, it is supposed by some linguists that it
"requires no modificatory sign ;" but the rule is not universal, eg., the
nasal above referred to in Hottentot cangha, Sechwana sip-tie, and in
Zulu words, -mnandi, -mtoti, -nwana, without their prefixes, show examples
to the contrary ; but when the two accessory elements of aspiration and
vocalisation are taken into account, as requiring diacritical expedients to
mark them, the importance of being particular in giving each its fixed letter
cannot be overrated. Two of the above Zulu examples, I suspect, ought to
be written mnandi, mtoti, or in any way that will show the m to be vocalised.

f That is, in the words of Dr. J. Miiller, <( by the application of the
dorsum of the tongue to the posterior part of the palate."

I This is admitted by Professor Miiller (p. xxx.), and by Dr. J. Miiller;
but, strange to say, the latter nevertheless denies that m is a labial con-
sonant" (p. 1047). See, in self-contradiction, his remark on Kempelen,
already quoted, p. 81, in which he calls it a " true mute sound."

the kindred faucal consonants (liquids), in that while there is a
contact of the same organs the breathing or vocalisation finds a
way of escape. In the case of the liquids, it escapes according to
the position of the organs in imperfect contact ; in the case of the
nasals, it is withheld and " retroverted " to the nasal aperture,*
which undergoes no alteration. Like the liquids, they must
then, necessarily, admit of vocalisation as well as aspiration.
We have thus presented the three possible forms of each organic
series.

Simp. Asp. Vocal.

Guttural g g }|

Lingual D. n H

Labial m m in

These all abound in the Sechwana language, with the excep-
tion of the aspirate forms, of which there are a few, but they
do not just occur to me. The following words, similarly
arranged, will suffice as examples :

p ata (a sheaf) phe (a species of ground-squirrel)
nala (loiter) ntsa (produce)

matla (gallop) tsama (a walking-stick)

NOTE. (1). The object-particles m, n, and y. (= me) of verbs are in-
variably vocalised in the Sechwana language, e.g., mpona, nnama, pkhatla.
(2.) In the case of the prefix mo being attached to the initial cognate
element b, which by synaeresisf becomes m, e.g., mobala (colour), is always
pronounced mala; mobutla (hare), is always mutla.

* Dr. J. Miiller thus describes them : " In the pronunciation of all the
three consonants of this order, the cavity of the mouth forms a blind
diverticulum of greater or less length from the throat and nasal canal.
This diverticulum is largest during the utterance of m, smallest when ng
is being pronounced (p. 1047). I have already copied this phraseology in
describing the formation of the other vocalised liquids I, r, &c., without
at the time noticing, what I now observe, that the vocalised nasals are in
this respect also analogues of those (proper) vocalised liquids.''

f It is probable that the initials n and y. are in some instances

(3.) In all cases like mohago (road-food), becoming mhago ; mo henya
(conquer him), becoming mo henya ; but iu examples such as the latter the
words are pronounced both ways. In this language these three nasal
consonants form initials independently of any other consonants
nam, mothu.

II. THE "PALATALES" (OF LEPSIUS) AND THE LETTER "q."

Next in order in the above General Alphabet of Dr. Lepsius, we find
the letter q, and (omitting the mute gutturals which have been fully under
consideration) a series answering to the gutturals, but with a diacritical mark
attached to distinguish them. Professor Miiller also says "Palatals
are modifications of gutturals, and therefore the most natural course would
l>e to express them by the guttural series, adding . . any . . sign to
indicate their modified value" (p. Ixi.)* Though acknowledged to be com-
pound sounds in the pronunciation of Europeans, Dr. Lepsius, with an
over-scrupulous regard for the graphic element in the sacred Davanagari
writing of the Indians, must, forsooth, pronounce them to be simple sounds,

analogously formed. See Kellic permutations referred to in sequel (Sec. ii.
ch. iv.) The above is perhaps the phonical process called eclipsis by
"the Irish grammarians," e.g., mbaile (town), pronounced matte. See
Garnett, p. 81.

* The palatals are indicated by the gutturals Jc, Ich, g, &c., modified
thus

By Professor Lepsius k f kh' g> &c.

Miiller k kh g &c.

With regard to the flatus or fricative form of these palatals, Dr. Lepsius
adheres to his graphical rule by borrowing the Greek letters thus x'\ x"
but Professor Miiller wanders from his own by introducing the sibilants
* and z italicised. The following is a list of some of the " sounds" com-
monly comprehended under the term palatals :

i ch in English church, German rutschen, Italian ceci.

jjin English join.

(sh in English sharp, French oh in chose, German sch in scharf.

tj & g in French joli, genou, English s in vision.

Ich in German ich.
g Jconig.

g taglioh.

gn in French besogne, &c., &c., &c., (= ny.")

Its imperfection will account for the amount of unsuccessful speculation
on the subject of these " sounds."

100

because in that system represented by simple signs. " This is, moreover,"
he says, " proved by their not rendering the preceding syllable long, and by
the possibility of doubling them." Professor Miiller, in other words,
says " All (the French tch, the Italian c, and the Russian *[), even the
German tsch, are meant to express simple consonants, which, with the
exception of the tenuis aspirata in Sanskrit, would not make a preceding
short vowel long" (p. Ixi.)* I have already alluded to this subject in the
preceding section, in attempting to show that there is no such element as a
separate palatal nasal, and shall therefore not attempt to anticipate by
more than a few remarks here, what will be fully treated on in its proper
place.

Just as " the name of x (chi) connects it with the vowel i" in the minds of
such as have given more attention to graphic differences, so the letter q,
from its Hebrew and Greek names qof and Jcoppa, seems to be connected
with the vowel o ; but it is necessary to disabuse the mind of the " syllabic
power" originally attached to such letters, and resolve them into their
elements. The former is merely an instance of a guttural being prefixed to
the prepositive vowel of a palatal diphthong, and the latter (in later ortho-
graphy) of the guttural k being attached to that of a labial diphthong. On
this account, neither they, nor any analogous syllabic combinations,! can
be said to fall under a classification either of the simple or compound con-
sonants, and are, therefore, reserved for the third part of this treatise.
Professor Miiller again says " Although, therefore, we are forced to
admit the palatals, as a separate class, side by side with the gutturals,
because most languages retain both sets, and use them for distinct etymo-
logical and grammatical purposes, still it will be well to remember that the
palatals are more nearly related to the gutturals than to any other class,
and that in most languages the two are still interchangeable" (p. xxxvi.)
It is to be hoped that before the writer has time to disprove this phonetical
tenet, of which some European linguists seem so tenacious, they will
anticipate the numberless proofs to the contrary which he hopes to produce
from a barbaric dialect viz., that " palatals" are as much modifications
of both dentals and labials as of gutturals, and that they are in no instances
simple consonants.

* He, however, elsewhere says " Frequently the pronunciation of the
palatals becomes so broad that they seem, and in some cases really are,
double consonants." (p. xxxvii.)

f There is very little doubt that all " sounds," called by Dr. Latham
" Unstable Combinations," fall under these. I find similar ' sounds" are
called by Professor Miiller " Specific Modifications of Gutturals"

101

The Oriental q ; probably the elementary form of the aspirated
consonant (German guttural ch, Sechwana g 2 of the pre-
ceding analysis.}

But the element which is intended by Dr. Lepsius to be indi-
cated by q is that out of which the modern letter is proved to
have grown, viz., the Arabic J (kdf), or the Hebrew p (qof) t
which, according to that learned writer, "is formed at the posterior
part of the soft palate." Besides the two languages above mentioned,
it occurs in the tables of the Persian, Hindustani, Turkish, and
Malayan, suspended in a doubtful position between his faucales
and gutturales ; and is uniformly represented by the same letter,
except by Wilson in the Hindustani, and Smith and Robertson
in the Arabic, who employ the pointed letter k, and Crawfurd in
the Malayan that of k. However, the same letter q is used by
Burnouf and Brockhaus in the Zend, and by Rosen in the
Georgian, to indicate another element, which is included by Dr.
Lepsius under his faucales; so that there is still a good deal of
uncertainty attached to its nature.* To a local student of South
African tongues it is very difficult to form a correct conception of it.

It is, therefore, the more necessary that some discussion should
be raised on the subject, as similar elements may occur in those
tongues which it will otherwise be a troublesome matter to
classify. The Oriental articulation in question is thus minutely
described by the learned Dr. Gilchrist, in a work on the
Hindustani :

" Q, or our k, articulated by raising the root of the tongue simply towards
the throat, which must not be in the smallest degree ruffled, as in forming Jch,

* Some objection may be urged against the repeated mention in this
work of the graphic symbols used by different writers ; but I refer to them
upon the ground of the probability that every linguist will have chosen a
certain letter, or diacritical modification of it, to indicate what he considered
an articulation approximating to that which it was usually intended to
indicate. In this respect, the tabular abstracts which I have made from
the numerovis tables of Dr. Lepsius cannot but be valuable.

102

or gk. The q may consequently be styled a deep but liquid linqual letter,*
produced by clinking tbe root of the tongue against the throat, so as to
cause a sort of nausea. The same sound will be recognised when pouring
water in a particular manner from a long-necked guglet, as the liquid
decanting may represent the lower part of the tongue acting upon the throat
or neck of the vessel in question, unruffled by the water gushing from it."
British Indian Monitor, vol. i., p. 12.

Again :

"Though q be called a guttural, I would rather name it a linqual letter,
because its formation is almost entirely owing to the root of the tongue being
raised to the roof of the palate or throat, which last is preserved perfectly
unruffled in this operation, whence the real difference between q and the other
gutturals already enumerated. Water poured in a particular manner from
a long-necked guglet, or the hiccup of a man more than half-seas over, will,
I believe, yield a sound very near the q, which, when duly articulated, has
the peculiar property of exciting a nausea in the learner. When followed
by u, the scholar must never, as in English, change u to w, but call words
like qulum (a pen), qazee (a judge), kulum, kazee, never qwulum, qwazee, &c. ;
nor qeer (pitch), queer, but keer, or rather qeer, &c., with the lingual q above
described alone."f Ibid. p. 20.

The above quotations only confirm the fact of the universally
acknowledged difficulty of giving a reader an adequate con-

* In taking exception to this term, I may as well state that the root of
the tongue is called into operation in the formation of all tbe gutturals, and
that there is no good reason for not including this element q among them.

f The italics are all substituted ; also in the examples throughout the
above two quotations.

Dr. Duff considers that the articulation has been happily described by
Dr. Gilchrist in the former. See " Original Papers," by Monier Williams,
M.A., p. 79.

Professor Miiller, I find, makes little reference to it. At p. xxxv. he calls
it " a low guttural," and speaks of its " superlative degree of explosiveness"
as " a characteristic peculiarity." In his Table, at p. xci , it is called a
guttural tenuissima, and indicated by q as a base letter, or by K as a " modi-
fication of the second degree." Rather indefinite ! When in Natal, in
August, 1861, I examined a Hindoo on the subject of this element. His
pronunciation answered to Dr. Gilchrist's description, and confirmed me
in the above conclusions. It did not appear to be more explodent in its
nature than I or r, and, so far as I was then able to judge, is certainly not
a mute. An acquaintance of mine, who spoke Bengali, considered it was
" a sort of liquid guttural."

103

ception of a consonant by mere verbal description ; they will at
all events enable us to form a pretty close comparison.

Dr. Forbes says of it :

" <_3 bears some resemblance to our c hard, in the words calm, cup ; with
this difference, that the is uttered from the lower muscles of the throat."
Hindust. Grammar, p. 6.

I have already given Dr. Gilchrist's description of his digraph
kh, alluded to in both of the above quotations, and shown it to
be identical to the ch in German lachen, or the Sechwana g of the
missionaries (^ of preceding analysis). The learned writer
describes the element q as only differing from this kh in the fact
that in its enunciation the organs in contact are unruffled, by
which may be understood that they undergo no vibration by
means of a strong aspiration of the breath. It is evident, from
the whole description, that the element intended to be indicated
by q, is a leiiis form of the simple explodent k that is, a liquid
modification of g, bearing the same relation to this letter as I
have proved r and I do to d, and as ch (lachen}, # 2 does to aspirate
g viz., g]

Now, in respect to all that has preceded, it must be borne in
mind that I have stated my objections to the terms fricatives or
continues merely as generical terms, in contradistinction from
explodents on the one hand, and aspirates on the other. My
meaning will, I trust, be apprehended, inasmuch as I have at
the outset attempted to show that, in respect to the gutturals, the
terms fricative and aspirate have been confounded, and to the
sibilants, that some of the so-called fricatives are simple " ex-
plodents," and the others aspirates. I am not aware of having
denied that either word, fricatives or continues, may be used in a
specifical sense, or that it may be introduced consistently into a
general classification, in contradistinction from the mutes. In
fact, the whole of my reasoning goes to prove that this is quite

104

practical, and that it amounts to a necessity, as the following
tableau of the results of my operations thus far will clearly show.
While I hold that certain elementary forms of r and Z, and those
elements that may be proved to be their analogues, are all simple
(tenues) liquids, I admit that all these must also have corres-
ponding aspirated forms (" modifications" of the first kind}, and
also vocalised forms (" modifications" of the second kind^to both of
which the terms fricatives or continues are not only applicable, but
have already been applied, unfortunately, by those who, in
hastily leaping at a physiological classification and terminology,
ignore demonstration of another kind. To borrow the opinion of
Dr. J. Miiller, the application of the principles upon which the
distinction between mutes and liquids is founded, has been imperfect
(Physiology, p. 1045); and so long as men refuse to abide by
these principles, it is impossible to arrive at correct results.

Simp. Exp.

Asp. Exp.

fortis.

lenis.

fortis.

lenis.

GUTTURALS

(Mutes k

'"(Liquids ...

(Mutes t

t C

LINGUALS . . .

Liquids... j

It must be evident to the reader of the above tableau, that if I
would maintain consistency in this . classification, an element is
actually wanting where I have placed the asterisk, that is, a
simple liquid " explodent " related to /*, as r and I to r and
T respectively ; or to g, as r and I to d. It is difficult to con-
ceive of any other position to which to assign the Oriental q in

* I am aware that, in consequence of q being a superfluous letter in the
European alphabets, it has been employed in the Hottentot as well as the
Kaffir to indicate one of the clicks ; but as it, or the letter from which it was
derived, has for ages indicated a guttural sound, it is not improperly em-
ployed in the above classification.

105

question, for it is evidently not an aspirate consonant. If my
conclusions prove correct, and the letter q be retained for this
peculiar element, in order to preserve analogy in their graphical
representation, we shall have another letter prepared to supply
the want of one to represent the g* of my analysis (ch in lachen,
Sechwana g) viz., g.

The foregoing remarks will, no doubt, shed a little light on the
following quotation from Mr. Donne's admirable dictionary of the
Zulu, and bring the subject of this consonant a little within the
range of observation of those whose studies are confined to the
South African tongue.

" G is a guttural, and has, in Zulu Kafir, two sounds. The first is the
hard sound e.g., igama, goba, as in English go, gab ; the second is soft
e.g., gapa, or a sound between g and k, or between g and r (soft}.* The
dialectic differences, however, respecting the gutturals, particularly in
Natal, render it extremely difficult, if not impossible, to assign to each
sound its proper limit, and hence only one character represents them both.
Besides, there is no provision made yet for the proper distinction of sounds
in the present state of orthography." Zulu Kafir Dictionary, letter G, p. 90.

By the aid of all these quotations, I am led to conjecture that
the Oriental q, and the soft form of the g of Mr. Dohne, are
probably identical consonants ; but to proceed beyond conjecture
is impossible, without the aid of more satisfactory data than have
been collected.

It is very probable that the Danish g (y or x' of Lepsius), which
I have only heard pronounced by one intelligent Dane, is the
same element, inasmuch as it is not an aspirate, but a simple
liquid articulation, though considered by Dr. Lepsius as the soft
form of ch (in lachen), the g of this arrangement, and called by
him Sifortis fricative. This assumption would only make them
both lenes, and, judging by the above fragment of a classi-
fication suggested by the Sechwana phonology, no fricative (or
continuous or liquid) consonant can be hard.

* The italics are substituted. By the r here must be distinctly under-
stood the arbitrary letter introduced by the missionaries to indicate the
Kaffir " soft guttural"

106

The result of the above analysis is, that the Oriental q is
probably the simple guttural liquid (tenuis); but Professor
Miiller writes (p. xxvi.), " in truth, a guttural liquid is not to be
distinguished from a guttural flatus, except in theory." I have,
however, proved the lingual aspirated liquids (lingual "flatus")
r and Z in Sechwana and Zulu respectively, to have corre-
sponding simple lingual liquids in the former language ; and the
probability is, that there is elsewhere to be found an analogous
guttural form, to which the three elements above alluded to are
very close approximations. The above very learned writer, as I
have already shown, confounds simple breathings with their
guttural modifications, which will account for his opinion just
quoted ; but the physiological fact should be borne in mind, that
the same disposition of the organs (i.e., partial contact of the
dorsum of the tongue with the posterior part of the palate) is re-
quired in the enunciation of the pure guttural liquid as in that of its
aspirated form, ch in loch, and its vocalised form, the Arabic .

However unsuccessful the reader may consider me in at-
tempting to prove this interesting point, he will no doubt admit
that the relations of, or differences between the powers of, the
three letters, o Arabic, g Danish, and the soft Zulu g of Dohne,
are worthy of further investigation.

III. CEREBRALES INDICJE (or LEPSIUS.)

Next in order we find the two series (1) the Cerebrales, said to be
peculiar to the Sanskrit ; and (2) the Linguales, pertaining as " exclusively
to the Arabic and cognate languages," both of which appear to be frequently
confounded, but are considered by Dr. Lepsius to be " entirely different."

Professor Miiller says " It is true that there is a difference between
the Sanskrit ~g and the Arabic J-,. In the former, the tongue is more con-
tracted than in the latter, but both are produced by contact between the
tongue, more or less contracted, and the palate. Their difference is so
slight, that here again an organ which is able to form the Sanskrit lingual
is generally unsuccessful in the formation of the Arabic lingual. In
Hindustani, therefore, where, owing to the mixture of Arabic and Sanskrit
words, both letters occur, no difference is made between the two." Wilson,
Indian Terms, p. xvi.

107

This able writer classes them as first and second modifications of the
" dentals" (common linguals) t, d, s, z, <tc. ; and (in his Table, p. xci.)
indicates the former (except the sibilant sh and the liquid r) by italicised
letters, and the latter by capitals. It is to be questioned, however, whether
these " modifications" (excepting the Sanskrit sibilant sh) are intended by
him to be analogous to those of the gutturals called palatals, as he would
appear to make them in his " Missionary Alphabet" by italicising them.
He says elsewhere (p. xxxix) " These linguals vary again in the degree of
obtuseness imparted to them in different dialects, a difference which evades
graphical representation."

In the absence of the living pronunciation, I shall, as usual, give the
reader an abstract of the various modes of indicating these consonants by the
able linguists who have given attention to those tongues, so as to enable
him to arrive at an approximately correct conception of their peculiarities.

The Series of" Indian Cerebrals" in the Orthographical Systems of different
Linguists.

"Standard Alphabet'
of Lepsius, used also
by him for the Sans
Jcrit and Sega

fort
t

Explosivce.

is. lenis. nasals

d n

FricativcB.
fortis. lenis

Ancipites.
r I

Aspirates,
f (t

Sanskrit (Orient, gr.)...

2
t

d n

T 35

t d?

Wilson

d n

th dh

Bengali :

d n

th dh

Zend:
(Oriental gr.)

s %

Burnouf

ch j

s' sh

BrocJchaus

sh i

Hindustani :
Yates

sh zh

Oikhrist (1796)
(1806)* ...
Wilson

sh zh
sh zh
sh zh

Malayan :

(s')

Javanese :
Crawfurd

* British Indian Monitor, vol. i.

108

Other letters are made to take the place of these consonants, under the
explodents of the Chinese.

Explosives.

FricativcR.

Ancipites.

AspiratcB.

"Standard Alphabet"

fortis. lenis. nasals.

fortis. lenis.

of Lepsius, used also

hy him for the Sans-

krit and Bega

t d n

r I

t' df

Chinese :

Revs. J. Gough ...")
T.McClatchie)

t(ts) d(d))

r z

t (ts)

Stephen JEndlicher ...

tc' tc'

sli sh

tc'

1. Explosives In the comparative table under this head, there is not
sufficient variety in the letters used for these articulations to assist one in
his conjectures. The following is Dr. Lepsius's physiological description
of their formation : " This class ... is formed by bringing the tip
of the tongue backwards, and upwards to the neighbourhood of the palatal
point, so as to produce there the explosion or friction.* To our ear, these
sounds are nearest to the dentals." Standard Alphabet, p. 44.

Of the same sounds, Sir John Stoddart says :

" The reason of applying the term cerebral to any of these letters, I never
could discover; nor does Dr. Lee's remark render it to me more intelligible-
He says of the Hebrew teth, ' it should be pronounced with the tip of the
tongue against the roof of the mouth, just as our own t is, and hence it may
be termed cerebral.' To the English ear, the sounds expressed by the two
Sanskrit series appear scarcely, if at all, distinguishable; but, to the native
ear, they are perceptibly different. According to some persons, this arises
from a slight lingual vibration in the (so-called) cerebral series, somewhat
approaching to the Mexican tl.\ It is said, however, that the native writers
employ the characters of this, but not of the other series,] to express the
t or d in English proper names." Glossology, p. 133.

According to Dr. Gilchrist, one of the most able authorities quoted, the
above articulations, which he indicated in 1796 by 1, d, and in 1806 by t, d,

* This description, as well as Dr. Gilchrist's, in the sequel, does not differ
from that of the formation of the cerebral click in Hottentot, by Mr. Tindall,
p. 12 of his Grammar, before referred to.

f As will be observed in the sequel, this sound is frequent in the Sechwana..

I There appears to be confusion in this quotation i.e., the Hebrew
teth to, according to scholars, is equivalent to the Arabic lingual Jb (not the
Sanskrit cerebral) ; so it is not certain to which series the remarks are
intended to apply.

109

are " formed by conveying the tongue forcibly against the roof of the mouth,
while articulating the common dentals t and d of our own alphabet, in
tub, duck, &c., which, he continues, are however softer than they. Elsewhere
he says " d, t, r, require that organ [the tongue] to be curved backwards,
and then struck against the roof of the mouth, as in dull, tub, rub," and
then further describes them as " a great deal harsher than our English
d, t, r."

Professor Miiller indicates them by I and d (always italicised), and
classes them as first modifications of the common European linguals, t, d.
The same with their aspirate forms.

Both Dr. Duff and Dr. Forbes concur with Dr. Gilchrist in his
description of these elements in Hindustani, excepting that the latter speaks
of the tongue being only well turned up towards the roof of the mouth in
their articulation. They also indicate them by t and d. All state that
they approximate more closely to the European linguals than do the Arabic
elements. According to the former linguist, the Sanskrit t and d are of far
less (ten times) frequent occurrence in Hindustani than the Arabic linguals
t, d. According to the latter, Sanskrit d is very much akin to Sanskrit r
(r of Duff) 5 " in the Davanagari, the same letter serves for both, and the
same applies to their aspirated forms," dh and rh,

Notwithstanding the above attempts of able linguists to explain the
nature of these two members of this so-called series of consonants, it is a
difficult matter to conceive of even an approximation to their " sounds."
I have already alluded, under the head of vocalisation, to the nature of
the articulation produced by an attempt to vocalise the mute consonants
viz., that the explosion of the element in each case is strengthened, and
a hiatus is caused between it and the following vowel. Some analogous
articulation has also been referred to as occurring in the Hottentot language.
In the absence of the living pronunciation, it is not easy to explain those
in question otherwise. Were they only described as faucal utterances, it
might be possible to explain them as modifications of lingual consonants,
or combinations of these with liquids; but the subject is rendered more
intricate by the addition of a nasal to the series viz., n, in the Sanskrit,
Sega,* and Bengali.

As the late Professor Wilson is one of the authorities quoted in the table,
the following remark from Sir John Stoddart, on the letter n, will enable
the reader to make an effort at hair-splitting in English phonology, based
upon the Sanskrit distinctions : In the latter (' proper n'), indeed, Pro-
fessor Hayman Wilson makes three distinctions (besides the other which
he galls guttural) viz., & palatal in singe, a cerebral in none, and a dental in
content ; but I must own that I see no ground for these distinctions in the

* An African language.

110

proper pronunciation of the English language."* Glossology, p. 142. ( See
Professor Wilsons " Sanscrit Gram.," p. 5, cited in that work.}

Professor Wilson (cited by Miiller) has four modifications of the
common n e.g.,

Sanskrit Lingual. (Tamil). Guttural. Palatal,

n n n n

I have already shown that the third, viz., n (= $ is decidedly a simple
consonant, as much as m or n ; and the fourth a compound articulation, as
much as my or y-y would be. To call either of them a modification of the
common n is only a graphical fallacy, arising from the assumption that the
sound of a nasal depends on the nature of the consonant which follows it.
It is strange that, though the Hindustani linguals t and d are borrowed
from the Sanskrit, the corresponding Sanskrit lingual nasal (so-called) did
not accompany it. As it is, phonically, only one, i.e., the common n is used
in Hindustani. The probability is that, as a phonic element, it has no
existence separate from that of the common n, as it is said to occur only in
the Sanskrit (and Bengali) ; and the former is a dead language, which,
according to Professor Monier Williams, "was never spoken, and
never intended to be spoken in the way it is at present written" (p. 262).

It may perhaps assist the reader in forming his own conclusions, if I add
that, besides the n's preceding a guttural, " dental," and labial, respectively,
and the " French nasal" in bon, all of which have already been considered,
Dr. Gilclirist gives, in his work on the Hindustani, frequently quoted
above, the following remarks on two additional forms of the letter:

" n") The first is peculiar to the Hindoos ; the second to the Arabs ; still
n)both are pronounced like the common n in run, nurse." Brit. Ind.
Mon., p. 47.

2. Fricative. Under this head (in above table), there is more material for
conjecture, assuming the comparative tables of Dr. Lepsius to be correct.

It is remarkable ih&tjive English linguists, and one German, have repre-
sented the fortis (s of Lepsius) by the digraph sh; one French writer by
ch, which in his language is well known to coincide with the above English
sound; Bopp by s, which, judging by the remarks at p. 34, Stand. Alph.,
is intended by him to indicate a sound equivalent to " the English sh, the

* Trusting that the reader will not suspect me of presuming to " abrogate "
English orthography, I venture the following remark :

The n's in the three words here instanced would, upon the principles of
his work, be written as follows singe, none, content. In the first example,
the "palatal" element is not n, but a portion of the combination of sounds
following it, which is improperly indicated by g. The is only vocalised,
as is also the second n in none ; both upon principles to be examined in
the sequel.

Ill

French oh, German sch ;" only one writer, Crawford, gives it a new letter
viz., '. In all the tables of the languages, except the Sanskrit and the
Bengali (the only two in which the so-called lingual n is said to occur),
these two fricative elements are confronted by Dr. Lepsius's exponents
* and z, not s and z ; so that it must be concluded the is also confined to
those two languages. It is therefore strange that, while the explodent forms
are represented as occurring in so many languages, only the Sanskrit should
contain an example of the fricative equivalent s (Lepsius).

Professor Miiller calls the fortes s a lingual flatus, and says it " is a
sound peculiar to the Sanskrit, and, owing to its hollow guttural pro-
nunciation, it may be expressed there, as it has been hitherto, by s, followed
by the guttural k (sh)." P. Ixxi.

Dr. Duff says of it, in Hindustani " it is pronounced in the same way
as sh ; and is so marked (sh) because it has a distinct letter in the Sanskrit."
" Original Papers," by Monier Williams, p. 84.

Dr. Lepsius himself, by placing * in the series in the Bega and Zend,
would seem to indicate that it indicates the same sound ; but the following
is his description of the difference :

"When the tongue recedes still further [than in the articulation of **],
so that behind the upper and lower teeth a greater hollow space remains,
this enlarged resounding space produces the sound s. . . . . The
Indian cerebral |, however, receives, from the peculiar flexion of the tongue,
which produces a double cavity in the mouth, a somewhat different ex-
pression, indicated by the cerebral point."

As to the lenis form, the fact of three English writers representing it by
zh, and one French writer by j, leads to the supposition that the intended
element approximates to z of Lepsius (zh, or s in vision).

So it is really in the Sanskrit and Bengali that the fricative s and z
belong to the series of sounds confronting them under the head of explosives.
Judging by the examples of the explodents in Chinese, to which the common
'' palatals" sh and zh, s and z , are apparently allied, the probability is that
s and 2 also will rather be included among the " palatals," otherwise called
the " unstable combinations" of Dr. Latham.

3. In respect to the letters under the head of ancipites and aspirate, I
shall merely, for the information of the reader, add the following remarks,
culled from Dr. Gilchrist's able work :

(a) " The letters r, rh, are rather nominal deviations from d, dh, than
formal characters." B. I. Mon., p. 23.

(&) " r, I, letters peculiar to the Hindoos in form, but exactly of the
same power as r and I" Ibid. p. 47.

Professor Miiller says " The liquid is the lingual r, produced by a
vibration of the curled tongue in which the Italians and Scotch excel, and

* Interpolated.

112

whiclj we find it difficult to imitate," p. xl. The same occurs in Welsh,
and is that I have attempted to show is the vocalised form, or second
modification of the unmodified liquid r in Sechwana morimo.

In the foregoing observations on the " Indian Cerebrals," it must be
evident to the reader that it is only with regard to those under the head of
explosive that any real difficulty is presented, which would lead one to doubt
whether they are merely simple modifications of t. At all events, no one
will deny that it is absolutely requisite that sounds should be resolved into
their elements before they can admit of classification.

IV. LINGUALES ARABICS (OF LEPSIUS).

The following table shows in a compendious form the different modes in
which the members of this series have been, indicated.

Tlie series of " Arabic Linguah" in the Orthographical Systems of different
Linguists.

" Stand. Alph." of Lepsius ...

Expl

fortis.
t

osivce.

lenis.

Fricativce.

fortis. lenis.
s z

Aspirates.
fortis.

(t)

t
t

&>
d

b
d

*
thw
?.

(11)
z
dv

37
ye

*
CO

z
th
z

z 1

Arabic

Smith and Robinson
Persian :
M M Ibrahim

Hindustani :
Yates

Oilchrist 1796

Wilson

Malayan :

Turkish :

Galla (African) :
JutscheJt....

Other letters are made to take the place of these consonants, under the
explodent of the

Chinese :

Revs. J. Gough ^
T.McClatchie ...)

t(ti) d(dz)
f

t (is)

Stephen EndlioJier

ts ts

113

In the above table, the examples which would lead men to suppose that
they are a series of compound sounds, is that they are represented as such
by Gilchrist in four members, by Crawfurd in one, and by Jutschek
in one.

Dr. Gilchrist elsewhere* remarks of the explodent forms "The Oriental
t, d, [are] formed with a slight protrusion of the tongue between the teeth.
Tub, duck, do ; tube, duke, dew, due, will convey a tolerable idea of the
difference between palatials [Indian cerebrals already noticed] and dentals
[Arabic linguals] f in the Eastern tongues, the t d of the four last even with
us being much softer than in the three first ; for, in fact, some people seem
to soften the liquified d and t with us so far as to say tshube,jooJe,jew for
due, &c. The lisp of children and others will convey a tolerable notion
of the very soft dentals d t in question, as essential sounds in the Oriental
tongues, &c."

Again, he writes} " The soft d, t, r [Arabic linguals], cannot be softener!
too much, and the harsh d, t, r, can hardly appear enough so, till their
opposite natures be sufficiently understood from practice."

Dr. Duff (Papers by Monier Williams, p. 80) writes " It resembles, says
Dr. Carey, the Yorkshire pronunciation of t in butter^ It also resembles
as nearly as possible the soft French dental t in tu." Of the d, " it is
formed with the point of the tongue pressed on the roots of the upper teeth,
nearly as in duke, due; or still more nearly as the French dental d in des."

Dr. Forbes, also, in the Hindustani, describes the pronunciation of
the fortis thus : " Softer and more dental than that of the English t, it
corresponds with the t of the Gaelic dialects, or that of the Italian in the
word sotto." Of the lenis d his description is literally the same.

Professor Max Muller considers these elements as modifications of
the second degree of the common linguals t, d, s, z, and represents them
thus

Tenuissima. Flatus Diacrit. Flatus Aspirate. Tenuissima Atsil.
T, Z; Z(fr), T.

From the above descriptions, it seems but reasonable to conclude that
the sounds in this series, if not consonantal diphthongs, can only be
regarded as those combinations of simple elements which Dr. Latham
would call " unstable," and which, under the provisional denomination
" palatal," I propose to reserve for the third part of this work.

* British Indian Monitor, vol. i., p. 10.
| Parts in brackets are interpolated.

* British Indian Monitor, vol. i., p. 17.

This is not unlike but-ther. See Chap. V.

114

In respect to both the Indian cerebrals and Arabic linguals,
and I may add the Namdn dicks, it is very difficult for any
European, without hearing the sounds thus indicated, by means
of numerous orthographic expedients, to decide whether they are
consonantal diphthongs* (mutce cvm liquidis), vocalised con-
sonants, or elements combined with the pure " palatals ;" but
there is nothing in the foregoing analysis to warrant the con-
clusion that any of them is a simple consonant.

In studying these peculiar consonants, it ought not to be for-
gotten that they belong to graphic systems which possess " such
a superabundance of characters that one sound has often three
letters," and in which the forms of the letters are " not less liable
to change" than the "powers of the letters are very absurdly
ever varying" in the European systems. In the words of
Professor Monier Williams, "what creates the difficulty [in
reading them] is, that every letter has four separate forms,
according as it is initial, medial, final, or detached; and that
groups of three, four, five, or even six letters are shaped exactly
alike, being only distinguishable from each other by the number
and position of their dots."f In comparing the opinions of the
learned on the distinctive natures of such letters, it is difficult
to discriminate between those of men, on the one hand, who
have studied a language critically that is, who have had
to do with its literature very much in the same way as
we, spite of the absence of the living pronunciation, manage to
command a knowledge of Greek and Latin ; and those, on the
other hand, who have acquired it practically, by "condescending
to learn the vulgar tongue," as spoken by the bulk of the people.
As an instance of the confusion arising from a superfluity of
letters being worse confounded by the contrary opinions of
learned linguists, I give the following :

* See Chap. III. on Compound Consonants.
+ " Original Papers," p. 260.

115

An interesting icriter, in a late journal
of literature *commentiny unfavour-
ably on a system of " romanising"
suggested by Professor Monier
Williams, writes
" He would only give one h for the

Hindustani two ;

Remarks.

" Only two t's for the Hindustani
three ;

One z for four z's in Hindustani

" Only two s's (viz., * and sh) for
the Hindustani three, and similarly
in other letters "*

As to the " two h's" I have already
given a quotation from the learned
Orientalist, Dr. Gilehrist, to the
effect that they are identical ele-
ments in Hindustani.

As to the " three t's," the preceding
analysis will show how far linguists
are at present justified in regarding
them as simple consonants, and, till
resolved into their elements, it is
impossible to distinguish them by
any uniform orthographic expedients.
These, viz , z, z, z_, which Dr.
Gilchrist gives, besides zh, he pro-
nounces " merely formal varieties of
the self-same sound."

These, viz , s and which Dr.
Gilchrist gives besides sh, he pro-
nounces " varieties of simple sounds
by different letters."!
Such glaring differences of opinion are of course to be attri-
buted to difference in the modes of acquiring Oriental tongues.
Indeed, unanimity cannot be expected to exist between the
closet student and the colloquial learner. In the case of the one,
the language is presented in written characters to the eye ; of
the other, in spoken sounds to the ear. Both modes of research
may be united, but how seldom is this convenient to men bent
on some third occupation.

Sources of Confusion, and the main Obstacles to Uniformity.
(a) The nature of the confusion which exists will be more
clearly illustrated by an examination of three series of letters in
the Hindustani alphabet.

* Evangelical Christendom, vol. i., p. 242.
f See British Indian Monitor, p. 47.

116

(1) The Sanskrit has a set of peculiar linguals (Linguales
Indicce of Lepsius.)

(2) The Arabic has also a set of peculiar linguals (Linguales
ArabicfB of Lepsius).

(3) Each of these languages, again, has a set corresponding
with the common European linguals, t, d, &c.

(4) Now the Graphic /System of the Hindustani is a composition
of the alphabets of both of these languages ; i.e. } every consonant in
either of them is represented graphically (that is, however, not to
say phonetically} by a corresponding new letter in Hindustani.

(5) But, phonically, the peculiar Sanskrit linguals are said by
linguists to approximate to the common European linguals.

(6) Moreover, phonically, the peculiar Arabic linguals are
said to be actually used by the natives of Hindustan to represent
t;he common European linguals.

(7) Therefore, phonetically, the Hindustani alphabet must be
deficient in some consonants, viz., some of the European linguals;
and is, after all, rather a sort of artificial or hybrid orthography
maintained out of traditional respect or veneration for those
more ancient ones of which it is composed.

(8) It seems to be the vain object of modern linguists to pre-
serve these merely graphical distinctions in any scheme for a
phonetic alphabet.

I append the following example among those alluded to :

/Lepsius ... t d

4SBy ***- T ?

*r&..^} ?.<-> ;

fits peculiar ) ,
linguals..)
ARABIC i

s z
Z T

s z (we)
\i

Its common | ^
( linguals..)

quiva en j- JTJ W ^ M J S ^ j

LT J

S 2

117

Now, according to the descriptions of most linguists the
phonical equivalents are

Of the former t d s z

(peculiar) (common Europ.)

Of the latter t d s z

(common Europ.) (common Europ.)

So that the only strange elements are really t and d of the
former ; and it is very probable that such an examination of the
Sanscrit linguals would have nearly a similar result. Thus, to
a legitimate number of phonetical equivalents are added hosts
of superfluous graphical equivalents, which must necessarily
suggest divergent articulations, and tend to mislead.

The above is a suitable example to show the degree to which
the subject has already been complicated by trying to maintain
a scrupulous respect for Oriental classifications. It is out of
such a jungle of graphical materials that it is necessary for the
linguist to extricate himself before he can fall into the plain
paths converging to a uniform phonetic alphabet.

(6) This very important error, which made the Hindustani
alphabet a mere graphical hodge-podge rather than a phonetic
model, must forsooth be followed up by European linguists in
the application of one graphic system (the Roman) to two dif-
ferent processes of transcription and transliteration.

By transcription must be understood the employment of certain
letters to represent certain articulations and sounds; by trans-
literation that of certain letters (e.g., Roman) to represent certain
other letters (Oriental or Hieroglyphic).

The one process, which is that of the principle of a phonetic
alphabet, would indicate

A simple articulation or sound by a simple letter.

A compound do. by a confound letter.

A modification of either by a modificatire mark.

118

The other process, however, is proposed to represent

" Every [Oriental or other] double letter, though in") , & ^^ letter>

pronunciation it may be simple >

E.g., that anything like the English word though, pronounced tho, be
transliterated to the same number of letters.

"A single letter [Oriental or other], although its ) , ft . fo letter."

pronunciation be that of a double letter >

E.g., that the Sanskrit 13| = ch in church, consequently a compound
articulation, should be represented by k' (Lepsius), or 4 by Jc (italicised, of
Miiller) ; or, that the Armenian e and o, which are pronounced ye and we,
be written e and o.

The difficulty which has no doubt given rise to this new
process is thus carefully described by Professor Lepsius :

It " is greatest in those systems of writing, which originating in an
earlier period of the language, and fully developed, have been retained
unaltered, whilst the pronunciation has undergone a change, as also in
those in which several reformations have left their traces. An instance of
this kind has already been mentioned in speaking of the Sanscrit palatals.
The differences of European orthography have mostly arisen from similar
circumstances. Some such difficulties, however, are presented by almost
all existing alphabets which are not of modern formation. As the object
of a standard transcription is to avoid as much as possible all such incon-
gruity of sound and sign, no other course remains open in such cases than
to fix upon a distinct period of the language in question, and to adapt its
transcriptions to the different purposes of rendering either the actual pro-
nunciation, or the ancient one which had been expressed by the alphabet,
and which may be deduced from it by linguistic researches." Standard
Alphabet, p. 53.

This able linguist chooses the latter alternative, in the words
which I have confronted with those of the late Baron de Bunsen
on the title-page of this treatise, viz., that " the linguistic
scholars will prefer to follow the written system fixed by litera-
ture, and to neglect the varying deviations and shades of modern
pronunciation." It must, however, be apparent to my intelli-
gent readers, that for Professor Lepsius to ignore the actual
pronunciation is tantamount to a denial that phonetic philology
is "a branch of inductive science" standing "on precisely the
same footing as geology, and those other sciences which are

119

' connected by this bond, that they endeavour to ascend to a
past state of things by the aid of the evidence of the present.'"*

Professor Miiller, in defence of this process, recently suggested
by him, remarks :

"If we attempted to represent the sounds in transcribing literary lan-
guages, we should be unable to tell how, in the original, sounds admitting
of several graphic representations were represented. In written languages,
therefore, we must rest satisfied with transliterating letters, and not attempt
to transcribe sounds."

Again, in other words :

"If, instead of imitating the letters, we attempted to represent their
proper pronunciation at a certain period of history, how should it -be
known, for instance, in transcribing the French of the nineteenth century,
whether ' su' stood for ' sou,' halfpenny, or ' sous,' under, or ' soul,' tipsy."f

Now, it cannot be denied that " in historical languages the
system of orthography is too important a point to be lost ;" but
it seems very natural to expect that by mere transliterations,
such as that suggested, it is most likely to be lost. And, just as
much as " it is a mistake to imagine that in living languages all
etymological understanding would be lost if phonetic reforms
were introduced,''^ so it would be equally a mistake to suppose
that in archaic literatures all etymological understanding would
bo lost if the historical orthography were maintained. A refer-
ence to the Greek classics will bear me out in this assertion ; for
in spite of our ignorance of the genuine pronunciations, the ety-
mology of this extraordinary language is to this day a rich field
of research, and will be rendered still more interesting in the light
shed on its structure by vernacular forms of speech. Professor
Lepsius tells us that " the Armenian alphabet has also undergone
peculiar alterations of pronunciation, which may be historically
proved ;" and also speaks of the ancient pronunciation of a tongue
being deduced from the actual one by linguistic researches. In

* Dr. Donaldson, Ency. Brit., vol. xvii., p. 53'.), citing Whewell.
f Troposals, &c., p. Ixxxix.
1 Ibid.

120

a similar manner, the difference between Professor Miiller's
examples, su of the present, and sou, sous, and soul respectively
of the past, could be arrived at by comparing French works of
the nineteenth century with past historical records of the lan-
guage. According to this learned philologer himself, historical
or antiquated alphabets are objects of archasological research ;
and I think that under this specification ought to come all
Oriental alphabets, most of which have already been reduced to
" romanized" forms. Vernacular alphabets, on the other hand,
no one will venture to deny, are, elementarily, objects of phonical
research. Transliterations may be all very well in the one
branch of inquiry, but transcription is certainly the process
which is indispensable to the record of phonetical data ; therefore
the two methods ought to be kept separate. The one scholar
ought to persist in Avriting the Greek eyyve as the Greeks wrote
it the other, if he adopt the Roman alphabet, to write it
according to the ear, engus or epgus, or to any preconcerted
standard of letters ; but neither of them is justified in trans-
literating it to arbitrary Roman characters (thus, eggus), in
which word the two gutturals are universally agreed to have
distinct phonical powers.

The two modes of recording the symbols of speech can be
pursued separately, the one rather aiding than detracting from
the other. In the one case, the philologer merely places and
transposes the dusty specimens of a valuable cabinet, and con-
tents himself with noting the chronological relations of materials
collected and successively shifted by his predecessors; in the
other case he plods, and observes, and notes as he treads the
situs of all that his attention is bent upon, while his mind con-
ceives and generalizes, and thus builds up a fabric of positive
science, which in its turn becomes a subjective means of fresh
and more important acquisitions.

In presuming to write in this strain, I have been actuated

121

merely by a desire to unveil to those versed in historical philo-
logy the difficulties which have occurred to myself at the
threshold of vernacular studies, in endeavouring to overcome
this troublesome subject, and certainly not by any meddlesome
inclination to question the mature opinions of men supposed to
know better.

I trust that what has preceded in this humble treatise will
show the reasonableness of my opinions, viz. (1) That if Dr.
Lepsius only aims at a standard historical alphabet, his maxim
holds good, and he remains consistent; but, if his aim is a
standard phonetic alphabet, he must abandon to the archaeologist
the second-hand materials of historical orthography, to consult
him only occasionally and supplementally, and rather look to
the situs of all living human speech for the only valid materials
of inductive phonology ; and (2) That the compromise suggested
by Professor Miiller, between the different modes of recording
historical and phonetical data, would only be another name for
" confusion worse confounded."

V. DENTALES (OF LEPSITJS).

It is under this series of Dr. Lepsius's classification that we
first meet with a few instances of simple consonants foreign to
the Sechwana, and in disposing of which, especially, I shall have
to alternate with a little speculation, the treatment of those
instances found in the language to admit of classification; as I
have already done with regard to the element q, under the head
of the gutturals.

Inquiry into the Nature of the Letters s and z, th (in thin) and
th (in thine) ; and the probable Existence of their Elementary
Forms.
I have elsewhere implied that the consonants I and r, which

are common in the Sechwana, form only part of a more complete

122

set of lenes elements, the remainder of which do not exist in the
language ; and the probability that the element s,* in the
English words parts, parks, harps, and the elementary form
of zAf in the word thin, which has frequently been indicated by
the letter 6, upon the supposition that it represented the same
consonant in the ancient Greek, both belong to it. It is, how-
ever, necessary for us to arrive at a conclusion on this head
more satisfactory than such mere conjectures. I am fully
aware that this would in any case involve a further conclusion,
as in the case of g and k, that had the said elements s and th
occurred in the language as initials of verbs, they would have
been commuted into t ; that, moreover, like r and I, they would
be expected to have their aspirated and vocalised forms. '

The consonant s, as before remarked, seems to be little under-
stood. The letter is chiefly employed in English to represent
two different consonants, as in this and these. It is by some
writers regarded as not admitting of classification, and of a
peculiar nature ; therefore not included under either the liquids^
or the mutes. Its affinity to r seems to be often admitted. Of
the Latin language, Zumpt writes

" S, like the Greek a, was pronounced more sharply than with us

" In the ancient pronunciation, there must have been a peculiar re-
semblance between the letters * and r; since it is mentioned by Varro

* That is, as a simple consonant, independently of any prepositive
element. It abounds in the language in composition with the letter t, as
in mocwetsana (a fountain), morwetsana (a damsel).

f For the purposes of this section, I shall in the text indicate th in thin
by th 2 , assuming it to be in most cases in English an aspirated liquid (con-
tinua) ; the other, in thine, by th 3 .

I Unless it be shown that this and the term continues are intended by
them to be descriptive of the same instances.

It is remarkable that Dr. Latham places under his aspirate mutes all
those sounds called by Dr. Lepsius fricative or continuous, except s and a,
which he includes among the lene mutes ; but upon what principle it would
be difficult to say. Much more difficult would it be to account for his
placing any continuous sounds under the mutes, unless by the latter term
he means " explodents," in the sense in which I use it throughout this work.

123

(De Ling. Lat. vii. 6), and others, that formerly that is, before the Latin
language had assumed a fixed form through its literature s was pronounced
in many words for which afterwards r was substituted, as in Papisius,
Valeslus, lases, eso, &c." Latin Gram, by Dr. Schiidtz, p. 6.

It is not improbable that also in the Sechwana barimo, and
the Suahili wasimo, dialectical variations of the same expression,
the letters r and s indicate only analogous elements.*

There is every reason to suppose that tJi (0) is related to ,f as I
is to r. In one respect, there is a remarkable analogy between the
two couples. What is often considered the natural infirmity, or
disability of pronouncing I instead of the r, called in Greek
rpuuXio-juoe, and in Latin balbuties, is precisely analogous to the
substitution of the articulation th for s in the words " mith for
miss, thpell for spell, and the like," by those who are said to lisp,
and which was called sonus blaesus among the Romans. There
can be but little. dispute as to s and th belonging to the lingual
series ; for though they may be both continuous and fricative,
or semivocal, in some of their modifications, like I and r, they
are, like these elements, essentially formed by a contact^ of two
organs, though not a complete stoppage of the breathing, there-
fore as much as any consonants explosive.

But the main difficulty appears to be in establishing their
quantity under the " explodents." Where, but under the lenes
forms of these linguals, can the two elements in question be

* In mozimo (-nto-rimo), the singular form of this word found among the
seaboard tribes of Sofala, the letter z indicates perhaps only a vocalised
modification of the Sechwana element.

f Dr. J. Miiller calls the English th a modification of s (p. 1049).

A writer in " Chambers's Cyclopaedia" (1860), under " Alphabet," says:
" The sound of th is very nearly allied to that of s (witness ' loves or loveth') ;
also the pronunciation of a person who lith^th," p 169. He suggests that
for each pair Z and r, s and th, one letter would suffice. And I may add,
that th 3 is similarly related to z. Sir John Stoddart (p. 134, Glossology)
writes " these two sounds, th 2 and th 3 , approximate less nearly to t and
d than to a and z."

\ For the elementary forms of these consonants I shall, for the purposes
of this section, employ the letters th 1 and s 1 .

124

placed? They differ physiologically in no respect whatever
from I and r, except the fact that the contact of the tongue, in
place of with the palate, is with the teeth ; and that while the
former may be called, organically, palatal linguals, the latter
are dental linguals. Moreover, admitting that they are liquids,
and, what may be inferred from all that precedes, that all liquids
are lenes, their quantity is explained. It is only by this train
of analogical reasoning that I can arrive at their classification as
kindred forms of r and I, and therefore as lenes. In the event
of its appearing unsatisfactory to the reader, I shall at any rate
venture to assume it ; for it is not till I come to treat of certain
combinations of simple consonants that the assumption will be
found to bear the test of anything like a legitimate proof.

The set of lenes simple " explodent"-linguals thus amounts to
five viz., l t r, d, s 1 , and th 1 . I have shown that I and r have
their corresponding aspirate forms ; the same must consequently
apply to s 1 and th 1 . It is easy to form a theoretical conception
of the aspirate forms of r and I, even had no objectively true
instances of these occurred, by simply supposing a forcible " aug-
mentation of the breathing" in every case, which is well indicated
to the eye by the spiritus asper, e.g., r,"l; but the difficulty is
to conceive of aspirate forms of s 1 and th 1 ;* and why? Simply
because of the absence of a preconceived notion of these elements,
in their simple " explodenf forms. This difficulty will remain so
long as the fact is lost sight of that all consonants are essentially
" explodent ;" that s 1 and th 1 , though liquid consonants, and may
be either aspirated or vocalised, are both completely formed
by a mere contact of organs, and, as I have shown, with
regard to I and r, to say that they are either continuous or sibilant,
or lisping, is to say that they are aspirated ; to say that they are
semivocal, ought to mean that they are vocalised. It would not

* Latham denominates the two English forms of th " so-called aspirates,"
and elsewhere classifies them as " aspirate mutes."

125

be going too far to say that, in English speech, just as the con-
ventional notions of the simple forms of I and r, which are so
prevalent in it, exclude any notion of the rough forms of these
elements found in other languages ;* so the rough forms of s 1
and th l (viz., s 2 and tfi\ usually called sibilants and lisping-dentals
respectively, and by others continues or fricatives, equally prevalent
in it, seem to preclude the formation of correct notions of their
simple liquid-" e&plodent" forms. It would be difficult, and almost
impossible, on the mere spur of this suggestion, for myself or
any one to distinguish cases in English phonology of the occur-
rence of either of these elements s or th in both their simple and
continuous (or aspirate) forms. But, in the Sechwana language,
I have been able to ascertain beyond a doubt that, in the com-
pound examples, mocwetSana (a spring of water), and natSane

* I have since met with the following remarks of the Rev. Kichard
Garnett: " An Englishman or German is apt to take a limited view of the
subject, because he only allows of one power of the letter I, and naturally
supposes that the same is the case in all other languages." Philological
Essays, p. 249. (See context.) Again " The same remark is equally
applicable to the other liquids, especially to r. A native of our Southern
counties, accustomed to enunciate this element with a delicate, sometimes
scarcely perceptible vibration, naturally thinks his pronunciation the
standard and only genuine one, and regards every marked deviation from
it aa a defect in utterance or a provincial peculiarity." Ibid. p. 254.
Again " In Welsh, the common soft r is unknown as a primary initial of
words, the aspirate form rh being invariably considered as the primitive."
Ibid. p. 255. Now, this articulation abounds in Sechwana, in both its
aspirated and vocalised forms. I have heard a very intelligent Welsh
missionary so much at home in it, his native sound, that it gave (even
independently of one or two of his native gutturals) a character of unusual
accuracy to his pronunciation, which of course elicited a good deal of
obsequious flattery, of which Bechwanas are so lavish ; but, at the same
time, this led him to misuse the tennis, or simple liquid r, in such words
as morimo, rusa, rila. Moreover, the Welsh I, as it is usually called, is a
fine instance of the aspirate I, also found in the Zulu language, and none
except a Welshman will be inclined to deny it, as I have heard one do,
by alleging that (like the clicks of other " barbarians," Hottentot and Kaffir)
the contact of the tongue with the palate is lateral !

126

(a young buffalo), the instances of s are respectively simple and
aspirate.* The fact of its being practicable to distinguish such
instances of s alone in the Sechwana phonology, satisfies me
that, could we arrive at them by a similar principle of ana-
lytical investigation, success would follow in all instances of both
elements. In the Sechwana, neither of the two forms of r
(r and r) need be left doubtful in any word of the language. I
am very sanguine the same remark will soon apply to s ; and
is it unreasonable to suppose that if the language contained
the element ih (0), it would be a mere matter of inductive
analysis to distinguish its different forms ?

Again : I have shown that I and r have their corresponding
vocalised forms. If s and th are analogues of those elements,
the same must apply to them. I have elsewhere stated that the
approximation to vocality is stronger in the aspirate lenes, inas-
much as vocalisation implies aspiration, which again means
something more than the mere momentary action of the breath,
required in the sudden collision and separation of two organs ;
therefore we must look to modifications of the aspirate s 1 and th*
for their vocalised forms. Now, s 2 and th* are commonly supposed
to have the same relation to z and th 3 ( of Lepsius) respectively
as t to d, k to g, &c. viz., that the former are fortes, and the
latter lenes ; indeed, as before stated, they are so classed by Dr.
Lepsius, Dr. Latham, and most writers. But, upon the prin-
ciples of this treatise, s 2 and th 2 are analogues of r and I] which
are lenes; therefore they are also lenes, and consequently also
analogues of d\ g] b\- but differing from some (two) of these
organically, and with all (three) specifically, in that these are
mutes, and they are liquids. They cannot, therefore, have the

* A proof is to be found in the fact that, in the former case, ts is a
euphonical modification of simple r 2 in the noun mociveri (r 1 ) ; in the latter,
of aspirate r j in the noun nairi (r 2 ).

127

same relation to z and tfo (3) as k to g, and t to d. The
forms z and th 3 must then be accounted for in some other way,
and is it unreasonable to suppose that they are the vocalised
forms of s 2 and th 2 ? The former of these latter elements I have
said exists in both its simple and aspirate forms in the Sechwana
language, in such words as morwetsana (a damsel), and nafsane
(a young buffalo), which examples are a sufficient proof that z is
not the simple aspirate form of s, as the words exhibit, exhort, in
English, in which x is considered a substitute for either ks or gs,
would lead one to suppose. It must be evident to any. one who
pays a moment's attention to the utterance of either z or th 3 , that
they are vocalised elements ; and it cannot be questioned that
the affinity between s and z, and th 2 and th 3 (0 an'd B of Lepsius),
bears a striking resemblance to that existing between r and r.
On this point I have the satisfaction of coinciding with Dr. Lepsius,
who considers that, in the pronunciation of z and th (his S), " a
vowel sound is produced in the larynx." All the modifications of
these two liquid elements will now be indicated as follows :
S 00 k c (s 2 ) S (>?orz}

th (th 1 ) th (th 2 in thin) th (th 3 in thine)

NOTE. If there be any truth in the results at which I have arrived, a
difficulty is here presented as to the graphical representation of these
modifications of the simple " explodent" element. If the mode of employing
a diacritical mark, in the form of a small circle, either above or below the
letter, he adhered to by learned linguists, who have adopted it in the case
of both I and r, to indicate the vocalisation of an element, z must he thrown
out, and rendered as superfluous and useless as q is long known to have
been. Again, the diagraph th, for the well-known consonants in English,
has been considered so objectionable, as to have ancient foreign letters
substituted for its two forms viz , 9 and S ; but there is little probability
of even these meeting with public favour. Admitting the correctness of
ray conclusions, it would not be going too far, I think, to suggest that the
letter z* be substituted for it; as such an innovation would at all events

* I find, as precedents, MM. Ibrahim in the Persian, Yates, Gilchrist,
and Wilson in the Hindustani, Crawfurd in the Malayan, Jaubert in the
Turkish, Hahn in ihe Herero, all adopt s and z to represent the two English

128

obviate the necessity of introducing a foreign letter into the Roman
alphabet. The series of modifications would then stand thus :

Z (W) z (th 9 ) z t 3

But this change would not be admissible in the event of all the vocalised
modifications ever being indicated by separate letters, as with z and the
letter v (still to be considered). The difficulty would then be increased
viz., to finding three simple letters for the vocalised forms of r, I, and q,
instead of r, l, q, and even th 3 , which, though phonically different from
th?, is not graphically distinguished from it in English phonology, not to
mention the vocalised forms m, n, and J-, already noticed.

I am fully aware that my opinions on the subject of these two
elements and their modifications are not likely to find general
favour, especially as I am unable to produce any objectively
true instances in support of my conclusions. However, what-
ever of an original character has preceded on the lingual series,
will, I trust, suggest that it is worth the consideration of such
philologers as are too well assured of the importance of the science
of language relatively to that of speech, to deem such a subject
as trivial, or " below the dignity of a philosophical inquiry."

The remaining articulations in Dr. Lepsius' series of dentales
viz., s and !z, usually indicated by sh and zh (French ch and j\
and by most writers called simple consonants, but unquestionably
" palatals," and, as much as any other instances furnished by
Dr. Latham, " unstable combinations," I reserve for the third
part of this work, as the Sechwana and other aboriginal tongues
contain satisfactory proofs of their formation from more than
one element.

forms of th. Professor Max Miiller represents them by th and dh (or th,
dh), or as " second modifications" of these ; thus TH and DH, and frankly
states that he is " at a loss how to mark" the elements (p. Ixvi.)

* Professor Miiller calls s the dental flatus asper, which is nothing more
than I have assumed it to be in the letter s (s 2 ). He says, " the more con-
sistent way of expressing the sonant flatus (z, his dental flatus lenis),
would be to put a spiritus lenis over the s," (p. Ixx.) I happen only to have
acted upon this by substituting for the spiritus lenis the circular dot to mark
its vocalisation. He, however, retains z.

VI. LABIALES (OF LEPSIUS).

Under this head Dr. Lepsius's only other simple consonants
not yet considered, are those commonly indicated by the letters
/ and v, to which it is probable a mode of classification will apply
similar to that employed in the case of the linguals above men-
tioned. Assuming, upon the grounds stated in that case, that/,
in its continuous and sibilant form, is an aspirate, the lenis form
of ph (p 4 ) it must also have its simple " explodent" form viz., the
lenis form of p ; and v, which Dr. Lepsius admits as containing
the indistinct vowel sound, must be its vocalised form, so that it
may be represented in its modifications thus

NOTE. _ Graphically considered, v would thus become superfluous, unless
it were decided to substitute it for / in all three instances ; but as there is
an instance of a pure liquid labial in the native languages of Greenlaed
and Mexico (before referred to), its retention for the purpose indicating this
would perhaps be considered desirable.

Another element is included by Professor Miiller among the labials as a
tenuissima, viz., the Ethiopic pait ; but as yet I have met with no description
to decide whether it is the tenuis (or simple liquid) form of either/, or the
so-called Mexican/, or whether a liquid at all.

As in the case of some other elements above referred to, it is
not till I come to treat of the combinations of these dental-labials

* Professor Miiller calls / the labial flatus asper. I have already
assumed the common /to be an analogue of f, and therefore write it "f.
He thinks this " soft labial" ought consistently to be written, as / with a
spiritus lenis. I have, as in the case of s already adopted /, to express
the sonant v. He fears, however, we must sacrifice consistency to expe-
diency, and therefore retains v.

130

that any proof will be afforded of their belonging to the class of
liquids, i.e., as analogues of r, I, s } &c. The Sechwana does not
contain these elements, and it is only by this analogical mode of
viewing them that I find it possible to dispose of them.

CHAPTER IV.
SUMMARY OF PRECEDING ANALYSES.

1. COMPENDIOUS VIEW OF THE SIMPLE CONSONANTS, AS
SUGGESTED BY THE PHONOLOGY OF THE SfiCHWANA LAN-
GUAGE.

I flatter myself that, in the analysis contained in the second
chapter, I have provided all the requisites of a complete in-
duction. The presumption with which I started, upon the basis
of a palpable inference from a few facts, has been verified in a
detailed and more complete examination. The third chapter
was an attempt to supply by speculation that is, by the as-
sumption of a continuity of principle, or by analogy other facts
in which the Sechwana language appeared to be deficient.

The following is a complete tableau* of the results :

FAUCAL.

NASAL.

Simple Explodent,

Aspirate

Expl., or

Tenues.

fortes, lenes.

. (Mutse... k e
Gutturales-L. .,

( Liquidas q

Aspiratae
fortes, lenes.

k c g

Voo.

lenes.

Ten. Asp. Voc.
g > g

Liquidae.. j r

Linguales

Mutae ... t d

t' d<

n n h

(s
Liquid^.. \

Z (thin)

S TZ proper)
Z /Mil)

Liquidae ... /

f(t. proper)

Labiales

Mutoe p b

P b<

m m ni

Liquidse ... v

* Those in bold type occur in the language ; the rest, excepting , are
foreign to it. Those in italics indicate elements not hitherto, to my know-
ledge, admitted in any classification.

132

The above are all the simple consonants suggested by the
phonology of this language, together with their aspirated or
vocalised forms, or first and second modifications.

NOTE. Professor Max Miiller writes: "All I insist on is this, that
there should be one class of simple or base-letters ; that there should be a
second and third class of modified letters, expressive of the first and second
degrees of modification, as explained in the physiological alphabet." Again,
" The three classes of simple and modified letters must be kept distinct."
(Proposals, d-c., p. Ixxiv.) However, it is an open question as to whether
the learned Professor has given even a physiological explanation of the
nature of his first and second degrees of modification. His description of
these is decidedly indefinite, and they are applicable to only two classes
of his consonants, the dentals and the gutturals. When Dr. LepsillS
uses the acute accent, and Professor Miiller the italics, to indicate the
palatals [thus, for ch in church, the one writes Je', and the other k (italicised)]
neither of them has given a scientific explanation of what a palatal is. The
s ame with the Sanscrit and Arabic linguals ; and though some of either of
these may yet be proved to be simple consonants, those able linguists, in
the case of the palatals, are virtually speaking of combinations of conso-
nants with vowels as modifications. Whereas, before speaking of evert
simple modifications, they are expected to refer us first to simple consonants,
or elements; e.g., in my table above, the gutturals, It, g, q; the linguals,
t, I, r, d, s, z (th') ; and the labials, p, b,f, are all simple consonants or
phonetic elements in fact, tenues proper. Again, the gutturals, k, g, q ; the
linguals, t] I] r, &c., &c., are first modifications by the spiritus of both
mutes and liquids. Again, , Z? r, s, &c., are second modifications (but
only of the liquids) by the element of vocalization. These are simple
modifications, applicable not only to simple consonants, but also to the
palatals of the above writers, which, as I have said, are combinations of
articulations, not simple consonants modified, e.g., t, and s, and y, are all
phonetic elements or simple articulations, which, without exception, may be
aspirated (1st mod), and the two latter may also be vocalised (2nd mod.),
and thus indicated respectively, t? s, y, s, y ;* but the same applies to their
combinations (not modifications) usually called palatals tsy (=ch in
church), sy (=sh in shall), y (=si in vision). These combinations would
be written by Professor Miiller, k, s, z (all italicised); by Professor
Lepsius, Je't s, 2, by simple letters, because they consider them simple
consonants modified by a change of the passive organ from the throat
to the palate. With regard to these palatals, Professor Muller's principle

* As this element y belongs to an analysis of the vowels, it is rather pre-
mature to indicate it by any arbitrary letter ; it is sufficient for me to
imply by above that I consider there is a vocalised form of y.

133

of a first modification is based upon a mere assumption that there are
modified forms of only gutturals and dentals ; whereas, as before stated, it
will be easy to prove that some of them are also analogously modified forms
of the linguals and labials. This very process of modification by the combi-
nation of simple articulations is that which has never yet been explained to
us. The question has never yet been answered as to what constitutes a
" palatal," whether called by one a "palatal," by another a " specific modifi-
cation," by a third an " unstable combination," or by a fourth a " divergent
articulation;" and, of course, it is impossible to arrive at a correct reply, till
we first ascertain all the simple articulations. It has been the object of this
treatise to arrive at that reply. Such_/?rs and second modifications as those
embraced in the above classification, I have attempted to explain in a clear
manner, which will, I trust, meet with the concurrence of the reader.

From this it would appear that there are sixteen elements,
or simple forms of articulation (temies)\ thirteen of which are
faucal, and three nasal. Of the faucals, seven are liquids,
leaving the same number of mutes as has always obtained viz.,
six. The nasals remain as usual, according to the latest autho-
rities viz., three.

NOTE. If there is any truth in my inferences from the Sechwana, and
therefore in this classification, the above table shows the practicability of
having a complete scale of simple consonants, without the introduction of
one foreign letter viz., by the mere transposition of z and v for other
letters, viz., the English th, and the Mexican liquid-labial, respectively. It
must be borne in mind that I have omitted in the above table all sounds
usually called "palatals," or any which can be resolved into a simple con-
sonant and a post-positive vowel element, under which may be included
both sh (Fr. ch, Germ, sch) and zTi (Fr.j), not to mention the Italian gl,
or the French II, mouille,-&c.

The result of the preceding analyses may be explained as
follows:

(a) All the faucal elements are simple articulations, formed
by the mere contact of two organs, and momentary (partial or
complete) interruption of the breath, whether viewed as liquids
or mutes, and therefore essentially "explodents."

(b) They all admit of an accessory element, called the spiritus
asper, and are therefore divisible into simple " explodents " and
aspirate " explodents," or, better expressed according to the old

134

nomenclature, tenues and aspirates. The probability is, that the
spiritus exists in binary quantities.

(c) They all exist in binary quantities viz., pairs, between
which there is an affinity, not only as universally acknowledged
to exist between k and g, t and d, p and b, but also between k,
and a liquid form of g, i.e., q ; between t, and several liquid
forms of d } i.e., r, /, s, and tk; and between p, and a liquid form
of b, i.e.,/. The same remark applies to their aspirate forms.*

(d) In the case of all these liquids, the terms continuous and
aspirate are synonymous.

NOTE. There is, after all, nothing new in the general arrangement thus
suggested by facts of a very original character; for it is but a return to
that of the ancients, with the addition of a second vertical set of aspirates
to correspond with the mediae, of these systems, or the lenes " explodents"
(tenues) of the more recent one.

ANCIENT SYSTEM.

Tenues. Mediae. Aspirate.

Gutturales

Linguales

Labiates

The only remarkable thing is, that the principle of binary quantities
which did not exist in the ancient classification, while it was the turning
point of a new one, became at the same time the erring point in the latter,
because based on a vulgar view of the correlation existing between certain
letters; and that the liquids were in both systems excluded from classi-
fication. The error of the ancients arose from ignorance of the binary
nature of some elements; their only alternative was to place the media
(lenes) into an intermediate position with respect k> the tenues and aspirates,
and to exclude the liquids, because they apparently bore no analogy to any
of these. That of the moderns consists in their making certain consonants,
partially similar to these mediae (as lenes), but different from them (as
continue), into a separate division, corresponding with one of mutes, in-
cluding both tenues and media, thus :

MODERN SYSTEM.

Explosives. Continues.

fortes.
(Ancient
Tenues.)

lenes.

(Ancient
Media.)

fortes.

(Aspirated
Liquids.)

lenea.
(Vocalised
Liquids.)

* See Proposals, dc., by Max Miiller, p. xxvi., where he remarks : " Pro-
fessor Wheatstone's researches prove that a distinguishing mark of the

135

Thus excluding from the tabular view the aspirated mutes (ancient aspiratte,)
as merely "explosive sounds which are pronounced with a simple but
audible breath,* as well as placing the commonly known liquids r and I
under the head of ancipites, because apparently both continuous \ and
explodent in their nature, and thus altogether leaving doubtful the important
fact that they also have tenues and aspirated forms, and ignoring altogether
the probability of there being several consonants analogous to these two.

The more natural classification, suggested by an inquiry into the pho-
nology of the Sechwana, is the following :

Tenues. Aspirates.

fortes. lenea. fortes. lenes.

Mutes

While it admits the division into tenues and aspirate of the ancient
arrangement, instead of the explosives and continue of the modern, it at the
same time admits the binary quantitative arrangement of the latter under

liquid semi- vowels consists in their having no corresponding mutes." My
conclusion above, if correct, would seem to prove that they have corre-
sponding mutes.

* Professor Miiller calls the aspirated fort is a "modified tenuis."
f It is easy to account for I and r being thus dubiously classified, and
alone of all the other liquids considered " explodent" by both Lepsius and
Miiller, by the fact that the sibilant nature of s, th,f, &c., and which has
led to their being classified as continue, is quite accidental to the breathing
required in their articulation i.e., arising from the permeable nature of the
set of teeth, and the proximity of the aperture of the mouth; any sibilation
in I and r being stifled by their formation above the aperture. It is, however,
remarkable that, conversely, the same fact will account for the " explodent"
or simply liquid nature of these analogous elements, s, th,f, &c., being lost
sight of, and which it is one of the objects of this treatise to uphold.

Professor Miiller considers that in the formation of these " sibilants"
" there is no contact at all, and the breath passes between the two organs
without being stopped, still not without giving rise to a certain friction in
passing that point of contact where guttural, dental and labial consonants
are formed." Proposals, So., p. xxvi. I leave the reader to test the cor-
rectness of this statement, by experimenting on his own organs. His
description, I believe, applies only to the formation of the spiritus, and the
German ich, which under the new name of flatus he confounds with the
liquid consonants of the different organs. See preceding note on Z and r
(p. 125). What I have aimed at proving is, that these fricatives of Lepsius
and/atas of Miiller, viz., Germ, ch (gutt), s, th, f, as well as the aspirate
form of r and Z otherwise called continuous consonants are in every case
liquid consonants or spiritus (or flatus or fricative}.

136

each head with respect to the six mutes ; but goes still further, as the
reader will presently see, with respect to the liquids, which it leads us to
suspect are not only more numerous than is usually supposed, but also
admit of consistent classification of another nature, under each organic
series viz., that they are all lenes, and none fortis. Of these liquids, it
proves that not only I and r are both explodent and continuous, but that,
viewed as the former, they are simple liquids, and, viewed as continues, they
are aspirated liquids ; and, moreover, that the same applies to all the
elements usually classed under continues.

(e) That / and r, commonly called liquids under the old as
well as new arrangement, but, in the latter, shelved under the
head of ancipites, are probably only two members- of a numerous
set of analogues viz., seven, including, one guttural,/owr linguals,
and two labials ; that they are tenues, being lenes forms of t, and
have their corresponding aspirate forms, which again are lenes
forms of (aspirate); and it is only in the case of these aspirated
liquids that the term continuous is at all applicable to con-
sonants.

NOTE. T would here remind the reader of these two elements, called
ancipites by Dr. Lepsius, and by him considered as both " explodent"
(formed by a contact and apertion of organs) and continuous, that, by my
inferences from the Sechwana, there are tenues forms of them both, and
also aspirate forms, which may occur separately and distinctly. According
to the principles of this treatise both are explodent,* whether tenues or
aspirate, but both are continuous only when aspirated.

(f) That certain articulations viz., Germ. gutt. ch, Sech. g,
s, tk (English), and /(above indicated by 5, s, z, and/'), usually
regarded as continuous consonants, and therefore, in such in-
stances, according to what has preceded, aspirate forms have
also their simple "explodent" (elementary liquid) forms (q, s,
2j/), just as some writers have supposed both I and r to have,
and as has been proved in the second chapter.

(<jr) The probability that both sf (in parts, parks, and harps},

* Professor Miiller, I observe, also admits that in the formation of liquids,
among which he includes r, I, and a guttural 'h, there is " an approach or a
very slight contact between the organs." Proposals, <&c., p. xxv.

f The following occurs in an able American periodical, the Bibliotheca

137

and the English th (in thin}, in their simple as well as aspirate
" explodent" forms (tenues and aspiratas) the former of which
are not usually included in phonological classifications are not
only analogues of the simple and aspirate forms of r and / re-
spectively, inasmuch as they are all liquids; but, moreover,
organically, kindred forms of these two elements, and therefore
also lenes forms of t and ; and that the correlation usually sup-
posed to exist between them, as indistinctly represented by
the letters s and th (thin), and the other kindred articulations z
and th (thine), is not analogous to that between each of the three
pairs of mutes k-g, t-d, p-b ; but that these simple and aspirate
forms are both lenes, and z and th (thine) vocalised, modifications
of these aspirate forms.

NOTE. It must have struck the reader that we have no satisfactory
terms by which to express the nature of the distinction between the in-
stances of any pairs in this binary arrangement of consonants, but it is
necessary that we should hear in mind the peculiar nature of those which

Sacra, so recently asOct, 1860 (p. 831), and affords another instance of theusual
loose mode of forming deductions from the physiology of the human voice :

" 8 is a sui generis sound, which, like the sponge, mediate, as it were,
between a vegetable and an animal, or the bat between birds and quad-
rupeds, occupies a sort of middle ground between a consonant and a vowel,
uniting the characteristics of them both." B. W. Dwight.

This is, of course, only a circumlocutory way of saying it is the only semi-
votvel or semi-consonant, but the same writer elsewhere calls $ and a lingual
or dental aspirates,) which I have admitted; but only on the principle that
the continuous forms are tenues-liquids + the spiritus asper) , and distinguishes
the latter from the spiritus as follows :

"H is not so much a consonant as a breathing. It differs from the
sibilant, physiologically, only in being a breathing through the whole open
mouth, with the tongue at rest on its base, and the teeth apart; while the
sibilant is a breathing through the teeth, in a nearly closed state, with the
tongue against the upper teeth. H and are therefore both breathings,
and differ only in the different positions of the tongue and teeth. The
sibilant and aspirate have accordingly an etymological as well as phonetic
parallelism with each other, &c." Ibid. p. 839. (The italics are substituted.)

Whereas, the actual difference may be simply expressed in the following
formula: h = spiritus asper. The sibilant (aspirated or continuous) s ~ a
tenuis-liquid (dental lingual) + the spiritus asper.

138

were called semi-vowels under the ancient arrangement, and continue under
the new, and their similarity, essentially r to all called liquid-lenes (whether
simple or aspirate)^ under that suggested by the Sechwana, and that, in
respect to the mutes, the terms fortis and lenis short and long are perfectly
compatible. When most of the writers who reason upon a physiological
basis admit that the difference between vowels and consonants consists in the
former being caused by free emissions of both the voice and the breath,
and the latter by the check, interruption, or impediment of the breath, and,
moreover, use such expressions as quickly slowly, complete incomplete,
intense firm, strong weak, in reference to the degree of closure, contact,
collision, or compression of the articulating organs, it seems strange that
the idea of relative time, which appears to be implied in every antithetical
expression, has not also suggested the terms short and long, which would
convey, in the case of the consonants, a difference, or a correlation, pre-
cisely analogous to that between the vowels; for in all cases the position
of the organs is the same in the respective pairs, whether vowels or con-
sonants. In the case of vowels, proportion of time is distinctly the element
of the difference ; in the consonants, it is only less distinctly the element
in the interruption of the breath. At all events, quantity is a term quite
as applicable to the binary arrangement of the pairs of consonants, as to
that of the vowels.* Admitting all this, it must be the more satisfactory
that the terms lenis and long are quite compatible to express the quantitative
nature of all liquids. For instance, it will be immaterial whether we give
the following nomenclature:

Tennis, Aspirate.

Mutes fortis-lenis fortis-lenis

Liquids lenis lenis

Or Mutes short-long short-long

Liquids l n g long

In order to enable the reader to compare the terminologies of Professors
Lepsius and Muller with that of this treatise in the classification of
the so-called semi-vowels, I arrange them as follows, the vertical columns
showing equivalents :

Lepsius Liquids. Fricatives or Continues.

(Including r and I.) fortis and lenis.

Merely alludes to continuous forms-
of r and I, and including both.
5 and z

*ft(in) th(ia)

eh (Germ, gutt.) e (Arabic)

/
and the spiritus lenis and aspirate.

* The following will not be out of place here, though intended by the

139

Max Muller Liquids. Flatus or [Sibilant*.

(Including r and I, asper and lenis.

and 'h, a breathing.) Not including modifications of r and
I, but all the rest of above consonants
in the same order, and the spiritus.

Arrangement

gested by the [ Liquidise.

tenuei aspiratas roc.

Including simple Including aspirated and vocalized
or elementary modifications of the simple forms
forms* of s, th, of r, I, as well as of *, th, ch, f,
/, Germ. g. ch, &c., and consequently all above
&c., as well as consonants, viz.:
of rand I (viz., s s ( proper)

*,,/,?.) 2(th 3 ) I (th)

/(Germ ch) q (Arabic )
"f ( v proper)

The pure spiritus is classed as a
separate and independent ele-
ment.
As to quantity, all three divisions are lenes.

(A) That the elements m and n, usually called liquids, are not
properly so in their tenues forms ; but, so far as regards the
organs by the contact of which they are articulated, both they
and their analogue y- (ng, n, of Lepsius) are strictly mutes. On
the other hand, viewed as nasals, their aspirate and vocalised
forms are liquids. But they differ from the proper liquids I, r,
&c., in that, as will be shown in the sequel, these are liquid
with all other faucal consonants, and form the post-positive ele-
ments in diphthongal combinations; whereas m, n, and <?,
are only liquid with their cognates b or p, t or d, and k or g,
respectively, and their aspirated or vocalised forms, and generally

learned writer to have quite another application : " The consonantal, like
the vowel, elements of speech, have their different degrees of weight; and
their weight is hut another name for the amount of their phonetic force, or
the density, as it were, of their phonetic substance." DwigTit, in Bibliotheca
Sacra, p. 272, April, 1860.

* These are the consonants to which I allude as being entirely ignored
in classifications.

140

form the prepositive element in such combinations e.g., mp, nk,
nt\ &c. ; but these do not form diphthongs. This is only the
case when, combined with their cognates, they form the post-
positive element, as pm, tn, k?.

(i) That all the liquids, as well as the three nasals, and also
the spiritus asper, admit of another accessory element called
vocalisation.

Principles at Work in variable Pronunciations.

It is generally supposed that " the exact place of contact" of
two organs " can never be fixed with geometrical precision, and
that by shifting this point forward or backward certain modifica-
tions will arise in the pronunciations of individuals, tribes, and
nations."* But this has the appearance of being too general a
conclusion ; at all events, it requires proof. If correct, it will, I
think, apply chiefly to cases in which simple consonantal articu-
lations are compounded with vowel dipthongs, i.e., the numerous
forms of the "palatals." It is strange that while only one
capacity of hearing is allowed, every shade of pronunciation is
ascribed to a different disposition of the organs; and it is fre-
quently forgotten that one element may sound variously to half-
a-dozen ears.

It is chiefly in regard to the gutturals that there is a difficulty
in identifying the powers of an element in different languages.
There is not half so much dispute in respect to the linguals, even
though these include as many articulations as the other organical
divisions together ; and still less about the labials, which on
this account are rather arbitrarily pronounced by Professor M.
Miiller, "the most constant sounds in all dialects," f whereas it
is simply because the mode of their formation is the most
palpable to the eye as well as to the ear.

However, I think that the variable pronunciations of different

* Proposals, &c., p. xxxiii. f Ibid.

141

tribes and individuals is regulated by certain principles, which,
if carefully examined, will very much simplify the subject ; and
which it is necessary for me to sum up in review of the preceding
classification suggested by the Sechwana language.

(1) I have referred to the term quantity, expressive of the
difference between k and g, t and d, &c., and between k, t, &c.,
and the liquid forms of g and d respectively (viz., q Oriental
gutt., and r, I, s, z(th) linguals). The correlation of the former
pairs is what has generally been admitted ; but that of the latter
has, I believe, been first pointed out in this treatise. Both g and
q differ from k in quantity,

(2) But between g and q (Oriental), which are both lenes
relatively to the fortis k, there is another difference arising from
some accessory or accidental property in the latter, viz., that of
being prolonged in the utterance ; the one is a mute, and the
latter a liquid. The same applies to the difference in the
linguals between the mute d and the liquids r, I, s, and z(th).
In the case of the gutturals, the position of the organs is the
same in every instance, but there is a quantitive difference
between k and g or q, and a difference in property between g
and q, though both of the same quantity.

(3) Again, we find that fiie mutes k, g t &c., can be pronounced
in another very different way, viz., with first modification of the
spiritus, e.g., k or g. But the Oriental q as a liquid can be
pronounced in two other very different ways, viz., with first
modification of the spiritus, as in the Germ, ch gutt., or Sech-
wana g (q of my table) ; and with the second modification of the
element of vocalisation as in the Arabic (q of my table,). In
these two or three modes of pronouncing the same element, the
same organs are at the exact place of contact ; * but the differ-

* If this does not apply similarly to the linguals, this is easily accounted
for by the greater mobility of the point of the tongue required in the forma-
tion of them.

142

ences consist in the modifications by accessory elements. We
have thus

Difference in quantity between fortis and lenis, k and g or q.
property lenes g and q.

modification ( tenue * as P irate \ q, q } q.
3 \ and voc. $w

In the case of the gutturals there are thus seven articulations
formed at one point of contact of the organs, " without shifting
this point forward or backward."

Now I have remarked, that "either the fortis or lenis form
of any consonant may be enunciated with different degrees of
distinctness; but its quantity relatively to that of the other
instance of the pair, remains constant." These remarks will
apply to any of the seven articulations in the above examples.
To repeat the useful description by Sir John Stoddart, " It is
not to be understood that either the one or the other articulation
in each pair does not .admit of nice shades and discriminatory
touches, as it were, perceptible to some ears and not to others."

It is to such variations of pronunciation that Professor M.
Miiller gives the appellation of "Dialectic modifications." He
writes

"Where this variety of pronunciation exists only in degree, without
affecting the nature and real character of a guttural or dental consonant,
we need not notice it. Gutturals from a Semitic throat have a deeper
sound than our own, and some grammarians have made a new class for
them hy calling them pectoral letters. The guttural flatus asper, as heard
in the Swiss ach, is deeper, and as it were more pectoral than the usual
German ch."

Again,

" Thp Swiss ch is, according to Wallin, page 21, the same as the
Arabic ."

Again,

" But though there is a distinction between the ch as heard in loch, and
the and of the Arabs, as described above, vet it is not necessary to
admit more than one type of the guttural flatus asper."*

* The italics are mine.

143

Now, the Arabic is, according to Lepsius, the equivalent of
the Germ. gutt. ch, which cannot therefore differ much from the
Swiss ch, nor any of these from the Scotch ch in loch, the Dutch
g in dag, the Sechwana g, the Naman gh of Tindall, and the
Kaffir r; every one of which is to be looked upon as a quantitive
equivalent of the others, i.e., an aspirated liquid guttural, and any
of them may be chosen as the type. The same will apply to the
tenuis form of this consonant, viz., the q (Oriental), which I have
suggested has equivalents in the form of the Kaffir g and the
Danish g; also to its vocalized form the Arabic the equivalents
of which are said to be the Northumberland burr and the
Hindustani gh. It is the business of the linguist to put these
facts to the test, and raise upon them something more solid than
mere conjecture.

Throughout these analyses it has been my aim to express my
meaning by a uniformity of nomenclature ; but where one has
to do with a plurality of systems, this is in a measure imprac-
ticable. The terms simple and aspirate " exploderit," and some
others, have merely been used for the purposes of this inves-
tigation ; but, as distinctive terms earlier in use, such as tenuis
and aspirate, liquid and mute, fortis and lenis, express all that is
required, I propose to adhere to these in future, in the treatment
of the various forms of compound consonants. Hitherto I have
only attempted to classify simple or elementary articulations,
and shall in the following section proceed to the consideration of
simple consonants in their mutual combinations.

144

11.

ON THE APPLICATION OF THE NOMENCLATURE AND ORTHOGRAPHY
ABOVE SUGGESTED, TO SYSTEMS OF CONSONANTAL PERMUTATIONS

IN OTHER LANGUAGES.

At the outset of this treatise, I had never given attention to
the subject of permutations as existing in other languages. I
merely referred to the remarks of Dr. Richardson, commenting
on Home Tooke, as rather crude ; and in a foot-note (p. 6) gave
a quotation from the North British Review, in which the learned
writer does not suppose that Grimm's law, with its nine equations,
is without a foundation in the history of language, but considers
it exaggerated by being run out into a vicious circle. The
following is what has met my eye since, and as it is contained
in the first Encyclopaedia of the day, and professes to give a
concise view of all that inductive philology has accomplished in
what may be called the science of the " embryogeny" of human
speech, I give it at length, especially as it is probable some of
my readers will not have the means of referring to that cumbrous
but invaluable work.

" Its claim (Philology), however, to rank as a branch of inductive science,
does not rest merely on its services in classifying the phenomena and
interpreting the facts of language. It has also proved itself able to discover,
like other inductive sciences, the general laws which prevail among the
phenomena. One of the most important of these general laws is that of
the ' transposition of sounds' (Lautversehiebung) ; or, as it is sometimes
called, ' the law of the interchange of the mutes," which had been im-
perfectly indicated by Rask, and which Grimm demonstrated completely
in its application to the Greek (Latin, Sanscrit), the Gothic, and the old
High German (Deutsche Grammatic, vol. i. pp. 584, sqq.) and which Bopp

145

has extended to the Zend and Lithuanian ( Vergleihend Grammatih, pp. 78,
foil.) The general law is thus stated :

Labials.
In Greek (Latin, Sanscrit)

P answers to the Gothic F, and the old High German B or V
B P f

F B, P

Dentals.
In Greek (Latin, Sanscrit)

T answers to the Gothic TH, and the old High German D
D T Z

TH D T

Gutturals.
In Greek (Latin, Sanscrit)

K answers to the Gothic -I Q. } 1D1 ,j\ rand old High German G

G K CH

CH (H) G K
" Or thus :*

Greek (Latin, Sanscrit.) Gothic. Old High German.

Tenuis. Aspirate. Medial.

Medial. Tenuis. Aspirate.

Aspirate. Medial. Tenuis.
" One example of each interchange will explain the application of this law

Greek. Gothic. Old High

(Latin, Sanscrit.) (Old Norse.) German.

P, F, V, (B)

Hovs=ir6S-s; Pes=*ped-s ; Padas,

Fotiis, Vuoz.

B, P, F,

OopuBos; turba,

thaurP, doroF.

| H }B,P,

fcjjyos, Fagus

Beyki, Puocha.

T, TH, D,

65ovs=6S6vT-s; dens=denT-s , danTas,

TunTHus,ZanD.

D, T, Z,

<5Aoi>s; Dens; Dantas,

Tunthw, Zand.

TH, D, T,

Qvyarno,

Dauhtar, Tohtar.

K,{H }G ,

iKvpos, soCer,

SvaiHra, SchwaGer.

G, K, CH,

TBVOS, Genus,

Kuni, CHunni.

H ^'}G,K,

Xopros, Hortus,

Gards, Karto.

* This summary, reduced to my nomenclature, would read as follows :
Fortis mute. Aspirated liquid. Lenis mute.

Lenis mute. Fortis mute. Aspirated liquid.

Aspirated liquid. Lenis mute. Fortis mute.

But there is no mention of temt-es liquids, which ahound in the European
tongues, and the existence of which, in a scale of sounds, it is the object
of this work to prove.

146

'' Mr. Winning has pointed out a curious interchange between the Greek
and the Gothic, with regard to the relations established by this law (Manual
of Comparative Philology, p. iii). Other modifications require to be intro-
duced ; and Dr. Guest attaches so much weight to the exceptions which
he has noticed, that he has arrived at a conviction of ' the general un-
soundness of these celebrated canons' ('On the Elements of Language,
their Arrangements and Accidents.' Proceedings of the Philological Society,
vol. iii., p. 180.) The great majority of philologers, however, acquiesce in
the general validity of this theorem of interchange. Blinsen calls it ' one
of the most fertile and triumphant discoveries of philological ethnology'
(Report of the British Association for 1847, p. 262) ; and Max Muller
accepts it as a proof of ' the systematic regularity, the almost absolute
certainty, to which the phonetic laws of different languages can be brought.'
(Edinburgh Review, October, 1851, p. 319.)" ENCYCLOPAEDIA BRITANNICA,
vol. xvii., p. 539.

The above statement of the general law of Rusk, Grimm, and
Bopp, is, after all, the mere record of a very general inference
that certain articulations in one language answer regularly to
those in others, and amounts to a simple synopsis of facts like
that at page 15 of this work, of the permutations in the Sechwana
language, discovered by the missionaries some forty years ago,
though of a different nature. The Indo-European law is based
upon a comparison of cognate dialects, which had lost all ap-
pearances of cognation by reason of historical vicissitudes, and
has been evolved from a mass of heterogeneous materials, which
it has become a subjective means of rendering more accessible
and comprehensible to the efforts of the mind. Its discovery in
the evolution has shed light on that cognation. It amounts to
a mere series of palpable facts, exhibiting the relations of other
facts ; but I am not aware that any attempts have been made
to reduce them to abstract laws or principles. The same, to a
certain extent, may be said of the laws of permutation in the
Scchwana. The first missionary who reduced the language to
writing stumbled upon these laws as so many difficulties ; but
the more he observed of them and their constancy, the more
easily he made a comprehensive grasp and maintained a firm hold

147

of his knowledge ; and by noting them, in fact made the language
more easily acquirable by others ; but, till the humble attempt
made in this work, I am not aware of any efforts to reduce these
palpable laws, as fresh objective facts, to still more abstract sub-
jective principles. One would think that we have to search in
such singular phenomena for " a phase of progress, of growth,
of history," for a department of science explanatory of both
lexicology and syntax.

Moreover, the above general law of what may rather be called
comparative or dialectical phonology than phonics, however
striking, labours under a disadvantage alluded to throughout
this work viz., that it is necessarily based upon a very copious
induction of materials of questionable analogy, culled from both
purely historical and purely phonetical facts, in dead and living,
archaic and vernacular examples of speech. Therefore the
remark of the North British Reviewer, that " it has been run out
into a vicious circle," seems not without foundation, and it can
therefore only be regarded as empirical. The late Richard
Garnett, an eminent philologer, considered that the above scale
of permutations would " admit of being considerably extended
beyond the limits assigned by Grimm and Pott." P. 254.

Of the two, the scale of permutations laid down by the humble
and plodding missionary is by far the more important in a
scientific point of view, inasmuch as it is based upon the ob-
servation of the " living traditional pronunciation," and that in
one language. When looked at in this light, what is usually
called " Grimm's law" seems at present destitute of value,
further than as a subjective means of rendering the relations
between the archaic forms of the Greek, Gothic, and old German,
and one or two other cognate dialects, more easily accessible for
the mind.

Of considerably more importance, and withal more satisfactory
than such general speculations, are the following peculiar per-

148

mutations* in Welsh, inasmuch as they also, like the Sechwana,
are gathered from the vernacular pronunciation :

. Gutturals.

Car (a kinsman) kf fortis "| _

Eigar (his ,. ) tog to lenis } TenU6S mutes '

Ei char (her ) toq to lenis Aspirate liquid.

Vy nghar (my ) to Arf to lenis / Aspirate route, with the cognate nasal
6 s I preceding.

Linguals.

Tad

(a father)

Ei dad (7ws
Ei thad 7ter

to d
to z

Vy nhad (my ) to nh to

Pen

Ei ben (his

Ei phen (her

V mben iy

LSame mutations.

(Simple aspirate (spiritus) with the cog-

1 nate nasal preceding, which is w<?a?-

( ised,probcMy by absorption with thed.

Labials.

(

1 Same mutations.

o
) to mh

1 As in preceding cxample,the\) being pro-
| bdbJy absorbed inthevocalised nasal.
The other examples cited are not so regular, but I append the following :

Gutturals.

Gwas (a servant) gw lenis mute.

Ei was (his
Vy ngwas (my

Duw (a god)
Ei dhuw (his )

Vy nuw (my )

Bar a ( bread)
Eibara (his )
Vy mara (my )

) to w lenis is lost.

, o r lenis is absorbed in the cognate
) to nw \
f I vocalised nasal.

Linguals.
d lenis.
to z (th 3 ) to lenis vocalised liquid,
r lenis absorbed in the cognate vocal-
n \ ised nasal.

Labials,
b \

to f (v) I Same mutations as in preceding case.

to m

* These, as they occur in the first column, I have found in Latham's
" Eng. Lang.," vol. i., p. 327, cited from Prichard ; but I have tested them
by the speech of a genuine and intelligent Welshman.

f The bold letters in the middle column are those arbitrarily assumed in my
tables, and the terms in the third column have reference to my nomenclature.

149

Of the above "regular mutation of initial consonants," called
by Garnett the " Celtic process," he states they " are changed
into others of the same organ, to denote a diversity of logical or
grammatical relation ;" and adds, " the entire system is, so far
as we know, peculiar to the Celtic tongues, and it exhibits a
phenomenon as curious as it is difficult to account for."

On examining the living pronunciation, I ascertained the Keltic
process to be strictly regular in the case of the tenues-fortes con-
sonants k, t, p, except in one instance. The second example
under each consonant is a change to the corresponding lenis ;
the third into the aspirate liquid (" fricative" or " flatus",), which
is also a lenis ; the fourth presents an anomaly, in that, in one
case (&), the consonant becomes an aspirate mute, and in two
cases, t and p, the mute itself is absorbed in the vocalised nasal
preceding e.g., p.g (instead of y-Ti) ; nh, mh.

In the case of the tenues-lenes g, d, b t one anomaly occurs
only under the guttural. The second example, under the lingual
and labial, is a change to the cognate vocalised liquid, i.e., Z (th?),
and f (v), for, by the analogy of these two examples, one would
have expected the corresponding articulation under the guttural
to have been qw viz., the voca&serf-liquid-guttural. Under
tlie lenes, again, the anomaly alluded to under the fortes does not
hold good, for here one finds the guttural, lingual, and labial,
respectively, becoming p, h, m.

It is, probably, principally in reference to this peculiar process
of permutation that Mr. Garnett writes of the Keltic language
It " appears to be the most ancient, the most singularly con-
structed, and the most true to its original form of all European
tongues." It' the greater perfection and constancy of the
Sechwana process of permutation be considered in conjunction
with its equally remarkable and normal syntactical structure,
the same remark will apply to it in comparison with other African
languages of the same family.

150

It is to be expected that a process of permutation in one lan-
guage must shed some light on that of another, especially if
arrived at by the same mode of observation. One case occurs
to me at this moment. The Sechwana contains proofs that
t and d, p and b, are correlatives or pairs, but it fails in data to
prove that k and g are similarly related (see p. 20) ; whereas,
the Keltic examples fear to ei-gar, tad to ei-dad, pen to ei-ben, are
conclusive and satisfactory in respect to all three pairs of con-
sonants. It would be difficult to find such a beautiful example
in Grimm's law to corroborate any inductions from even the
Welsh or Sechwana, and yet the above result, at which we
arrive in both these languages, is founded upon what Professor
Miiller would once,* perhaps, have called " the irregular utterance

It is remarkable that, in Sechwana, precisely the same initial
permutations, .presented in the table of verbs at page 15 of this
work, are repeated with great exactness in the formation of
plurals in Li from nouns in Lo, with one or two omissions only,
and also, in some cases, a few exceptions in addition to the
normal forms.

Mutations. Singular. Plural,

btop Lobopo (form) Lipopo

PlxcEPTioN. ob to mp e.g., Loobu (brack ground), Limpu.

* 1855. The following, however, written six years later, shows less
scholastic prejudice in the investigation of truth. "In the science of
languages, languages are not treated as a means; language itself becomes
the sole object of scientific inquiry. Dialects which have never produced
any literature at all, the jargons of savage tribes, the clicks of the Hottentots,
and the vocal modulations of the Indo-Chinese, are as important, nay, for
the solution of some of our problems, more important, than the poetry of
Homer or the prose of Cicero." Lectures on tJie Science of Language, by
Max Mutter, p. 23.

151

Mutations. Singular. Plural,

c* immutable Locwiisa (cuticle side of Licwiisa

a sldn)

g to kh (*') Logon ( (wood) Likhon (n)

htoph(p") LohaTo (fissure) Lipharo

h to kh (k") Lohihin (umbilical cord) Likhihiri or Liphihiri

NOTE. I have already remarked that this mutation is probably euphonical,
a,nd was the only exception in the case of the verbs.

Ar immutable Loketla (fragment or Liketla

chip)

kh (**) Lokbaru (eggshell) Likharu

J to t Loleme {tongue) Liteme

m immutable Lomati (plank) Limati

n Lonaka (horn) Linaka

n (P)>, ~Lop&m\i( flank) Lijfamu

P Lopalo (scab, in goats) Lipalo

P n (p") Lopheqo, (hollow pots- Lipheqo

herd)
r to t

EXCEPTION. r to nt eg., Lori (cord), Linti
r to th (t') Lorole (dust) Litbole

s to ts Losika (generation) Litsika

EXCEPTION. s to nts e.g., Losi (eyelash), Lintsi.
h to c Loshwaela (skeleton) Licwaela

EXCEPTION. sh to no e.g., Lusho (spoon), Linco.
t immutable Lotolo (fire steel) Litolo

th (t') Lotheka (loin) Litheka

EXCEPTION. th (t*) to nth e.g., Lotha (membrane), Lintha
tl immutable Lotlowa (net) Litlowa

tlh(tl'),, Lotlbakore (side) Litlhakore

ts Lotsatsa (thin) Litsatsa

to k Loato, or kato (increase) Likato, or maato

EXCEPTION. to ny e-g., Loetb (journey), Linyeto.

But this will not apply to the formation of other plurals, save
in a few exceptional cases i.e., nouns in Bo, with plurals in
Ma; in Se, with plurals in Li; or in Mo, with plurals in Ba t
Ma, or Me, as a general and constant rule, form these without
a change of the initial.

* C of the missionaries, as before stated, equivalent to ch in Charles,

152

Under Se, we find the exceptions

I to t Seleru (beard) Literu

, ,

or to ^ Seqaqa (fc&) li

ft j I SebhaqaJ

In the formation of plurals in Ma, from nouns in Le, there
are more rariations ; but these are only exceptions, which can
be confronted with regular forms.
8 (prob. s) to r Lesapo (bone) Marapo

R ea forms -I Lesa " ri (night g arment )

\L0rotobolo (done) Marotobolo

O (prob. t;) to r Lecoha (hole) Maroha

.. _ J Lerothori (drop of rain) Marothon

~\Leouti (cloud-shadow) Maeuti
C to b Lecoyo (arm) Mabo'qo

Reg. forms .-{ L T ^ man ( e <? ck > Maeomane

(Lebori (rat) Mabort

CW to dy Lencwe (stone) Madye

Reg. forms :- (

.
voice)

ts to 1 Letsatsi (swn) Malatsi

,. _ j Letsatsa (squirrel-hole) Matsatsa

~ \Lelata (female slave) Malata
ts to b Letsele (grain of corn) Mabele

,. ( Letsela (rag) Matsela

Reg. forms :-\ ^.^ ^ }

Letsaina, or) Marama, or

tsh(ts)tor LeUma \ (cheek) Mft|ama

Reg. forms : | Lets * ts *> or (louse-nit) Matsetse

\ Lentsetse Mantsetse

sh (prob.sy) to r Leshophi (o?rf stfc) Marophi

Reg.forms. Leshoril (n T eof ^^Af^/.oH
(cuhar custom) /

Among other permutations in Sechwana, we find
C to r Cwa (come out) perfect rule

Causative rusa (Stock,
with milk, of a cow

before calving)

1 to r Cwela (come out towards) mocweri (fountain)

8 to sh Cy) Risa (fcmZ) tisho (herding)

tS to c Eetsa (listen) theco (listening)

ts to s Tsimo (garden) rnasimo (gardens)

153

Nwaqa (year)
Leba (look, Tr.)

linyaqa (years)
ledywa (pass.)

huh* (pay wages)

ruywa (pass.)

Hapa (bind)
Mocweri (fountain)
Mariri (hairy)

hacwa (pass.)
mocwetsana (spring)
maritshane

g w (Uw) to ny

b (prob. bi) to dy

h (prob. hi) to y (Germ.
ch in ich)

p (prob. pi) to c

r to ts
r tots

All the preceding regular permutations are those upon which
the principles of this work have been based. The exceptional
forms (all in bold type) are those other instances of interchanging
consonants to which I referred in a note at page 15, and which,
together with the following, showing the affinities of the Sech-
wana and Isi-Zulu, represents pretty nearly the general character
of the numerous irregular dialectical variations observable in a
comparison of all cognate African languages.

List of Permutations showing the Affinities of the Sechwana and Isi-Zulu
Languages.*

Sechwana. Meaning inMh Language,. Isi-Zulu. S. Z.

Atlhola. (adjudge)

Dy& (eat)

Tl& (come)

(crush or mince
with the teeth)
(well, beautiful,

etc.)
(backbite)

(laugh)
(shorten)

(rinse out)

Sentle

mega
SAepola

Cokotsa,

Atlh&m&
Qano^a
tffcalemela

Isi-Zulu.

A/tZola

Hl&

HIM&&

Lsihle

Jtl

Hleba,

S(prob.s) *

/ZZeboka

Hleka,

#7epola

sh (sy)

flZukuWa

fc(-toy)

Its J

(open wide) Afcama

^CanuAra
(reprove abruptly) Kulimela

tl c

* All the Zulu examples are from Doline.

f Kakatha is also used with the same meaning.

The click.

154

Seckwana. Meaning m both Language,. Isl-Zulu.

S. Z.

A^ara

(become clothed)

Am&ata

p xnb

Atla.

(kiss)

Anga

tl )

Befca

(jerk flesh)

Eeng&

k i

ItSG

(know)

Azi

ts 1

Setlhoh

is. (dangerous \
person) L Ihlo*i
k. (a ghost) )

Tlhwafi

(species of make) Hlwa^i r *

MoruZi

(shade)

UmtunM

Ru//a

( s. (pay wages) j
(z. (reward) f Vu * a

27ala

(become full)

^ala

Letsha.

fs. (lagoon)
1 z. (ri^pk)

ts 1

. z

/s. (breathe

hard, or

Shuma.

J hiss of a

j serpent)

' // uinu

sh (= sy)

jz. (toAre %

V surprise)

Mosari

(woman)

Um^ali

tfona

(themselves)

Zona,

C ( = tsy)

Phaceqa

(bespatter)

Baceka

Q (= %) ^1

<2ama

(milk, v.)

(7ama

JOamelo

(pail)

Isicamelo

kh y c (cZicA)

Seba

(slander)

(7eba

Tima

(extinguish)

Cima

t )

Boruiu

(dulness)

Ubutuntu

t nt

Sebate

f s. (rag)
(z. (scar)

r Isibawda

t \ nd

Serene

(^Z)

Isitende

th(O J "

.ffTiaola

(CM Off)

<?aula

kh (&) )

tfobo

(mantle)

Gubo

k 1 g

Pitae

(the quagga)

Ibisi

Is. (the gums) \

Marini

z. (gap of a

Isisini

tooth) \

Coma

(speaTt a strange oma

C (= tsy)

language)

Ka mosTiof

(to-morrow)

Ngomso

sh (= sy)

* All the

Zulu examples are

from Db'hne.

f The sh

here is often pronounced like/u-.

155

Jtfwmfa6ott Language.

Isi-Zulu.

s. z.

(gather)

.Buta

ph ) ,

(think)

Goioda

P J

Is. (take a good)

handful) I

tabula

z. (separate) )

(male)

Doda.

(root out)

Z>omula

(think)

Goboda

(die)
(burn)

},a

Sh (=sy)

(spit out)

^ela

(female)

Isi/azi

(sweet cane)

Itn/e

C (= toy)

fs. (bury) ~)
jz. (conceal) )

.Pihla

(s. (person) ")
\z. (stranger) }

Um/o

(Z. (&Z0H7 tffo I

Mnya

(cloud)

Ili/u

j"s. (linger) )
(.z. (consider) )

Zin^Za

tl nhl

(a lewd female)

Isibanica

ntl nx (dick

(beard)

Idevu

r ^

(open)

Fula

b f v

(sheep)

Mim

k J

(regular body, or

MAadu

ph h

uniform class)

(dog)

Nja

(= toy) j (=<%)

(boil over)

Pu/)uma

h p

(once)

Kawye

w ny

\ s. (pay tribute) \

KeZela

(name, v.)

(love, v.)

Tanda

r U

(get)

Tola

C (= ty)

(dulness)

Ubuiuntu

(eat)

Tya

156

SecJiwana.
Hafci

-KVtucama

JTotama

Meaning in both Languages. Isi- Zulu .

j s. (earth)
(z. (down)

Pansi

s. (fall on the \

knees') L Qotjama
z. (crouch) )
(squat) Qotama

S.
ts

kh
k

q (ofoft)

The above permutations are, however, not all constant, for s in the
Sechwana often corresponds to s in Isi-Zuiu, p to p, b to b, I to I, n to n,
and TO to in. This list is a very different one to that given by Dr. Bleek
(See p. 40 Sir O. Greys Lib. S.A. Languages) ; but that able linguist
laboured under a great disadvantage viz., the absence of any complete
dictionary of the Sechwana language a thing which actually does not
exist, except, perhaps, in the manuscripts of two or three individuals. I
have had the great privilege of comparing my own copious vocabulary of
the Sechwana with the excellent Zulu dictionaries of Mr. D6h.ne and
Bishop Colenso, which has enabled me to give a more satisfactory
summary than has hitherto been published. It is, nevertheless, very far
from complete, having been transferred to these pages from a few rough
pencil notes on the margins of Mr. Dohne'S interesting work. I have pre-
viously stated the drawbacks to a careful comparison of these two important
languages, and in the above list, however precise I have endeavoured to
be in representing the powers of the Sechwana articulations, it has been
impossible for me to be sure of those in the Isi-Zulu, especially after having
met with a few discrepancies between the actual pronunciation and that
indicated in the orthography of some missionaries. An asterisk points out
every instance in which I believe an aspirate has been lost sight of, though
I may be wrong in some instances on verifying them by the actual
pronunciation.

The examples of initial permutation in the Sechwana, upon
which I have based the principles of a new classification, are
absolute permutations of either simple consonants, or their
simple modifications, or simple combinations, for which there
appears to be no other mode of accounting. But, of most of
the exceptional cases above shown (and of the relations between
the Sechwana and the Isi-Zulu) this cannot be said. Such
permutations must be accounted for upon principles or analogies
of either phonical constructions or phonical corruptions. They
show the extent to which one or other of these processes has

157

interfered with the normal forms of speech, and it must be
interesting to the European scholar to know that these very
processes, which he so delights in observing and describing in
the Arian family of languages, are at work in the Sechwana and
its cognate barbaric dialects.

Among the permutations showing the former process viz.,
that of phonical construction or growth, may be classed the
palatals, as they are commonly called, otherwise the " unstable
combinations"* of Dr. Latham, or the "specific modifications"}
of Professor Max Miiller. The Sechwana abounds in such
instances, which may all be traced to combinations of consonants,
with a post positive y or w, though in the root neither of these
elements may be perceptible. Such mutations as above, of
p to c (= tsy), h to yl b to dyw, ts to c (= tsy), s to gh (= sy),
are all permutations of this kind.

Contractive p^^ Wntten^the

In Hapa (bind), the change to the passive is hapiwa halsywa haeoa
Ruha (pay wages) ruhiwa \~uywa rushoa

Leba (look, Tr.) ,, lebiwa ledywa leyoa

The above analogy, again, would seem to shed some light on the probable

existence of an i prepositive to the o, in the formation of verbal nouns

derived from verbs with terminal sa.

Retsa (listen) Theco (listening) prob. constructive form, Thetsio

'Risa, (herd, of cattle) Tis^b (herding) Tisio

" The Neapolitan echiu, from piu,"\ is an analogous instance.
Another is " presented by the Spanish language, in which the
Latin li not unfrequently becomes a pure guttural, as in muger
for mulier, and hoja for folium, MoXtc and /uoyec exhibit the same
species of affinity." In Sechwana, for instance, dya\ is often

* Vide " English Language," vol. ii., p. 8, &c.
f Vide ' Proposals. &c" p. xxxv., &c.
\ R. Garnett, p. 241.
Ibid. p. 251.

j| Written by the missionaries ya and yoa ; but the d, in both oases,
thoagh mollified, is distinctly audible.

158

substituted for lea, in the possessive particle of nouns eg.,
lehuku dya ml (my word) ; the same with dywa for boa e.g.,
bogbbe dywa qaqive (his bread). Again, the Sechwana has sybna
for sebna, tsybna for tsebna* In cases where the spiritus occurs,
the combination is less perceptible, and the conjunct articulations
have more the appearance of a single sound. But, as has been
repeatedly implied, a proper investigation of these compound
permutations presupposes a thorough analysis of the vowel
System, and ought to form the subject of the third part of a
work of this nature ; for it is in all such examples that we have
to continue tracing a process of construction, and deriving from
this the phonical laws* by which the Sechwana language dis-
closes its own peculiar but natural growth from the common
elements of all human speech.

The permutations comprehended under the process of phonetic
corruptions are of greater vai'iety ; and they appear to form that
branch upon which the labours of the European philologer are
chiefly expended. While those included under the first process
are to be found in all their simplicity in barbaric languages,
these pertain chiefly to languages whose accidental or material
forms have alternately undergone disintegration and recon-
struction by the numerous circumstances attending the vicissitudes
of nations; though both processes have been at work in all
tongues, as the Sechwana, distinguished by the prevalence of
normal forms, and the paucity of exceptions, will alone show.

This subject would of itself supply materials for a whole
treatise, but a hasty survey of the different classes of irregular
permutations may not be uninteresting to the reader, and will at
all events throw light on some apparent anomalies in Sechwana.

1. There are those instances in which the now conjunct con-
sonants were formerly merely initials of conjunct monosyllables.

* Written by the missionaries shona and cbna respectively.

159

The following quotation, from the invaluable and fascinating
work of the late Richard Garnett, will clearly illustrate this
principle :

" Even many of the words usually regarded as Sanscrit roots are capable
of being resolved into still simpler elements. For instance, the root i
denotes to go (Latin i-re, Greek ievai) ; ri, also to go, may very possibly
be a compound of ra + i = per g ere ; tri (to pass), ta + ri q. d , go
thither; stri,to strew, or spread, a further formation with the particle sa t
and so of many others. Our readers will find much ingenious speculation
on this subject in Potts' ' Etymologische Forschungen.' We consider many
of his conclusions as highly deserving of attention ; but we do not feel disposed
to agree with him in referring the above prefixes to the Sanscrit prepositions
in their present form, which is evidently not their primeval one. We think-
for example, that tri is probably compounded, not, however, with the
preposition ati, but with the pronominal or prepositional root ta. We
freely admit that all this is, in a great measure, conjectural, and requires to
be confirmed by a more copious induction from cognate dialects. Could the
fact be sufficiently established, it would afford scope for much curious
discussion respecting the formation of language, and might perhaps serve
as a clue in tracing the affinities of tongues commonly supposed to be
entirely unconnected. It is scarcely possible for two languages to be more
unlike than Sanscrit and Chinese, but it is by no means improbable that
both were at a very early period much in the same condition, and partly
composed of the same elements. Both consist of monosyllabic roots ; and
a few more pronouns and particles, employed copiously in the connexion
and composition of words, might have made the latter not unlike the former,
But while the component elements of Greek and Sanscrit have, as it were,
ci-ystallised into beautiful forms, Chinese, as an oral language, has remained
perfectly stationary, and is still, as it was 3000 years ago, ' arena sine calce.'"
Pldlolorjical Essays, p. 108

It is doubtful whether such a principle would always apply in
the case of a mute with a post-positive liquid. A Mochwana
only knows of two combinations of this kind viz., tl and ts
(tla, come ; tsimo, garden) ; but if you give him a foreign word
to pronounce, such as Bethleliem or Esther, he will invariably
syllabicise every consonant or spiritus, thus Be-te-U-he-me,
E-se-te-re'. At all events, he will only use such combinations as
those to which he is accustomed in his own speech, and even
interpolate these with vowels in some instances. This fact, crude

160

and superficial as it may appear, ought not to be beneath the
notice of the philologer. I know of no objectively true instance
in which either of the Sechwana combinations tla or tea can be
explained by this principle, but it is possible some may yet be
found.

A reference to my former remarks in the present work, in
respect to the influence of syllabic quantity on combinations of
consonants, will, no doubt, in connection with the above quo-
tation, be suggestive.

A proof that quantity does give rise to certain combinations of
consonants, is afforded by the following facts, which cannot be
ignored in the examination of this important subject. In the
English language there is very little difficulty in recognising the
pronunciation of words when the vowels are all elided, and
many persons avail themselves of this mode of writing as a means
of short-hand. Take, for instance, the sentence
T vd tli mntny Imst nsprblfrm sch sbjct.*

If we introduce only the accented or long vowel, in every
instance, the pronunciation is indicated with double distinctness.
T void tli mnotnj slmst nseprblfrm such subjct.

NOTE. (a) It may be observed that where any liquids, faucal or nasal,
occur as initials, the insertion of the preceding vowel is unnecessary, unless
it is accented, as nseprbl, almost.

(b) When a mute is followed by any post positive liquid, the intervening
vowel, unless accented, is absorbed in the consonantal diphthong e.g., pr
and Jcr in nseprable and mdckrel.

2. There are those instances of permutation in which the
organically distinct articulations " are in reality derivative sounds,
descended from a more complex element capable of producing both"
e.g., Greek &T and Latin bis, in relation to Sanskrit dwis ; Latin
bellum, for ancient form duellwn, &c., &c.f It is to such that

* This was sugested by the perusal of an article in " Evangelical Chris-
tendom."
f Garnett.

161

Dr. Donaldson, who was the first to notice the principle, has
given the name of "divergent articulations" It cannot be better
described than in the words of the learned discoverer, for ap-
pending which I make no apology, as they are contained in a
concise form in the Encyclopaedia Britannica, which may be
inaccessible to some of my readers.

" The older grammarians had only one name, metalepsis, for all inter-
changes, whether regular and easily explicable, as from p to b, or irregular,
and at first sight inexplicable, as from p to AT. The present writer was led
to an explanation of these divergent interchanges by an inquiry into the
nature of the Greek letter called the digamma, which he proved to be a
complex sound, consisting of a guttural combined with a labial (New
Cratylm, p. 110), and he extended the same principle to all cases in which
two words, undoubtedly of the same origin, exhibit articulations which could
not have been interchanged. In all such cases, he concluded that the
original form exhibited a combination of the two sounds. The brief but
decisive induction by which this law was established, in 1839 (New Cratylus,
1st Edition, p. 136), was greatly extended by Mr. Garrett in his valuable
paper ' On certain Initial Letter-Changes in the Indo-European Lan-
guages' (Proceedings of the Philological Society for 1846, vol. ii., p. 233,
sqq.) A simple example or two will show the application of this law. The
Sanscrit paktas corresponds exactly in meaning to the Latin coctus, and the
Greek iretrTos. But as p cannot pass into k, the Latin differs from the
Sanscrit in the initial, and from the Greek in the included sound, or, in
Grimm's useful terminology, they differ reciprocally in anlaut and irilaut
Now the Latin coquo shows us that the guttural in this case was not pure,
but that it was followed by a vocalised labial ; and it is known that even in
Cicero's time, coquus was pronounced quoquus (Quintil. Inst. Or. vi. 3, 47.)
The divergent articulations p and k, converge, therefore, in the compound
sound qv = kp, and the three words are accordingly reducible to an identity
of origin as well as of meaning. Again, we have in Greek Ke\aiv6s as a
synonym of /xsXas, ptXaiva, fikXav ; and with the exception of the initial or
anlaut, the words are identical in root or crude form. But we cannot derive
k from p, or vice versa ; and, according to the law, we must assume the
complex sound kp as the origin of these divergent articulations. For-
tunately, we are not left to an inference in this case, for Pamphilus, of the
school of Aristarchus, recorded the fact, that fie\a9pa, meaning " the rafters
blackened by the smoke,'' were anciently called Jc/tsXeflpa (Etymol. Magn ,
p. 521, 33.) Lastly, to take an instance in which we have all forms of the
process, the Latin vivus exhibits no traces of a guttural in combination with
the labial. But the perfect vixi, from the corresponding verb vivo, shows
that the inlaut at least involved a k sound ; whereas a comparison with the
Gothic quios = vivus indicates that qv was also the original type of the
M

162

anlaut, or initial articulation ; and thus we arrive with perfect confidence
at the conclusion, that vivus => qviqvus was ultimately identical in meaning,
as it is in signification, with the old Norse quikr, old Saxon quic, and
modern English quick." Ency. Britann., 8th Edition, vol. xvii., p. 540.

The above process was suggested to Dr. Donaldson by
an inquiry into the nature of the digamma, which he concludes
to be a combination of a guttural consonant with a post-positive
w ; in fact, a " palatal." And it is not improbable that many
of these complex articulations referred to will be found to have
been " palatals" (whether the post-positive letter was formerly y
or w), as well as mere combinations of simple consonants. For
instances of several languages having the included articulation,
while others have both it and the initial, see Garnett (Phil.
Essays, pp. 108, 250, 251, 258, 259, &c., &c.

I am inclined to apply the process detected by Dr. Donald-
son in the Indo-European tongues, to the explanation of an
anomaly occurring in the Sechwana system of permutation."
See table, p. 16, where I have

Verb. Verbal Noun. Verb with the Object- Particlet. Mutations.
i = self. m, n, n = me.

16. Sila Tsilo Itsila Ntsila s to ts

22. Tsenya Tsenyo Itsenya Ntsenya ts immutable

NOTE. Nos. 2, Cola (teyola), and 17, ShoJca (syoka, perhaps ly), are, so
far as regards the simple consonants, in combination, precisely analogous
examples, but, being palatals, I cannot notice them here.

It is true, in the Sechwana the mutation implies a diversity
of grammatical relation ; but, inasmuch as I have proved that
the whole series of Sechwana permutations of simple consonants
may be reduced to phonic laws, is it not legitimate to conclude
that these laws will explain any anomalies in their compound
forms ? May we not presume that the complex form of s was
ds,* and that the primary change is from dsilo to tsilo ? It is
possible that the former may yet be discovered in some interior
dialects. The example cited by Dr. Donaldson is almost

* My reader may be inclined to insist on dz, but a perusal of the 5th
chapter may convince him of the correctness of my conclusion.

163

analogous viz., m and km in the examples nt\adpa, K/jif\f6pa
In the Sechwana examples, the mute which is elided is pre-
positive to afaucal liquid ; in the Greek, to a nasal liquid. The
CEolic fipoSov for poSov, ppia for pii^a, are merely analogous
phenomena. I have, in the preceding pages, attempted to prove
that s, z (th), /, q (Oriental), are analogues of r and I, which
appear to occur more frequently in European tongues as the
included element ; therefore the same explanation must apply to
them. Of all these analogues, only s, r, and I occur in the
language ; and the only combinations in which a liquid is post-
positive, are ts (see above examples), and tl in the following, as
well as their aspirated forms (See p. 16):

Tlotla, Tlotlo, Itlotla, Ntlotla, we have tl immutable.
In the Sechwana, however, we have no example of an I or dl
being commuted into tl, so that there are no means of either
proving or disproving above conclusions. They will, however,
be suggestive to other minds. I am content to admit my
ignorance of the consequences deducible from the mere statement
of this anomaly, and from my lame attempt to explain it

At pages 20 and 53 I have^referred to the insertion of a k,
in the case of verbs commencing with a vowel or spiritus (See
p. 16

28. Ila, Kilo, Ikila, Nkila, is commuted into k ;)
and explained it on the principle of euphony ; but the following,
from Garnett, may suggest a different course of speculation
on the subject.

" Formerly the only method of connecting aXivSsu and xa\iv&w together,
was by supposing that a guttural had been dropped or assumed. But the
knowledge that the former anciently had the digamma, places the matter
in a different light, and makes it at all events probable that they are in
reality collateral formations." (Phil. Essays, p. 248.)

I have also^in the course of this work, referred to another
peculiar phenomenon in the Sechwana language, which may
if 2

164

perhaps be included under the permutations observed by Dr.
Donaldson viz., the tendency to substitute the spiritus or
pure aspirate for the aspirated mute or liquid, i.e., h for either
q (gutt. ch~),r, or b\ e.g., hbnafor 'gbna (there); hae for 'qae (home) ;
he for re (we) ; hela for bhela (sweep). By the analogies pointed
out, I conclude of course that the same will be found to apply to
I (Zulu and Welsh), // * s, z (th 2 ), d] and In proof of this
conclusion, however, I have not found any objectively true
instances in Sechwana speech, but have little doubt of the pro-
bability of their occurrence to the students, on a comparative
survey of cognate dialects.

Similar phenomena abound in the Indo-European tongues.
In the Anglo Saxon hrced, according to Garnett, " h represents
a more ancient guttural."f "The Slavonian greblo (an oar),
would, in Bohemian, become hreblo."^ " S, in Latin, almost
invariably" corresponds with the spiritus asper in Greek e.g.,
we? and super are exactly equivalent." Dr. Prichard writes :

" It is to be observed that h never stands as the initial of a word in Erse
in the primitive form, or is never, in fact, an independent radical letter.
It is merely a secondary form, or representative, of some other initial viz.,
/or s. It must likewise be noticed, that the same words which begin with
s or /as their primitive initial in the Erse, taking h in their secondary
form, have in Welsh h as their primitive initial. This fact affords an
instance exactly parallel to the substitution in Greek of the rough and
soft breathings for the (Eolic digamma, and in other words for the sigma.
Oivy, as is well known, stands for Fairy, ^'Evirepos for Feoirepos, and tirra
probably replaced a more ancient form of the same word, viz., aeTrrd ;' |
stands for oig ; vs and i'pTrw for <ri~s and erp7rw."||

In all such cases, it is frequently supposed that the spiritus is
merely substituted for the simple consonant, but all such facts

* The Sesuto, an impure dialect, has /' where the Sechwana has both h
and the still purer form bh (&').

f Phil Essays, p. 245.

I Ibid. p. 257. It is not stated whether the g in the Slavonian word is a
liquid Germ. gutt. ch (q"), or a mute.

Ibid. p. 107.

j| Quoted by Garnett. Ibid. p. 83.

165

as I have met with in the Sechwana go to prove that the spintus
has lost its consonantal form i.e<, the mute, which it formerly
modified, has disappeared ; in fact, that , or/, or th, or r, or I, in
all such cases, must have been aspirated consonants, and so far
complex elements. To be more explicit : upon the basis of the
Sechwana examples, he for re, hbna for qbna. I believe that the
fact of h being the secondary form, is a proof that the primary
forms were not simply the tenues r and q (Oriental), but the
aspirates r and \ (gutt. cli) ; or, in other words, the former + the
spiritus. In a comparison of dialects, it will not improbably be
found that an aspirated consonant in the one will be represented
by an aspirated consonant in the other e.g., nari (buffalo) is
usually written by the missionaries nari, and the Isi-Zulu equi-
valent is written by the Zulu missionaries nyati, but I would
venture to assert that it is pronounced ny&i ; for the Isi-Zulu
equivalent of the Sechwana 're (we) is not, as usually written, tina,
but rather tkina (t), as I have heard it distinctly pronounced.
The importance of precision in these matters cannot be over-
rated, as the following example will show, in any attempts to
make a subjective use of the analogy. UthiXo is the Zulu name
for Deity (not Uti^p as usually written). Orija is that mentioned
by visitors to the lake ^habi,* as used by the natives there. If in
this word the r is aspirated (/), there can be little doubt that
Orija and UthiKo] are cognate varieties of the same expression,
and that neither has any connection with Morimo of the Sechwana,
in which the r is a tennis.

It would be an easy matter to carry the illustration further,
and not an unprofitable task to attempt to arrive at other prin-
ciples in a comparative survey of South African dialects ; but
when a student is acquainted with the actual pronunciation of

* Usually written Ngami, but pronounced as above,
f Tbe click (x) in the one is probably a substitute for the palatal in the
other.

166

only one dialect in a comparative survey, there is much to
detract from the pleasure of his researches.

I shall close by placing before the reader an extraordinary
permutation occurring in the Sechwana language which may not
be uninteresting viz., the change of the tenuis combination ri
to ts (mocweri, a fountain ; mocwetsana, a spring), and of the
aspirate form ri to ts (nari, buffalo ; natsane, young buffalo).
For instance, one would little have suspected that the verb rusa,
to stock, of milk (of a cow before calving), was the causative
form of cwa (come out), but for the above analogy. The
occultation of the fact arises merely from the confusion of ortho-
graphies. Were cwa written with all its constituent elementary
letters, tsywa, it would be more evident. Precisely by the same
analogy is it that rule (has come out) is the perfect of cwa, and
burule (ripe, ready) of bucwa (ripen). The Batlhapin have again
cona (tsyona), for rona of the other tribes.

I shall not attempt to explain this phenomenon in the Sech-
wana, but leave it to the speculations of those who have more
collateral knowledge at command. The following, from
Garnett, may not be out of place :

" We may here suggest that it would be a matter of curious speculation
to trace the Indo-European words commencing with r, or its combinations,
to their equivalents in the Tartarian dialects, supposing any to exist. It is
clear that, if they are to be found, it must be under some other forms, and
the identification of those forms could not fail to clear up points in philology
which are at present involved in obscurity." Phil. Essays, p. 258.

This able writer held that " an accurate knowledge of the
permutations of sound in cognate languages is the very foundation
of all rational etymology."* I trust that the few rays of light
which I have attempted to throw on the subject, by a careful
explanation of the phonology of this barbaric dialect, will in-
troduce the Sechwana to the attention of those really able to
make use of its stores of new material.

* Phil. Essays, p. 179

CHAPTER Y.

COMBINATIONS OF SIMPLE CONSONANTS.

To the eye, the number of combinations of simple consonants
may well appear infinite, because they are arbitrary ; but to the
ear, the laws of articulation will very materially prescribe their
limits. I have in this chapter nothing to do with any com-
binations of consonants with vowels, a subject which will occupy
the third and most difficult part of this work. This will alone
materially limit the number for present consideration. Before
proceeding to treat of such as really occur in the Sechwana
language, I shall do what is practicable to every closet student
viz., to try the extent of the various combinations of which the
simple consonants of my classification are susceptible, in order to
arrive at the principles which regulate their use. Those of the
mutes (k, t, p ; g, d, b) are as follows :

kg Jed kb gk dk bk

tg td tb gt dt bt

pg pel pb gp dp bp

gd gb db dg bg bd

kt kp tp tk pk pt

In all, thirty** But it is usually held that, in attempting to
articulate any one of them, it will be found that though it may
be possible to make such combinations to the eye to the ear
the first " must assume the property (quantity) of the second."
That is, with regard to these mutes, a principle is observed that

* Not including the double form of each consonant, making six more.

168

a fortis 'can only stand in opposition to a fortis, arid a lenis to a
lenis,* or the corresponding aspirated forms of the respective
instances, bj which limitation all those represented in italics are
usually excluded, and the remaining twelve combinations to be
regarded as proper phonetic compounds.

Dr. Latham's " Law of Accommodation"

It is to this principle in the combination of the mutes that
Dr. Latham has applied the term " law of accommodation," and
others " law of homogeneousness." Even supposing it to be
correct and constant in respect to the mutes, that learned writer,
however, carries it to a far greater extent than I believe is
warranted by legitimate proofs. Excluding his liquids and semi-
vowels, the following is his system of consonants, all of which he
calls mutes.'^

Lene. Aspirate.

sharp. flat. sharp. flat.

p b f v

k g xfa*?) y(q?)

sz a (sh) (zh)

Now, of these he adds :

" Certain combinations of articulate sounds are incapable of being pro-
nounced ..... Two (or more) mutes, of different degrees of sharpness
and flatness, are incapable of coming together in the same syllable. For
instance, b, v, d, g, z, &c., being flat, and p,f, t, k,s, &c., being sharp, such
combinations as abt, avt, apd, afd, agt, akd, atz, ads, &c., are unpro-
nounceable. Spelt, indeed, they may be ; but attempts at pronunciation
end in a change of the combination " Eng. Lang., vol. i., p. Ixiii.

So far as regards the first three pairs, which I have separated
from the rest by a horizontal line, it would perhaps not be
possible to contradict Dr. Latham ; and were he to use the word

* This would only be a more correct and concise rule than that in vogue
among Greek grammarians viz., " a tennis can stand only before a tennis,
an axpirala before an aspirata, and a media before a media.''

f Eng. Lang., vol. i., p Ixiii. Vide my note in reference to this term.

; The explanatory letters in brackets I have inserted.

169

"mutes "in the proper and restricted sense, it might be proper
to endorse his remark further on " It is only with the mutes
that there is an impossibility of pronouncing the heterogeneous
combinations above mentioned." That the law is in force in
Greek, in such examples as y^ouTroe ; in Latin, in such as
scriptum, rectum (stems scrib-, reg-} ; and in English in stept,
pluckt, &c., &c., there is no denying. But in English we have
the combinations Jed in backdoor, pd in lapdog, kb in blackbird, tb
in nutbrown, dk in woodcock, which are all pronounced exactly
as written, whether we accent the first or second syllable, and
show that such combinations are not phonically impossible ;*
therefore I rather think that an absolute rule does not exist, and
that the " law of accommodation" is guided by the particular
habit of any language. It is remarkable that, while in Greek
it occurs in the beginning and middle of words, in Latin it only
occurs in the middle, and in English most frequently at the end.
Dr. Latham himself admits that there are no general rules for
determining which of the two letters accommodates itself to the
other. The phenomenon must therefore be accounted for upon
some more satisfactory principle.

But when Dr. Latham extends the law to the continuous
sounds of other writers, it is time to demur. The following is a
case in which he applies it to such :

" Tlie letter s. 111 a very large class of words the letter s is used in
spelling where the real sound is that of the letter z. Words like stags,
balls, peas, &c., are pronounced stays, ballz, peaz. It is very important to
be familiar with this orthographical substitution of s for z.

" The reason for it is as follows :

* The example black-guard, pronounced Haggard, is an exception ; but
we have "black-gum, in which the articulations are distinct, though kindred.
The above examples may be spurned because in compound words ; but
surely they are quite as legitimate as the usual examples of nuthook, uphold,
&c., employed by scholars to illustrate to an Englishman the nature of the
aspirated t* p* &c., occurring in Oriental languages.

170

" The words where it is so sounded are either possessive cases, or plural
nominatives ; as stag's, stags ; slab's, slabs, &c.

" Now, in these words (and in words like them), the sounds of g (in stag),
and of b (in slab), come in immediate contact with the sound of the letter s.

" But the sound of the letter s is sharp, whilst those of g and b are flat, so
that the comhinations gs, bs, are unpronounceable. Hence s is sounded as
z." Eng. Lang., vol. ii., p. 69.

The tenor of this reasoning may be all very logical, but, upon
the principle laid down in the preceding chapter, I am inclined
to question the truth of part of the premises viz., that s is sharp
(or fortis). I would urge that the change of s to z, in the above
examples, may be accounted for upon another ground, i.e., there
is nothing to disprove that s, in the words stags and peas, is a
vocalised form of the letter ; in the word balls, that either s itself,
or the I in apposition, is also vocalised. In fact, a proof of this
very argument forms the context to the above quotation from
his work :

" In the old stages of the English language, a vowel was interposed
between the last letter of the word and the letter s, and when that vowel
was sounded, s was sounded also.

" Hence s is retained, although its sound is the sound of z." Ibid.

I have already attempted to prove that z (in my tableau I) is
only a vocalised form of s. Is it not, then, as likely that the
change of the sound in particular English words is to be
accounted for by the fact of an indistinct vowel sound, or element
of vocalisation, taking place of the old full vowel sound, and
merging in the consonantal articulation, so as to form a vocalised
consonant ? That, in many cases, the permutation of s to z occurs
in combination with the lenes-mutes g, d, b, I do not deny, and
shall rather attempt to account for the fact ; but would allege
that, (1) there are other instances in which such changes take
place without any other consonant in apposition, e.g., disable,
dismal, prison, easy, cosmetic, visible, resemble, misery, reason,
presence, &c., in all which cases the s could be indicated by z
(or s of my tableau); (2) there are also instances in which, after

171

such consonants, the mutation does not take place, e.g., eggshell,
absent, landslip, godsend, abstruse, nutshell. Rosetree, again, is an
example in which a lenis, s, precedes a fortis, t.

There is the more necessity for our arriving at correct views
on this subject, inasmuch as, according to the principles of this
treatise, what is said of s by Dr. Latham ought to apply to what
I have attempted to prove are its analogues viz., q (Oriental ),
r, I, f, th l (z of tableau). In the following table there are, for
instance, forty-two combinations of the simple consonants of my
tableau, in which a mute is the prepositive element :

TABLE I.

kz (tb/) kf

kv (Mex.)

gv ,

tl*

tv ,

dv ,

P v .

bv ,

Now, according to Dr. Latham, all those combinations in
italics are unpronounceable, because, in his system, the conjunct
elements are of different quantities. But he holds that gr, gl ;
dr, dl ; br, bl, are pronounceable, because both elements are flat
(lenes)\ and here he has the testimony of Professor Max Miiller,
who says of I and r, " they are soft, like the mediae, owing to the
process of their formation." However, upon the basis of new
materials in Sechwana phonology, I have already proved, satis-
factorily I trust, that the mutes g, d, b, and the liquids I and r
are similar in quantity. I have also made an effort to show that
q (Oriental) is an analogue of the elementary forms of both r and
I. Moreover, by analogy, I have attempted to prove that both
s and th l (z of tableau), as well as the labial/, are, in their tenues
or elementary and unmodified forms (not usually admitted into

* These are in bold letters, as particular reference will be made to them
in the sequel.

172

classification), analogues of both r, I, and q ("Oriental) ; and, con-
sequently, that the combinations of all these with the mutes
g, d, b, are also pronounceable. As to s, the English examples

bigseat gladsome absent
not not not

bigzeat gladzome dbzent
nor nor nor

bikseat glatsome apsent

will alone show that the combinations gs, ds, bs, are quite as
pronounceable as gl, dl, bl. If the combinations of these mutes
with the other liquids, q, th l , and /, existed in English or in the
Sechwana, it would no doubt be an easy matter to furnish similar
examples.

But in maintaining that coincidence in quantity is necessary
to the proper pronunciation of conjunct articulations, has Dr.
Latham not utterly forgotten that in such combinations as kl, kr ;
tl, tr; pi, pr, which abound in the English language, in such
words as cloth, crown ; little, tree ; place, pride, the elements, so
far as regards quantity, are heterogeneous, as well as in such
combinations as ts and if, which he also sanctions ?

In the following table, again, are forty-two combinations, in
which the liquids are pre-positive :

TABLE II.

sk (th')

fk (Mex.)

Here, though Dr. Latham has above made such an oversight
as that noticed in instances where the mutes are prepositive, he
admits that the liquids I and r, though flat (lenes), may be followed
by a sharp consonant, as alp, alt. The same is allowed on the

173

principles of this treatise ; but still more viz., that, inasmuch as
I have proved q, s, z (th 1 ), and /, to be analogues of r and I, and
therefore all flat (or lenes), their combinations, with a sharp
consonant, whether this be post or pre-positive, are all pro-
nounceable.

This treatise not only holds that the liquids of its classification,
although lenes, can be pronounced in apposition to the lenes-
mutes (which Dr. Latham denies in considering all, except I and
r, as sharp), but it also maintains that they are distinctly pro-
nounceable in connection with fortes-mutes ; in fact, that the
liquids r, I, s, z (th 1 ), q, and/, are pronounceable with any simple
mutes (unmodified by either' the spiritus or element of vocalisation),
whether post or pre-positive. If this can be disproved, the con-
clusions at which 1 have arrived respecting the nature of these
liquid elements are invalidated. If not, any conventional notions
regarding our habitual use of only some of their combinations
must fall to tne ground.

Now, it must be borne in mind that, in the above groups of
combinations, I have only had to do with the tenues forms 'of all
consonants, i.e., their absolutely simple or elementary forms,
unmodified by either the spiritus or vocalisation ; and, moreover,
that among these consonants, those of s, z (th 1 ), /, and q (Oriental),
as tenues forms, are not usually admitted into classifications.
Men are in the habit of looking at the aspirated forms of these
latter consonants viz., the common s, th, f, and the Germ, ch
(gutt.), having only their sonant modifications z (common),
th 3 (this), v (common), and the Arabic (Northumberland burr").
It is possible that, by now considering all the combinations of
these tenues forms, as affected respectively by the spiritus and
element of vocalisation, we may be enabled to approximate to
correct conclusions, and a satisfactory analysis of the difficulty.

174

Consonantal Diphthongs.

The remark of Dr. Lepsius, concerning " a complete and accu-
rate theory of transcription" in respect to the vowels, is equally
applicable to the consonants, inasmuch as it is necessary to make
a distinction between what are and are not diphthongs among their
compound forms. Even the subject of vowel diphthongs is in-
volved in such confusion, that I need not wonder at finding
myself in a labyrinth in the case of the consonants. The fol-
lowing quotation will display the unscientific and random style
in which these are usually disposed of.

(Here ends the Manuscript.)

" The diphthongal consonants are r, w, y, j, q, oh, wh." Author of Article
" Stammer," Penny Cyclopaedia.

-HI II

<AU!BRARY0/

s s

o -e^ ^ ^

v/nBAiNiHtf*

=5 o

& 1

> ^OF-CAIIFO% ^\\E-UNIVER%.

B =
ff %

* i
i |

I 5

1 a

s 8

s 3

5 8

! I

% ^

g i

2 ?

s B

S a

5 1

< 5
s <=>

S >

~~ *-

1 I

i S^l Iffic! IJUff

}''

i n

A 000020383 6

<*HIBRARY0

S I

i l

3 V

^\\EUNIVER% ^

.95 . G* & &

^OJIWOJO^
^OF-CAIIFO%

*"^l

*.A\H!

I fc

vvlOS-ANCflfXA ^JE-ll

"^ ^3 -f

^\\E-UNIVER%
^ *, -^"g

= I
^