THE LIBRARY

OF

THE UNIVERSITY
OF CALIFORNIA

LOS ANGELES

L -,*

LIVING SPEECH

IN CENTRAL AND SOUTH AFRICA

BY

A. C. MADAN, M.A.

STUDENT OF CHRIST CHURCH
OXFORD

OXFORD

AT THE CLARENDON PRESS
1911

HENRY FROWDE, M.A.

PUBLISHER TO THE UNIVERSITY OF OXFORD

LONDON, EDINBURGH, NEW YORK

TORONTO AND MELBOURNE

PL

PREFACE

THIS Essay, it is hoped, may serve to interest the general
reader in a subject of some importance in its bearing on the
future of Central and South Africa, to invite the attention of
students of Language to a form of it which seems to deserve
more study than it at present receives, and to appeal to all
acquainted with Bantu in any one of its many dialects to
compare the statements here made with the facts known to
themselves. In the end it may be that justice will be done
to a subject, which if only from a geographical point of view
has some claim to be called a great one.

This threefold purpose has involved some difficulty of treat-
ment, and will lead to criticism in details. For instance, Bantu
words have been only sparingly introduced for purposes of
illustration. The first class of readers would probably pass
them over, the second might well expect more, the third will
have a stock of their own to refer to. But the consequence is
that much here stated will seem to have little to support it.
The fact is, that the subject of each chapter, even of many
single paragraphs, requires a whole treatise for its separate eluci-
dation, and a single Essay could in no case adequately represent
the total bulk of literature on Bantu subjects, which is large,
scattered, and often difficult to find.

Here it will be enough to say briefly that the dialects best
known, though at best imperfectly known, to the writer are
the Swahili of Zanzibar and sundry dialects of the East Coast,
Nyasaland, and Northern Rhodesia, and that the books at hand
for further information and reference have been but few. The
following may be mentioned among those most accessible to
English readers Sir H. H. Johnston's article on the Bantu
Language in the Encyclop&dia Britannica, and chapters on
Languages in his books on various parts of Tropical Africa,
especially the latest, George Grenfell and the Congo ; Father

A 2

1105413

4 PREFACE

Torrend's Comparative Grammar of South African Langtiages,
written nearly twenty years ago, but still valuable as a com-
pendium of facts then known, ably surveyed ; Professor Tucker's
Introduction to the Natural History of Language, and among
the many volumes dealing with particular dialects, Stapleton's
Comparative Handbook of Congo Languages, and Scott's
Cyclopedic Dictionary of Mang'anja. The facts in this last
volume, and the striking suggestions interspersed throughout it,
have been freely utilized in this Essay. It is a store-house,
which no student of Bantu can afford to neglect.

But apart from the exact knowledge of the details of dialect,
one or many, constant familiar contact with Bantu speech in
one form or another for nearly thirty years constitutes a basis
which gives to general impressions and conclusions a force
greater to the writer than his actual command of facts, or the
meagre selection of them here presented to the reader, would
convey. And the object has been to write an Introduction to
Bantu, not to its formal Grammars or Dictionaries, but to
a great block of Human Speech, viewed as a whole, under
necessary limitations as to the point of view.

For convenience, the Bantu Family of Languages as a whole is
often referred to as Bantu, and the people speaking any dialect
of it as Bantus.

A.C. MADAN.

FORT JAMESON,

NORTH-EAST RHODESIA.

CONTENTS

PART I
INTRODUCTORY

CHAPTER PAGE

I. PRESENT STATE OF KNOWLEDGE OF BANTU . 7

II. BANTU AS A SUBJECT OF STUDY ... 12

III. BANTU IN RELATION TO GENERAL PHILOLOGY . 17

PART II

WORD-BIRTH

IV. GERM-SOUNDS, OR MONOSONANTS . . ... 21

V. MONOSONANTS, INTERACTIONAL AND ADVERBIAL 26

VI. ROOTS ........ 35

VII. VERBS AND NOUNS 38

PART III

WORD-GROWTH

VIII. VERBS ........ 46

IX. VERB-SUFFIXES 51

X. VERB-PREFIXES . 59

XL NOUNS 64

XII. NOUN-PREFIXES 68

CONTENTS

CHAPTER PAGE

XIII. ADJECTIVES .... . . 78

XIV. PRONOUNS .81

XV. ADVERBS, PREPOSITIONS, CONJUNCTIONS . . 86

XVI. CONCLUSION 91

PART I. INTRODUCTORY

CHAPTER I

PRESENT STATE OF KNOWLEDGE OF BANTU

AFRICA has produced many a new thing in the world's
history, but never more than in the last half-century. Yet the
remarkable Family of human speech which is most character-
istic of it as a continent, that which at present occupies almost
exclusively its central and southern parts, can hardly be said to
have received the attention which, if only as a novelty, it might
be expected to command. Even now, when the Native Question
in the whole of this region takes its place among the foremost in
the social and political horizon not only in view of the recent
union of South African States under the British Flag, but of
the interests of the European nations who find vast tracts of
the continent both east and west under their control the im-
portance of the Bantu language seems but little recognized.

Partly, no doubt, the fact is due to the position of Philological
science in general. It appears to be hardly regarded, at any
rate in England, as a science seriously entitled to the name.
For periodicals garnering the products of pioneer work and
isolated studies of portions of the vast agglomerate of Bantu
dialects; for books dealing with such monographs, digesting and
arranging their results, and bringing them into relation with
other branches of the world-wide science of Human Speech ; for
Professors engaged in research, or in teaching the results of
work done in all parts of the world, for all these it is to Italy
and France and above all Germany, that attention must be
mainly turned. Even when the British Association met a few
years ago in South Africa, and held sessions so far north as
the Victoria Falls, and when every one connected with it must
perforce have come in contact with the Bantu in speech and

8 INTRODUCTORY PT. I

person, and some must have been struck with the portent of
' Kitchen-Kaffir ', which does duty in many parts as linguistic
medium between black and white, even then it appears that no
recognition was thought worth according to the great linguistic
ocean in which they were plunged. A few more or less ethno-
logical aspects of the native population were allowed consider-
ation, but of studies of Bantu speech, as such, in general or
particular, not one.

Yet it cannot be said that the matter is an unimportant one.
Unity of speech, or at least the possibility of mutual under-
standing and easy inter-communication, is one of the most
potent factors making for or against the political unification
and social progress of the black race. Again, few indications are
more suggestive than language, of the actual and latent powers
of the race whose nature it reflects ; the actual powers oi
thought, feeling, and expression ; and still more the potentiality
of development in these and other directions. It would prob-
ably surprise some statesmen as well as scientists, if it should
appear that the Bantu Family of speech is by no means
a promiscuous agglomerate of barbarous dialects, multiplying
with the wild and wayward luxuriance of tropical growth, and
diverging rapidly into mutual unintelligibility or even de-
graded caricatures of the utterance of civilized beings. On the
contrary, there is much evidence to show that a striking unity
both of grammar and vocabulary underlies the speech of the
whole vast region, a unity which under favourable conditions of
general peace and increasing inter-communication might rapidly
be reinforced, re-emerging (as it were) where almost lost in
backwaters and under-currents, reasserting itself when obscured
by strange phonetic disguises due often to influences only
transient or local, a unity which, under the stimulus of
education, social aspiration, and the growth of a native press,
might in no distant future give large sections of the Bantu race
a racial self-consciousness and power of united action, against
which its present rulers might do well to be forewarned, if not
fore-armed. Much more so, if it should further appear, that
Bantu speech is such as in fact indicates considerable possibi-
lities in the people, of whose mental characteristics it is the

CH. I PRESENT KNOWLEDGE OF BANTU 9

reflex and to which it is the key, possibilities which, with due
allowance of time and opportunity for development on the one
hancl, and allowance for the total absence hitherto of writing
(alphabet, books, records of any kind) on the other hand, may
yet win a place in the world's regard. Its richness, flexibility,
expressiveness, and present living power of forming at will and
selecting suitable forms for the embodiment of new general
ideas and new objects, as well as of varieties and refinements of
those already existing, will be attested by all who are really
well acquainted with almost any single dialect of Bantu. That
its powers are so little exerted as to bring their very existence
into doubt, is hardly more remarkable than that Demosthenes
did not deliver the De Corona from his cradle.

To scientists, Bantu speech may be expected to bring novel
and useful contributions towards the solution of some of the
most important as well as the most remote and difficult problems
of human language, its origin, lines, and stages of develop-
ment, modes of differentiating roots, meaning of the forms of
declension, conjugation, and of some of the Demonstrative
(pronominal) parts of speech. And lastly, these contributions
can hardly fail to modify and extend knowledge of human
nature itself, ever striving to express its own limitless aspirations
within the limits of human capacity for producing sound.

It must be granted that at present, and probably for many
years to come, a complete and exact survey of Bantu speech
cannot be looked for, and general conclusions about it must be
in some degree hypothetical. Geographically, its habitat is
large. Practically, it is made very much larger by the relative
inaccessibility of large portions of it. The number of dialects
(whatever definition of dialect may be adopted) and even of
groups of dialects is also large. Few individuals make them
a subject of serious and careful study. Officials have many
other things to attend to. Settlers have to make their living.
Travellers seldom can go below the surface. Most of the work
is done by missionaries, simply because their immediate objects
make it indispensable to understand and be understood. More-
over, few students of any class start with any special equipment
of philological knowledge or training for such work, or even

io INTRODUCTORY PT. I

with any generally accepted guide to correct representation in
writing of the sounds, which are the sole material to be dealt
with. Hence even such results as are obtained are often diffi-
cult to understand and compare. Nevertheless a large amount
of work has been done, and its volume rapidly increases.
Any bibliography of works dealing with or written in Bantu
runs into hundreds of items, but completeness is hard to obtain,
because for the most part the works themselves are produced
by local presses or by special societies, with a practical and
limited aim and no close relations with the world of general
literature. The majority of such works are small Grammars
and Vocabularies in handbook form, collections of native
stories, accounts taken down (sometimes by phonograph) from
natives of native customs and history, and attempts at trans-
lation into a Bantu dialect of educational and religious books.

On such material, collected and compared, are founded
a few works of larger scope, (in English very few,) Bleek's
pioneer Grammar of half a century ago, Kolbe's Language-
Study of Bantu, and Torrend's Comparative Grammar of
Sotith African Languages (1891), Stapleton's more recent Com-
parative Handbook of Congo Languages. As an approximate
survey of the whole field, nothing seems at once so comprehen-
sive and so worthy of attention as Sir H. H. Johnston's Essay
in the British Encyclopxdia, and another in the latest of his
great series of volumes dealing with different regions of Central
Africa, George Grenfell and the Congo, though in this last work
his immediate results relate mainly to tracing the source and
movements of the great tide of Bantu speech, and defining in
a map its position and boundaries. It might have been sup-
posed that the mere growth of official contact between England
and Africa, the responsibilities of administration, and the
pressing need for mutual understanding between white and
black, in the East Africa and Nyasaland Protectorates, as well
as in the South and Rhodesia, would have forced the subject
to the front, as a matter of policy and obvious precaution.
But the facilities for learning in England any of the most im-
portant dialects appear to be extremely small ; the Universities,
old and new, make no provision for its study or teaching, and

CH. I PRESENT KNOWLEDGE OF BANTU n

the limited attention paid to Bantu in so recent and important
a book as Professor Tucker's Introduction to the Natural
History of Language shows how little the whole subject is
before the learned world.

Perhaps the most useful contribution that can be at present
offered to the subject as a whole would be to assume the two
works last mentioned (Sir H. H. Johnston's and Professor
Tucker's), as for the time defining it from the point of view of
history and geography on the one hand and of scientific
philology on the other, and then to confirm or modify them
by a view based on dialects forming a part, though only a
part, of the whole mass of Bantu.

CHAPTER II

BANTU AS A SUBJECT OF STUDY

IF somewhat scanty attention has so far been given in
England to the Bantu Family of Language, it will be hardly
denied that as a phenomenon, gradually revealed in the last
forty or fifty years, it is sufficiently striking.

To take first its external features only : as a whole it is new,
it is on a vast scale, by its isolation and its solidarity it is
especially fitted as a subject for scientific examination, and
it is the last of the kind, combining all these features, which
the world can provide for the solution of the important
problems of philological study. Beside all this, there is its
intrinsic character and individuality as language.

Gradually the results of the labours of hundreds of workers,
beginning with the early days of Portuguese discovery and
enterprise, increasing slowly till the last century, and then
with rapidity towards its close, have been accumulated and
compared. They came from independent sources and regions
far apart the Cape, the Congo, the Cameroons, Zanzibar,
Uganda, Nyasaland first from the coast regions, then gradually
from points here and there in the huge intervening spaces. At
last it became clear that all Africa from a line somewhat north
of the Equator to the Cape (with the exception of scattered
groups of Pygmies and Bushmen, of Malays and Hottentots
of the south, and Arabs) was in a real sense of one speech.
The northward boundary just referred to was drawn with
some precision and finality in 1908. [See the map in George
Grenfell and the Congo, by Sir H. H. Johnston.] The bound-
ary is doubly guarded. The broad belt of sandy desert, which
severs the north coast of Africa from the rest of the continent
and is prolonged into Arabia, pierced only by the Nile valley
on the east and various caravan routes in other parts, is itself
a barrier against pressure from the races and languages of the

STUDY OF BANTU 13

outer world. South of the desert comes indeed a broad fringe
of tribes of diverse origin and diverse speech, but all alike in so
far as the contrast is in the main clear between them and the
Bantus of the south, and valuable not only as another barrier,
but as a measure showing how little they are mutually affected
by contact except actually on the border line.

Southward, the coastline of the continent is the only boun-
dary of Bantu, and this also carries important consequences.
It is of course a coastline of singular uniformity, unbroken
by inland seas, gulfs, or even (with few exceptions) by good
harbours, while the rivers are almost all made difficult by
shallows and sand-bars at their outlet, and in no case allow of
navigation by large vessels for any great distance from the sea.
Thus another great natural barrier has preserved Bantu speech
from external influences which might affect its purity and the
course of its natural development. Full allowance may be
made for the infiltration of some small element of foreign
words from the coast, whether Phoenician, Persian, Indian, or
Portuguese. But the only alien elements in the Bantu area
itself, and those of no great importance, are those already
referred to the Malays of Cape Colony, who are not numerous,
the Hottentots, who are rapidly losing all separateness as a
race, the scattered and scanty hordes of Bushmen, marsh-
dwellers, and forest Pygmies, and the Arabs, who as settlers
and slave-traders have for centuries penetrated more deeply than
any race into Central Africa, and whose marked influence on
speech will be referred to again in connexion with SwahilL

On the whole there can be no part of the world where so
vast a volume of human speech has been longer or more
effectively safeguarded by natural barriers from any large
measure of external interference than the area occupied by
Bantu.

But it is not only its scale, its comparative isolation, and
its presumable freedom from adulteration, that gives the
phenomenon exceptional interest as a subject for scientific
study. However remote in time its origin may be, and there
are certainly indications of its being of a primitive type, there
is some ground for believing that its chief epoch of expansion

14 INTRODUCTORY PT. I

in Central and South Africa is comparatively recent. In
British Central Africa (1898, p. 480) Sir H. H. Johnston
gives reasons and promises many others for fixing that epoch
as beginning about 2,000 years ago. Taken as a working
hypothesis, this may mean that (in addition to the fact of its
present geographical position and distribution in Africa) the
spread, movements, and developments of Bantu in that area
have taken place within fairly definite chronological limits, and
those relatively small. Hence a new interest and possible
value of Bantu, as throwing light on the kind and amount
of dialectic variation and phonetic change taking place in
a comparatively short time under conditions favourable to
natural life and growth in speech. Philological grounds are
also given by Sir H. H. Johnston (who has had unequalled
opportunities for forming an opinion, both from his official
career in many parts of the Bantu area, and from his command
of the assistance of others) for fixing approximately several
important points. The chief are the starting-point from which
Bantu speech proceeded to extend itself, viz. the region of
Mount Elgon in Northern Uganda, and the lines of successive
extensions, each now showing different characteristics, to the
westward, to the south-west, and to the centre and south-east
(see George Grenfell and the Congo]. Further, it appears that
the two earlier advances, including the dialects of nearly the
whole Congo region and the German South-west African
province, show phonetic differences in the direction of decay
and obscuration of the original type, more marked in propor-
tion to their distance in time and space from the starting-point.
The same may be said generally of leading dialects of the east
and south-eastern regions, Yao, Makua, Tonga (of Delagoa
Bay), and Zulu, as compared with the comparatively pure
and unchanged current of a primitive type found along the line,
and on the east, of the great Central Lakes from Uganda to
the Zambesi. Hence it appears probable that this central
chain of dialects, including those of Lake Nyasa and parts of
Northern Rhodesia, provides a good field for the study of the
purer type of Bantu. In one or other of them, both the main
elements, radical and formative, appear to exist in forms more

CH. II STUDY OF BANTU 15

favourable than elsewhere for comparison with others of the
east, south, and western regions, and for provisional use as a
standard by which to recognize and estimate the various
linguistic processes illustrated by them, whether of decay,
regeneration, or analytical development analogous to that
which is widely recognized in other types of speech.

The above sketch is, it will be remembered, of the nature of a
reasonable working hypothesis, subject to criticism and correc-
tion, but useful meanwhile as at least a theory for verification as
time goes on. Final conclusions must depend on the complete-
ness and accuracy of a comparative survey of the whole vast
field. The difficulty of such a survey was due at first to lack of
material. Now the mere volume of accessible facts to be sifted
and mastered makes the difficulty even greater. Meanwhile,
the hope of advance appears to lie in each worker in the field
contributing his own quota of fact and theory to the common
stock, defining his own limitations of knowledge, and waiting
for similar bodies of evidence to accumulate, till local and
limited views are placed in their right light, and truth emerges
from a just estimate of them all.

It is needless here to enter into the difficulties of merely
stating facts of language. The facts are sounds, and as sounds
can only conventionally be represented at all on paper. At
every step the first-hand student is baffled. The same word
will sound differently, as pronounced by the lips of different
speakers. It may also sound differently to the ears of different
hearers. The sound as heard, or supposed to be heard, may
be symbolized by many different ways of writing it. Never-
theless, the phonograph can never supersede writing as the
vehicle of common communication of knowledge, and it re-
mains to acquiesce from the beginning in a merely approximate
representation of the facts at issue. Symbols must be used
for sounds, and used under the ordinary limitations of type and
space, without even the help of some kind of musical notation
to relieve the dismal contrast between a dead level of black
letters and the infinite modulations of any human utterance.

Again, it is impossible for a student of language to adduce
all the facts on which his conclusions rest. All that could be

16 INTRODUCTORY

written down would form but a small proportion of the whole
which go to form a theory or (still more) to train the judgement.
The majority pass through the mind, or at least the ear, and
are forgotten perhaps beyond recovery, but none the less they
leave their impress, strengthen or weaken (perhaps uncon-
sciously) previous impressions, and give edge to the tool in
dealing with new facts that follow. Consequently, in such
a subject as Language, and except for the purposes of a strictly
scientific treatise,. a selection only of facts can be presented to
a reader, examples, in short, quite disproportionately few as
compared with arguments and conclusions resting on them, but
valuable in proportion as they are really examples and not
exceptions, types and not oddities of speech, in proportion (in
fact) to the care and insight shown in their selection.

As to the chapters which follow, thirty years of contact
with Bantu speech, beginning with the Swahili of the East
Coast, and ending with various dialects of Nyasaland and
Northern Rhodesia, guarantee a certain familiarity with some
details of the great whole, but are hardly compatible with the
wide study in books exhibited by others, which might put
such details in a truer light. Nevertheless, if advance is to be
won, it seems necessary that partial and imperfect contributions
should be made and weighed. The present object is not to
introduce Bantu in connexion with any special dialect, or even
to outline with any formal completeness its characteristic
grammatical system, but rather to give a view of it as a distinct
type of human speech and a possible key to some of its deepest
problems, under the limitations above described.

CHAPTER III

BANTU IN RELATION TO GENERAL PHILOLOGY

FROM the nature of the case the study of Bantu, in
proportion as it becomes more exact and more complete, may
be expected to throw more light and new light on many of the
deeper problems of Philology, more especially on the processes
of language-making in all its stages.

The translation of thought and impression into sound, the
interaction of mind and matter, the blending of spirit and
mechanism incarnation in fact, seems in Bantu to be going on
before the eye. Free in a remarkable degree (as has been shown)
from interference of non-Bantu influences or intrusion of alien
modes of expression, free also from the fetters of writing and
the exacting demands of literary style, even free for the most
part from the conditions imposed by continued residence of
large communities in the same spot under settled conditions,
Bantu speech has every chance of reflecting life, the natural
movements, the instinctive efforts of the human spirit, exercised
if not developed in an environment which encourages in every
way the free interaction of nature and humanity, the mutual
influences of man and the world he lives in. The only sounds
supplied from without are those of nature winds, woods,
and waters, insects, animals, and birds. Besides social inter-
course and the operations of domestic life, war, hunting, and
agriculture are the chief experiences acting on the mind and
challenging expression in sound, i. e. in words. Each tribe or
group of tribes, each clan, even each village is often free to
adopt, reject, change its words to suit its own needs of mutual
understanding, without reference to any other consideration
whatever. If self-preservation did not in some regions step
in to force tribal organization into prominence, (with all its
necessary consequences of dwelling together in large villages,
and surrender of individual rights in face of the constant menace
of impending war and possible extermination by powerful
neighbours,) rapid and complete divergence in speech would

1314 B

i8 INTRODUCTORY PT. I

naturally follow even more largely than is the actual case, (see
Stapleton, Handbook of Congo Languages, Part ii. A. Dialecti-
cal Growth). Where the need of organized and united self-
protection is reduced to a minimum by the adoption of some
natural barrier against a hostile world the deserts by the
Bushmen, the swamps and marshes by the Watwa, the forests
by the Pygmies it is not to be wondered at if language within
the barrier serves its purpose there, but is unintelligible beyond.
Hence Philology might be expected to turn to Bantu as
a priori likely to exhibit human speech in forms and under condi-
tions which would throw light on its mode of development and
even its origin, as well as on the uniformity of the workings of
the human spirit in its endeavours after self-expression. Such
expectation will certainly not be wholly disappointed. It may
safely be said that there are few questions of interest in the
whole range of philological science on which Bantu has not
a bearing : whether remote or unimportant, may be left to those
who are most competent to decide. Some of the positive
features of Bantu may be shortly summarized. The sound-
system, based on the syllable consisting of a consonant, followed
usually by a vowel, is easy, flowing, even melodious, and would
be more so if not for the prevalence of the ' hard ' consonants
and the vowel A. In this respect, Swahili has been compared
to Italian. It is also singularly free from difficult sounds or
combinations, and such as exist are largely local and traceable
to outside (non-Bantu) influences, e.g. the clicks of Zulu to
Hottentots or Bushmen, the gutturals sometimes heard inSwahili
to direct imitation of the Arab. The richness and fertility of
sound-combination, actual and potential, directly fitted for
effective expression, implying as they do a language capable
of keeping pace with and supplying the needs of minds
growing in capacity for differentiation of their own ideas and
the assimilation of new ones, have impressed every student
of a Bantu dialect in proportion to his acquaintance with it.
The process of word-building, or rather of word-growth, is
singularly transparent. The raw materials of speech are
largely recognizable, the roots to which all forms point and to
which many can be traced, the interjections which form by

CH. Ill BANTU IN RELATION TO PHILOLOGY 19

themselves a notable and large element in speech and are the
base of a much larger. The prefixes, infixes, and suffixes one
or more of which are associated with every word are on the
whole clearly defined in form and to some extent in meaning.
At the same time, most of the grammatical terms used in
relation to European and other languages can without strain
be applied to Bantu speech, and receive a measure of illustra-
tion and even explanation from it, the form and meaning of
declension, the differentiation of stems, the systems of tense-
forms, moods, conjugations, even syntax, these and others are
points in which study may expect to be repaid. An attempt
will be made to indicate some of the directions in which results
more or less important may be looked for and without difficulty
obtained. Such results are of course largely confirmatory of
previous philological conclusions, but suggest new grounds for
some as well as modifications in others.

A few words are necessary here as to Swahili, the East
Coast dialect of Bantu. It still serves admirably the purposes
of a lingua franca in Central Africa, surviving the breakdown
of the old commercial system based on the slave-trade, which
was the chief source of its vitality in the interior, and made
it widely and commonly known. Faithfully reflecting the
mixture of African and Arab elements in the coast-population,
carried by their representatives along every trade-route of the
many which up to the present generation converged from all
parts of the region of the great Central Lakes and the Congo
basin on Kilwa, Mombasa, and Zanzibar, readily adopting,
assimilating, and making familiar Arabic words for all new
ideas and many old ones, never losing but in some ways
materially simplifying its own essentially Bantu grammar, the
loss of which would have sacrificed its power as a medium
between the African and the outer world, Swahili is for these
reasons excellent as an introduction to Bantu, though not
the best for scientific study of it. The requirements of the
foreign (almost entirely Arab) element not only introduced
many new words and even for the commonest things displaced
many old ones, but the Bantu element on the Coast was itself
not a pure one. Its vocabulary was largely influenced by the

B a

20 INTRODUCTORY

drifting of thousands of slaves from every tribe within reach
of the trade into its always mixed population. The slaves
brought numbers of words and forms together, which never
coexisted in any one of the dialects of the interior, and which,
though true Bantu, do not reflect faithfully any particular
locality or stage of development in Bantu speech. Moreover
(as has been said), the grammar, and especially the tense-
system, had to be simplified for the use of the Coast, and while
achieving a real gain in definiteness of meaning, lost much of
the richness, vividness, and variety, which is at once the beauty
and the difficulty of many a dialect in its own home in the
heart of Africa.

Swahili, then, is of great practical use and importance, not
only as largely the official language of the British East Africa
Protectorates, the German province of East Africa, and the
Belgian Congo, but also as an easy and adequate example of
true Bantu grammar, and a key to it wherever used. But for
the purest examples of natural and unadulterated developments
of the Bantu word and sentence, the truest reflection of the
Bantu mind, feelings, and capacities in its speech, attention
would be better turned to a dialect in the region of the Great
Lakes Victoria Nyanza, Tanganyika, and Nyasa. Singularly
valuable in this connexion is the Cyclopaedic Dictionary of the
Manganja Language by the late Rev. David Clement Scott.
What is of value in the following chapters on Bantu will be
largely due to the facts so accurately and fully recorded in
that book, and still more to its singular and inspiring suggestive-
ness. No other book on a Bantu dialect seems to approach it
as an extraordinarily able and truthful collection of the facts
of a single dialect, illuminated by the sympathetic insight of
a sensitive and philosophic mind, enthusiastically appreciative
of the capacity and promise of native thought. That it is
not a mere Dictionary has been a stumbling-block in the way
of its use by some, but the only subject of real regret is that
the author did not live or have the leisure to connect and follow
up the scattered hints and suggestions with which it abounds,
and so give its true completeness to a work of such unique
interest and value.

PART II. WORD-BIRTH

CHAPTER IV

GERM-SOUNDS, OR MONOSONANTS

LIFE life in sound, living speech, speech which is not to be
found in books or is half dead if put there, which refuses to be
photographed in type, and even if phonographed fails to make
a trumpet really speak, speech which postulates life, the living
speaker, the living hearer, and the life around them, as its very
atmosphere and condition of its intelligibility, indeed of its
very existence as language this if not the first, is the last and
truest impression given by Bantu. It is certainly the dominant
idea under which to attempt its interpretation. Life rather
than mechanism, word-birth and word-growth rather than
word-formation or word-building, and organic living rather than
mere constructive processes, are the expressions suggested by
the survey of even written specimens of Bantu speech. Not only
roots and stems, words familiar enough in all works on language,
but germs and seeds, branches, flowers, and fruit are terms
almost demanding use as fittest for true descriptive purpose.
From the first a written letter or symbol is felt not only as
a partial and inadequate representation of the truth of a sound,
but as an actual restraint on the power to grasp and realize its
nature. It challenges and necessitates criticism, as implying
greater fixity than exists, and a machine at work rather than
living spirit playing on an instrument. Hence in Bantu, more
perhaps than in other languages, endless and often fruitless
arguments are apt to rise and flourish as to the nature of cer-
tain sounds in themselves often more definite than definable,
the changes, interchanges, substitutions, modifications, which
they admit of, in a dialect or as between different dialects ,
arguments tending as usual to become more dogmatic in pro-
portion to the impossibility of demonstrative proof.

22 WORD-BIRTH PT. II

Symbols, however, have to be selected and used. And it is
fortunate that the letters of the English alphabet are on the
whole singularly sufficient as a guide to practical appreciation
of Bantu speech, however inadequate to reflect its actual pro-
nunciation and accent. In fact, several English consonants can
be dispensed with, and at most a few rather unusual combina-
tions of them required, though every student in actual contact
with Bantu will probably desiderate one or two special
characters in addition, if only to protect himself from the
charge of making fanciful or even impossible statements as to
phonetic fact. Here it will be assumed that as there is a close
general similarity between the organs of speech in all in-
dividuals of the human race, so in Bantu, at any rate, the sounds
produced by them for the purpose of mutual understanding
can be classed on the principles, and symbolized by the letters,
commonly given in English works dealing with the subject of
phonology. While English vowels will be limited to the more
fixed values * given them on the Continent and especially in
Southern Europe, the consonants are given their common
values, limited to one where more than one is in common use.

If we start, then, with the English alphabet and omit only
C (except in the combination CH\ Q, and X, three superfluous
letters sometimes usefully employed to represent clicks where
these are found, it will be convenient to arrange the letters
in the way likely to be found most suggestive and useful in
dealing with Bantu, as well as in the main most commonly
accepted as a sound physiological basis,

This arrangement may be taken to suggest, so far as Bantu
is concerned, that the vowels A, /, U are (so to speak) the
cardinal points of the vowel-system, and that E, O, if also
original, are also vowels in some degree secondary and inter-
mediate. They often are representative of fusion or coales-
cences of A, 7, and A, U respectively, where such fusion is
admissible. In all questions involving E and O in Bantu, it
is well to bear in mind the possibility of their resolution into

* A ah, E = ay, O = oh, I = ee, U = oo.

CH. iv GERM-SOUNDS 33

a more original form, involving two sound-elements instead
of one.

The connexion of the three cardinal vowels A } /, U, with
three leading groups of consonants is so close and important
as to make it advisable to include them in the following Table.
The various descriptive terms included suggest in some degree
the basis of a classification, which will further justify itself
practically in application to Bantu.

TABLE A.

(W)

(i)

(a)

(3)

Palatal

Dental

Labial

(back)

(middle)

(front)

Vowels

A (E)

/ (0)

U

Semi-vowels

H

Y

W

Nasals

NG'

N

M

Stops, full

K (CH)

T

P

half

G (?)

D

B

Liquids

L t R

Spirants

IF

V

( C

Sibilants \ ~

As this Table is used for reference in much that follows,
a few explanatory remarks are here added by way of
anticipation.

i. Each of the above consonants has its ordinary English
value, and represents a sound common in Bantu, except that

H is not used in some dialects.

G is not used for ? (as in age), nor 5 for Z (as in bees).

M and N have two values, not always easy to distinguish,
(a) as consonants, nasalizing the sound that follows, and pro-
nounced in close connexion with it, (b) as sonants or semi-
vowels, having a distinct syllabic value, as if\[/M, 7N, and
generally representing a distinct formative element in a word.
The distinction is important but not commonly indicated in
type. So also is the close connexion of sonant M with U and
W, and N with / and F, in Bantu phonetics, so close that each
may represent the other, as well as be combined with it.

Moreover, important phonetic (mainly euphonic) changes are
connected with the combination of M or N with other sounds.

24 WORD-BIRTH PT. II

L and R, though often distinguished in sound, are practically
interchangeable, and both are often represented in pronunciation
byD.

W, B t and V are common ways of pronouncing a single
peculiar sound, distinct from them all, and widely used in
Bantu, one of the few (beside local clicks) which are not easy
for a non-Bantu to pronounce. It is in fact a W pronounced
with an approximation to B on the one hand, or to V on the
other. It is here called modified W t written, when necessary,
W. The B in the word Bantu is a convenient illustration,
represented as it is in various dialects by B, W, or F, or some-
times altogether dropped.

a. It must not be supposed that Bantu is limited to syllables
beginning with a single consonant from the above Table.
Subject to some phonetic limitations, the following may be
noted here :

The commonest and most characteristic syllables are a nasal
(M or TV) before, or the corresponding semi-vowel ( W or Y)
after, any of the consonants, or both together, e. g. mp-a, pw-a t
Py-a,) mpw-a, mpy-a.

In some dialects the sibilants (S, Z, also Sff, ZH), and
Spirants (F t V) are often ' hardened ' by prefixing a ' stop '
consonant (T, P, D, B, not K t G), e.g. ts-a y dz-a, pf-a, bv-a.
And these combinations may be nasalized as above.

The use of H is limited to certain dialects, but is common
and of great importance in Swahili. But a faint aspirate or
' check ' is heard in some dialects after the ' stop ' consonants
(especially) which sometimes indicates a difference of meaning,
and is written h or ', pJia or p'a, tha or fa, kha or ft a. It is not
often of importance.

The above is not an exhaustive account of possible combina-
tions of consonants in Bantu, but sufficient for purposes of
illustration in what follows.

If we take, then, Table A as a useful aid, and remember once
for all the limitations already described (Chap. II), under which
all general theories about Bantu as a whole in this book are
made, it may be said that every Bantu word, however lengthy
and elaborate its form may be, is likely to be traceable to
a single sound, which may be regarded as the germ of the

CH. iv GERM-SOUNDS 35

word. As a sound, it may be called a Monosonant i. e. any
sound capable of separate pronunciation, whether represented
by a vowel, semi-vowel, or consonant rather than a mono-
syllable. A monosyllable usually implies a vowel, or a com-
bination of vowel and consonant. But all consonants, except K,
P, T, are capable of a degree of separate sonances, and the root-
meaning of a word in Bantu is often attached to a single con-
sonant, followed indeed by a vowel (necessarily in the case of K,
T, P), but by different vowels, carrying the possibility of implied
difference in meaning. Hence Monosonant seems a fitter term
than Monosyllable to describe the rudimentary germ of speech.

Such monosonants appear to be the ultimate basis of speech
in Bantu, and in the next chapter grounds will be given for
believing that they can reasonably be recognized as such, also
that a monosonant starts from the first with many capacities
for differentiation and consequent expressiveness, that it can
acquire further definiteness and distinction by juxtaposition
with a similar element, and that at length the simplest form of
what can be called a word in Bantu emerges by a union of two
such germs. Once united as parts of a whole, a new distinction
appears between them, one taking (as it were) the lead and
expressing the main or root idea of the combination, the other
in a subsidiary position as qualifying the first, giving it a special
aspect or bringing it into relation with other words. These
may conveniently be called the Radical and Formative elements
of a word, as commonly recognized in other Families of speech.

In Bantu, however, it may be found that both radical and
formative, and especially the latter, are more clearly recogniz-
able throughout ; their relations and functions in word-building
more regular and characteristic ; the power of selection and
employment at will for appropriate expression of ideas more
free. Hence it is possible that the study of these basal or
germ sounds may throw light on the early history of human
speech among other things on the nature of the distinctions
commonly known as declensions in nouns, conjugations, moods,
tenses, &c., in verbs, to explain which Comparative Philology
has not hitherto been able to penetrate far into the past.

CHAPTER V

MONOSONANTS, INTERACTIONAL AND
ADVERBIAL

So far it has been stated, mainly as theory and by anticipa-
tion, that the recognizable basis of all Bantu speech is to be
looked for in single sounds, called here Monosonants. At this
point it will be convenient to present the simplest of them in
a tabular form, arranged alphabetically, and in each case with
a vowel, though the vowel is not in itself necessary (except to
give sonance to K, P, T). This Table B is in fact a rearrange-
ment of the letters given in Table A.

TABLE B.

A

E

I

U

Ba.

Be

Bi

Bo

Bu

CHa

CHe

CHi

CHo

CHu

Da

D'e

Di

Do

Du

Fa

Fe

Fi

Fo

Fu

Ga

Ge

Gi

Go

Gu

Ha

He

Hi

Ho

Hu

Ja

7e

Ji

Jo

Ju

Ka

Ke

Ki

Ko

Ku

La

Le

Li

Lo

Lu

Ma

Me

Mi

Mo

Mu

Na

Ne

Ni

No

Nu

NG'a

NG'e

NG'i

NGo

NGu

Pa

Pe

Pi

Po

Pu

Ra

Re

Ri

Ro

Ru

Sa

Se

Si

So

Su

Ta

Te

Ti

To

Tu

Va

Ve

Vi

Vo

Vu

Wa

We

Wi

Wo

Wu

Wa

We

Wi

Wo

Wu

Ya

Ye

Yi

Yo

Yu

Za

Ze

Zi

Zo

Zu

This table gives no monosyllables. Are these the basis of
all Bantu speech ? If so, how so ?

MONOSONANTS 37

The answer, if any, can only be found by reference to Bantu
as it exists at the present day. Probably no sort of written
record of Bantu speech, in Arabic or European letters, goes
back further than a few centuries (see Preface to Torrend's
Comparative Grammar], and a record of an earlier age on any
scale useful for the purpose of drawing general conclusions is
hardly conceivable. The case is very different with regard to
some other families of speech, Sanskrit, Hebrew, and ancient
Egyptian. Even then the origin of human speech seems far
too remote and inaccessible to make 400 or 4,000 years any
real difference. The important matter is not lapse of time,
but degree of change necessarily involved, not the age but the
stage of a given specimen of speech. Thus an answer suggested
by dialects now spoken in Central Africa may be as satis-
factory (or unsatisfactory) as any other, considered as evidence
bearing on their original form. It is all that can be offered
and must be left to justify itself, without discussion at present
as to their inherent right to be regarded as specially primitive
in type, but only with the remark that there is no inherent
impossibility of one dialect of one family of speech being, if
not primitive, a fairly preserved survival of much that is primi-
tive, and so far a useful guide to sound conclusions.

Mang'anja is the name given to a Bantu dialect spoken in the
region immediately south of Lake Nyasa and along the valley
of its outlet, the river Shire*, and is closely allied to others on
the shores of the lake, especially on the western side. It is
the subject of the book already specially described, Scott's
Cyclopedic Dictionary of Mang'anja. In this book over 150
such monosyllables as those in Table B are given, with short
descriptions and some illustrations, i.e. many more than appear
in that table. The additions may be accounted for by reference
to the explanations appended to Table A, showing the many
combinations of two or more consonants possible, and in some
dialects (Mang'anja among them) commonly used.

Not only do these 150 monosyllables exist as separate ele-
ments of speech, but at least three differences can be recognized
in their common use, differences which seem to be a valuable
indication of stages in linguistic history. They are used as

38 WORD-BIRTH PT. II

(a) simple interjections, (b) interjectional adverbs, (c) verbs,
words which are verb-stems, not merely verb-roots but this
third use is limited strictly to the first column, i.e. monosyllables
ending with the vowel -a. This limitation will be shown here-
after to be important, and in fact marks the difference of a whole
stage in linguistic development.

The name ' interjection ' is not adopted from its suitability
in this case, but because it is the common grammatical term
for a c part of speech ' which is not in a grammatical sense
a word at all, i. e. not capable of entering into such exact gram-
matical relations as the others. Hence, it is commonly credited
with a minimum of value and meaning, and is often almost
entirely overlooked. Certainly this has been hitherto the case
in all works on Bantu which are most generally known, except
in Scott's Mang'anja. Even in Stapleton's Comparative
Grammar of Congo Languages, the whole subject is represented
by a single section under the title ' Onomatopoetic vocables '.
The fuller meaning of interjection, as used of these monosyllabic
rudiments of speech, will be shown later in this chapter.

A single illustration may here be given of the three uses of
monosyllables from the syllable (sound) Pa.

As a simple interjection, Pa may suggest a blow, and a light
rather than a heavy one, on a soft object rather than a hard one.

As an interjectional adverb, Pa may qualify a preceding
noun or adjective, by which its range of possible suggestion
is in a degree limited, e. g. following utsi (smoke) pa suggests
spreading in clouds, following nyumba (house) pa suggests
fullness, crowding. Or it may be used with a verb, e. g. mtengo
(tree), uli (is), pa, straight, upright, in place.

As a verb, Pa may mean variously, do, place, put apart, kill.
But as will be shown hereafter, Pa as a verb is a composite
syllable, and in some meanings certainly a contracted one, and
this is also the case with its further common uses as an adverb
of place, ' here,' and as a preposition of place, * at.'

For the present, the features to be specially noted are these.
As a simple interjection, the monosyllable so used is vaguely
expressive. As a sound, it suggests without by itself defining.
As an interjectional adverb it is qualitative, its suggestiveness

CH. v MONOSONANTS 29

is used to qualify, descriptively or otherwise, another element
of speech, which in turn reacts on its suggestiveness by limiting
it to the particular case. The smoke (e. g.) would not be de-
scribed as upright, nor the house as spreading. In connexion
with its use as verb, it must be remembered that while any of
the syllables in Table B may be, andmost are, used as inter-
jections and adverbs, only those ending in -a are (with rare
exceptions) used as verbs. When this is taken in connexion
with the general rule in Bantu that all verbal forms (except in
the subjunctive mood) end in -a, the conclusion clearly follows
that in this case the final -a involves a second element, i. e. Pa
as a verb is probably p(a) -a. This stage of combination of at
least two simple elements will be considered later. For the
present, the two simpler uses of the monosyllable require
further explanation, and first as to the nature, use, and possi-
bilities of the simple interjection in human speech generally.

Simply as sound, produced by human organs, an interjection
need not differ from other sounds whether of inorganic or organic
nature. It is often simply an imitation or reproduction of them.
But used by a man, as means of self-expression or of communi-
cation with fellow men, the interjection cannot but, from the
first, reflect the human spirit. It may be merely the sound of
the open vowel A, as pronounced (e.g.) in Southern Europe.
But as uttered by a man, and (it may be assumed) addressed
to a man, in an environment including at least two human
personalities, with all their endowments of passion, intelligence,
and will, in the atmosphere and surroundings of a particular
moment in the life of each, with every sense and faculty alert
(however unconsciously) to understand and be understood,
then a mere sound takes on a wholly new and important
character, and may have effects as telling, for illumination
and action, as an electric spark. As uttered by a man, its
force and meaning is inevitably coloured by attitude, gesture,
and look, it is defined by the particular organ or sense affected,
employed, or appealed to, by objects present, instruments used,
and further by the special occasion and general circumstances,
immediate cause or obvious purpose of the utterance, the
speaker's relation to his hearer, his knowledge of the intelli-

30 WORD-BIRTH PT. II

gence and sympathy of the other, and the probable effect upon
him, all the complex elements, in fact, which surround and
enter into the simplest movements of human life. Moreover,
no sound is so simple as not to admit of many degrees and
shades of change. The sound A may be uttered in a colour-
less inexpressive way, be slowly drawled, or have a sharp,
clear-cut effect. In either case its pitch may be high or low,
its tone smooth or rough, musical or grating, its emphasis weak
or strong. It may be uttered singly or repeated, and when
repeated be repeated with one of the changes just described.
In fact, a whole conversation can be carried on, case argued,
transaction discussed and concluded, by sole use of interjections
or ejaculations, largely even by natural variations of one.
Possible combinations of such elements as those mentioned
suggest almost limitless possibilities of a single monosyllable.
Vague, ineffective, almost useless as interjectional speech
may and must seem to those not familiar with it or with men
who habitually use it, it is nevertheless a fact which deserves
far more thorough study than it has received, if ever the
problems of origin and growth in human language are to be
solved. It is possible that the ' speaking with tongues ' of the
new-born Apostolic Church was the interjectional reflection,
through human organs, of ecstatic glimpses of the transfiguring
glories of Divine Incarnation in human life, requiring sym-
pathetic interpretation by one himself on the same high level
of ' inspiration ' to become intelligible on the plane of ordinary
human hearing. Similarly on a lower plane it might be ex-
pected that interjections are well understood by a hearer on
the same level of civilization as the speaker, and yet require
further interpretation to be usefully intelligible to the ear of
the white man. Where writing is absolutely unknown, and
no visual image in the mind affects sound-production, a new
meaning attaches to the expression Living Speech. Compared
with Bantu in its present fluid or gaseous condition, the literary
languages of the modern world are relatively as hard and life-
less as those only known to us in the writings of the distant
past. Bantu language is always and only sound ; speaking, not
writing. Hence the futility of judging of its possibilities by

CH.V MONOSONANTS 31

its present condition only, or of criticizing its supposed clumsi-
ness, limited vocabulary, and lack of exact expression, without
due reference to its present circumstances and the stage of
civilization to which it belongs. Those who know its more
than sufficiency for all present needs are not likely to doubt its
promise in the future.

What has been said so far of the interjection in general, may
now be applied to the present case.

A simple interjection in Bantu may be described as a single
sound or syllable, used as an element of human speech, with
a meaning sufficiently defined by

(a) intrinsic differences as sound, such as duration, pitch,
stress, and quality (the rate, amplitude, and form of air-waves),

(b) special accompaniments, such as facial expression,
gesture or movement, attitude or position, on the part of the
speaker,

(c) general conditions, such as time, place, and circum-
stances, including characters and characteristics of the speaker
and also of the hearer.

The number of possible combinations available for suggesting
difference of meaning may be illustrated as follows :

The syllable Pa may be pronounced either normally or
long or short, i.e. Pa, Pa (faa), PS, (pah). These three
varieties of pronunciation may be attached to P in combination
with each of five other consonants (combinations common in
Mang'anja, for instance, PW t PY, PF, PS, PH = P') pro-
ducing three varieties of the six syllables Pa, PWa, PYa,
PFa, PSa, P'a, eighteen in all. Each of these syllables are
commonly nasalized in Bantu by M prefixed. The thirty-six
resulting syllables may easily be differentiated by light or
strong stress or emphasis, producing a total of seventy-
two. And if further differences (say. three only) of pitch, or
place in the scale, are not used (as in Chinese) to distinguish
words, it shows that Bantu does not exhaust all its natural
resources when it uses seventy-two monosyllables based on the
letters Pa alone. Pe, Pi, Po t Pu can of course be similarly
treated.

Each of the seventy-two can further take on a different shade

3 a WORD-BIRTH PT. n

of meaning, according to the sense appealed to, or affected-
sight, smell, taste, touch, as well as hearing, and each of these be
effectively determined by the various special accompaniments
(of look, gesture, movement, attitude) and general conditions
just enumerated.

If these considerations be applied (as they may be) to each
of the monosyllables in Table B, it will be seen that the
interjectional basis assumed for Bantu is sufficiently broad.
It is obvious, moreover, that objection cannot reasonably be
brought against the interjection on the score of being neces-
sarily too vague for intelligent use. It cannot be set aside or
disregarded as other than a real, even the most real, element in
human speech. Any human utterance, however brief, that is
adequate to the occasion, i. e., sufficiently indicates what is
intended to be conveyed, has as good a claim to be a true
word as many a polysyllabic and sesquipedalian product of
modern scientific nomenclature. The study of it may be
difficult, indeed it has probably been kept out of sight largely
by its difficulty, but it certainly may claim to be useful and
even necessary. From the nature of the case, it can only be
studied on the spot, i.e. among Bantus, and by a qualified
observer, not only familiar with a dialect but sympathetically
appreciative of their modes of self-expression, and such obser-
vation (it must be remarked) is just what cannot be reproduced
by the phonograph. The sound alone is not the whole fact as
a rudiment of speech. And it is not the least valuable feature
of the Mang'anja dictionary so often referred to, that it seriously
attempts to show the value of the interjection, and record
systematically facts throwing light upon its uses and nature.

Again, the interjection does not deserve disregard, much
less ridicule, on the ground that (so far as now appears) the
same sound has from the first very different applications,
supplies the root or basis, in fact, of large and distinct groups
of words, springing from the same root but from different
original meanings of it. In fact, the interjection cannot with-
out violence be detached from its surroundings. Richly sug-
gestive in itself, it takes colour instantaneously from any and
every element in the environment, and illustrates the intense,

CH. v MONOSONANTS 33

if somewhat chaotic and unrestrained vitality of new creations.
It is indeed more living than its subsequent developments,
depends less on itself as sound for its own expressiveness,
challenges the human spirit to understand the human spirit,
with a minimum between them as vehicle of communication.
Its implied attitude is not so much Do you not hear ? as, Do
you not see ? Do you not feel my meaning ?

More than enough has now perhaps been said to indicate the
claims of the interjection on the attention of the student of
Bantu. They may be strengthened by what follows in later
chapters. The significance of its use as an ' interjectional
adverb ' requires now to be briefly recalled and re-stated.
This use is in fact the commonest in existing dialects of
Bantu. A simple interjection placed after a word qualifies
its meaning by added description or merely added emphasis.
Its own meaning is at the same time limited to what is appro-
priate to the idea of the word, to the function of qualifying it.
There is a gain on both sides from the juxtaposition. The
word gains in definiteness, the interjection gains also in useful-
ness. It has a function, although in a sense a subordinate one,
in relation to another element of speech. It is, at an early
stage of language, what the adverb, with its characteristic
grammatical use as qualifying verbs, adjectives, or other
adverbs, is at a later. But it has a further and perhaps im-
portant significance. It suggests the nature of the first step
in the process by which simple interjections in Bantu become
words in the full sense, capable of having and developing definite
relations with other words. In Bantu of the present day, after
doubtless many centuries of existence and change, words have
long come into existence, and the interjection appears mainly
as qualifying words. But it is not unreasonable to infer that
in an earlier and more primitive stage, in which interjections
were the current coin of speech and words had not won their
way to supersede them, interjections were similarly used to
qualify interjections with corresponding effects of action and
reaction upon each, the first of a pair being qualified by the
second, the latter being limited to illustrating and defining the
former. Here it is possible to see the beginning of what at

1214 C

34 WORD-BIRTH

the later stage are called the Radical and Formative elements
of speech, juxtaposition involving mutual limitation and
then gradual differentiation of function, the first member of
the pair giving the dominant idea, the latter so far subordinate
and utilitarian as to give it a further aspect, and eventually
relation. It is easy to see how one tends to represent the
living germ of meaning, and the other the also living but
also more mechanical element in its presentation as a word.
'Missing links ' will probably remain a feature in this as well
as other chains of biological development. But in Bantu at
any rate the gap to be filled may be found less difficult to
bridge than elsewhere. At any rate the suggestion just made
seems in harmony with other facts characteristic of Bantu noted
in following chapters.

It remains to notice in passing another class which may be
called ' Interjectional Adverbs ', belonging to a later stage of
growth, but largely illustrated in the Mang'anja dictionary
and of importance as a feature of Bantu. They are formed
themselves from fully-formed words, mostly verbs, but put back
(as it were) into the earlier (interjectional) stage, not so much
by the instinct of reversion to old type as of attempt to utilize
and extend it. They are used side by side with other ad-
verbs, and generally consist of two or three syllables. Simple
examples are: mpira (a football) utumuka (verb, is blown
up) tumti) tight, related to the verb tumuka. Kavalo (the
horse) aenda (verb, goes) gogodo gogodo, cantering along,
related to the verb gogodola.

CHAPTER VI

ROOTS

THE monosyllables of Table B, assumed as the basis of
Bantu speech, have been so far treated as furnishing interjec-
tions of two classes, the simple and the adverbial, differing only
in use, not in form. And the suggestion has been made, that
in the juxtaposition and (at a later stage) union of interjections
in these two uses is to be found the origin of the (so-called)
Radical and Formative elements, both of which are in Bantu
necessary to constitute a word, whether verb or noun, or at
least are found invariably to coexist in all such words.

The next step is further to examine this all-important
phenomenon, the emergence of the word from the stage of
interjection. If germ, seed, embryo, are terms suited to describe
the interjection, it now becomes inevitable to use the term
Root.

A Bantu root may be described as a monosonant, which is
strictly not yet a word or even in all cases a pronounceable
sound, but a rudiment or germ of speech, interjectional in the
comprehensive sense above explained, sufficiently definite in
meaning to be capable of and invite combination with another
of a similar nature. Such union, when consummated, issues in
a word, capable of definite logical and grammatical relations
with other similar words. Hence the appropriateness of the
word root. Germ and seed lead on naturally to a new stage
of vitality, still (as it were) underground, but demanding
emergence in a new form, based on their union and (as it were)
mutual absorption, the result being a product belonging to
a new region of life the word, verb or noun, with its endless
possibilities of varied growth, modification, and uses for the
purposes of expression. Whatever the interjectional germ,
seed, or root may be, it does not in Bantu become suitable

C 2

36 WORD-BIRTH FT. II

to play the part of a word, without union with another. To
become a verb, a root is insufficient without -a, as final vowel.
To become a noun, it must indeed have a final vowel, not
necessarily a second root, but also and primarily an initial
sound of one of several well-marked and distinct forms. In
other words, a second element must appear in every word
beside the root, in the case of a verb a suffix, in the case of
a noun a prefix.

The view has been implied so far, that the two elements
forming every word are of similar origin, i. e. originally of the
nature of interjections, and it will probably be admitted as
a reasonable working hypothesis that they are so. Incidentally,
reasons may appear for thinking that in some cases they may
be clearly recognized as each equally, if vaguely, significant of
an idea. But taking Bantu words as they must be taken, in
the present stage of development of the whole Family of
speech, with unknown centuries of growth and changes behind
it, it is clear that in such words one element conveys the main
idea, and may be usefully called Root or Radical, while the
other is not indeed without meaning, but in fact and function
secondary to the first, its meaning being used as a qualificative,
perhaps of an adverbial kind. If its qualificative use is mainly
kept in view, for whatever reason, then this second element
may well receive its usual name of Formative, without for-
getting that it has or had a root-meaning of its own.

A further step in definition is useful and almost necessary in
view of the leading characteristics of Bantu speech. The whole
word-building, or word-growth of Bantu rests on the use of
formatives, in the sense just described. On their nature and
position as regards the radical element depend all the distinc-
tions corresponding to such common grammatical terms as
stem, voice, conjugation, mood, tense, person in verbs, and
class or declension and number in nouns. On them rests the
whole system of Bantu concord, and they enter largely into
the important group of words usually classed together under
the name Pronouns.

Formatives are used at least as much after as before the
root, and may therefore be divided into prefixes and suffixes.

CH. vi ROOTS 37

There are some, however, of importance, which in verbs are
never initial or final elements, and when this class requires
separate distinction it may be said to consist of Infixes. As
a rule, however, it is convenient to call formative elements
preceding the root Prefixes, and those following it Suffixes.
More exactly they may be divided thus :

1. Prefixes, including

(a) initial prefixes

(b) pre-radical infixes.

2. Suffixes, including

(a) post-radical infixes

(b) final suffixes, or affixes.

Most finite verb-forms contain all these, often several infixes.
Most nouns have one prefix and one suffix.

The relation of a root, as described at the beginning of this
chapter, to the interjection will appear more clearly in the
chapter following, and later will be given reasons for re-
garding the so-called Prefixes of nouns as entitled in reality to
be called the root-syllables (Chapter XII).

CHAPTER VII

VERBS AND NOUNS

THE position so far reached may by way of recapitulation be
briefly summarized.

Table B has been taken as a list, typical but not exhaustive,
of monosyllables forming the basis of Bantu speech, and Pa
as an example of them.

Bantu, as it now exists, uses Pa

(a) as an interjection, e. g. describing a blow ;

(b) as an adverb, following and qualifying a word, e. g. utsi
pa, smoke in clouds ;

(c) as a verb, with various meanings in different dialects,
but always involving p(a) as conveying the root-meaning in
each case, and a formative termination -a, common to verbs
generally. (Pa has also other important uses.)

The explanation, offered as a key to the existing facts,
suggests the following phases of development :

(a} First, interjections, such as Pa, as the only original
material of speech, suggestive themselves and defined by
environment.

(b} Second, two interjections, used together and reacting
on each other, the first suggestive, the latter tending to be
qualitative, the typical formula being pa pa.

(c) Third, two interjections, united and tending to fusion,
the former conveying the general dominant idea (root), the
latter supplying the (formative) element necessary for relating
it to other similar combinations, the typical formula being
pa-pa, whence p(a-p}a, pa. This union constitutes a word,
verb or noun, and all forms of these and other words consist
of recognizable formative elements used before or after
a root.

So far as ' root ' suggests root-meaning, it will of course be
recognized in the first and second as well as in the third stage,

VERBS AND NOUNS 39

in the interjection as well as in verbs and nouns. But its
importance lies in its use as a particular element in words,
which begin (ex hypothesi) with the third stage. Hence it is
used here mainly of this third and of subsequent stages of
word-growth.

It may now be broadly stated, that each of the sounds given
in Table B, taken as samples of all the more important, and
not as a total of Bantu resources, appears recognizable in one
or other of the dialects of Bantu, as possessed of three distinct
values (reflecting, it would seem, stages of growth), viz. : Inter-
jectional, Radical, and Formative, and that the most character-
istic expression of Bantu as Living Speech is found in the full
and free use of these materials in word-formation, especially
of the latter two. There is no reason for expecting that any
one dialect will alone supply evidence to justify the statement.
What is suggested is, that as the survey of Bantu becomes more
comprehensive and complete, evidence is likely to accumulate,
and meanwhile any single dialect or group of dialects may
furnish a quite sufficient foundation.

A simple illustration, founded on the consonant T, may now
be given by way of anticipation, and afterwards in a more ex-
panded form:

Tidnatititititi. This group of sounds might be written
phonetically for English readers teeyahna teety teeteetee. Used
in Living Speech it may mean ' the children are tapping light-
ly', or again, 'the children are putting themselves in line,'
or bear other meanings, determined quite sufficiently by
the circumstances or a simple gesture, a movement of the
finger or a sweep of the hand. Here it shows a similar sound
ti (tee), used as interjection, formative, and radical.

The group may be divided grammatically into Tiana titi
tititi, and analytically, Ti-ana ti-ti, ti ti ti. The first of these
words is in Nyanja a noun, root -ana 'child ', with the formative
prefix ti, indicating plural number and with a diminutive
meaning. Thus tiana ' little children '. In the second word,
the first ti is the same as in ti-ana but now the formative is
prefixed to a verb, effecting grammatical concord between the
verb and its subject tiana, while the second is a verb, common

40 WORD-BIRTH PT. n

indeed in Bantu but peculiar in form, as one of the very few not
ending in -a. The meaning of the verb is general, ' do, act,'
and the union of prefix and root without any infix to mark
tense indicates action simply. Thus, titi ' they (children) do ! '
The repetition three times of the interjection suggests a re-
peated action of one of the kinds suggested by the sound ti,
such as a series of taps (cf. tick tick tick) or a series of the
objects present (one, two, three).

No identity of origin is here asserted as to the similar sounds
used as interjection, formative, and radical. The two latter
must have long linguistic histories behind them. What may
be noticed is, that in Living Speech sounds vague in themselves
convey in each particular case a meaning quite sufficiently
definite for the speaker's purpose, being defined by his
tone, gesture, or look, by the context of the communication,
the hearer's sympathetic attention, and the circumstances
generally. Vary any of the concomitants, and the sounds
may bear a different but not less clear meaning. The sound-
material meagre in form, rich in suggestion is not so much a
word-germ as a multitude of germs, and planted in the soil of
infinitely varying circumstances gives rise quite naturally to
words of similar and also of widely different meaning. This
is an explanation, as far as it goes, of the fact in Bantu and
in other languages, that roots identical in sound may have
diverse significance. The key to original divergence, if such
it was, has passed with the circumstances of the period. Once
begun, its history may not be recoverable. Nevertheless it
may well be a fruitful study to trace as far as possible the con-
nexion of groups of words obviously springing from a common
root, and estimate the trend in each case back towards a
single source.

Such study is endless and to the general reader irksome.
At the risk of defeating one of the objects of these chapters,
some further illustrations are added, based on the same letter
as before, T, and again drawn mainty from Nyanja.

T, combined with the vowels, gives the syllables TA, TE,
7Y, TO, TU. Each of these may be recognized in Nyanja
as interjection, radical, and formative.

CH.VII VERBS AND NOUNS 41

i. Interjection, simple and adverbial.

Ta, with its many possible varieties simply as sound,
conveys the idea (among others) of striking, with suggestive-
ness as to force used, object struck, instrument employed.
It conveys also (for whatever reason) an idea of stretching,
open, extended, flat an idea observable in many verb stems
based on this sound. (Cf. Latin jacZre andjacere, strike and
struck flat.)

Te suggests the idea of cutting or tearing a firm material
or substance, implying a certain degree of force. Tati te
means ' we tear (it) '. It denotes also, with a particular in-
tonation probably, clear, nice, exact definition or completion,
e. g. in describing the roundness of the full moon, the com-
pletion of a garden plot kuzungtdila te ' neat rounding off' ;
also with a sharper intonation, splitting and scattering, e. g. of
dry seed-pods, some of which 'go-off' like a pistol-shot in the
African forest. The meaning of openness, stretching, clear
space, is reflected in many words based on 7>, as well as Ta.

It must be remembered that in Te, as in other cases, E
often represents a coalescing of A and /.

TV, as has been observed on p. 40, is used as a simple inter-
jection. As an adverbial, it qualifies with an idea of large
amount, or rather completeness, fullness, tightness, e. g. kujala
ti ' to be full up ', madzi ti ' water to the brim '.

To, pronounced shortly as tdh, describes certain sounds,
e. g. discharge of gun, and acts corresponding to them. Also
as adverbial, it has an idea of fullness, compactness, complete-
ness, similar to Ti.

Tu, pronounced shortly as too, is used similarly to To,
e. g. of a gun-discharge, and adverbially of a leaping move-
ment (as of a locust, or in ball-play) or of falling rain.
Repeated two or three times it suggests objects in series or
succession, as TV.

Such interjections as Twe, Twi, Tyo, Twa, Tya, Tso, Tst,
Tse, can also be illustrated from Nyanja. (See Scott's
Dictionary.)

42 WORD-BIRTH PT. II

2. Radical.

In passing from interjection to radical, the stage of words
is entered. In Bantu, all words consist of a radical and a
formative element, as already stated, but though as elements
of a verb or noun they are inseparable, they may be here con-
sidered separately.

Of the five syllables Ta, Te } Ti, To, Tu, the first, Ta, is used
as a verb in Nyanja. It is the only possible regularly formed
verb in the series. As a word it is composite, including
a radical and formative element. As a verb, the final -a is
formative, being characteristic of verbs as such, but admitting
of variation in some parts of them. The radical element is
therefore Ta (or possibly T only), and as a verb Ta represents
T(d)-a. It has at present in Nyanja two meanings, (a) ' do ',
a general meaning, as might be expected from an interjec-
tional origin and simple monosyllabic form, and (b) ' finish ', a
meaning less vague, and easily connected with the first.

Te-a or Te-ya, Ti-a or Ti-ya, To-a, or To-wa, Tu-a or Tu-
zva, all occur as verbs in Bantu, but as disyllables may be
passed by here. Their form suggests that Ta also, as a verb,
involves two syllables, as well as elements, Ta-ya or Ta-wa,
whatever the second (formative) may be. Both Ya and Wa
are also common verbs of general meaning in Bantu.

Ti is a syllable deserving further remark. It is used as
a verb, and as such is one of the very few which appear to
consist of a single element, and at any rate does not, like other
verbs, take a final formative -a. It is remarkable also as
extremely general in meaning, indicating little more than an
activity, which may be translated according to circumstances
as ' do, act, say, think', and moreover in various forms supplies
the place of an almost universal subordinative conjunction,
i.e. a form suggesting the connexion of a dependent on a
principal sentence without in itself defining the meaning of
the connexion. (See Chapter XV.)

CH.VII VERBS AND NOUNS 43

3. Formative.

As a formative, the syllable Ta is used in the Nyanja
dialect in at least three ways :

(a) as a verb-suffix, both as Ta and (nasalized) Nta y as in
the verbs ta-nta, ka-ta, pa-ta, and many more.

(b) as a verb-prefix, giving emphasis to an imperative
form, as ta-taya ' do throw away ', ta-tanta ' do go across '.

(c) as a tense-infix, denoting completion, e. g. i-ta-ta, it is
finished ('done finish'). In Swahili -ta- is the regular tense-
sign of the future.

In (b) and (c] it is easy to recognize in Ta the force of the
same syllable as a verb ' do, finish '. In fact it is the verb used
as a formative. In (a) the connexion of meaning is not so
clear.

But the whole subject of formatives is a natural transition to
that of word-growth in general, and of all those developments
which are the chief subject of Grammars. It will therefore
be dealt with in the next chapter. But to complete the sketch
of the use of the T- monosyllables as an element in Bantu
speech, it may be briefly traced here in some noun-forms,
though these also will be more fully treated hereafter.

It has been already stated that to form a verb, a root must
have -a attached to its final vowel. To form a noun, it must
normally have one of a limited number of prefixes. It is
interesting to note that each of the monosyllables TA, TE,
77, TO, TU appears in common Bantu words as the base or
radical which with a noun-prefix forms a noun. Without con-
jectures (though some are obvious and tempting) as to the
connexions of meaning between the noun and its radical (and
interjectional) base, it is enough to adduce the following
from Nyanja and other dialects, each with the noun-prefix U-.

U-td) a bow (weapon).

U-te, spittle.

U-ti t inner part, or substance of a tree. M-ti, the same
radical with another noun-prefix, is a common word for ' tree '
in Bantu.

44 WORD-BIRTH PT. n

U-to, any sticky fluid, gum, &c., like u-te but more general.
M-to, is a common word for ' river '.

U-tu, nature or substance of a thing, sometimes human
nature, m-tu being the common word for ' man ', while ki-tu is
' thing'. In some dialects mu-tu means 'head', and in these
mu-ntu (tu nasalized) is used for ' man '.

This illustration of the uses of the syllables based on T
might be largely expanded, and similar illustrations of the
other syllables in Table B be multiplied almost indefinitely,
each being taken in turn and traced in the different forms and
meanings of words of which it is the base or root. P and K
would prove especially fruitful of interest and suggestiveness,
followed out in any good dictionary of a Bantu dialect. But
it is in its grammar that the characteristic features of Bantu, as
a distinct Family of speech, are mainly to be looked, and it is
to the development of the formative elements, which is for the
most part singularly clear and striking, rather than to that of
meaning, which is often intricate and obscure, that attention
will mainly be drawn in the chapters which follow.

Before leaving this part of the subject, two further illustra-
tions may be briefly given :

With one or two possible exceptions, all the monosyllables
ending with -a in Table B actually occur in Bantu as verbs, in
their simple or nasalized form, or both. Many of them convey
such simple or general meanings as might be expected from
their short and simple form, e. g. ' be, do, make, act, come, go ' ;
others have special meanings and some several meanings, which
suggest that their present monosyllabic forms are not original,
but represent contraction or actual phonetic decay.

Secondly, the interjectional meanings of Pa were briefly
illustrated at pp. 28-9, but a note may be added on the other
monosyllables formed on P with a vowel, as found in Nyanja.

Pe, (a) pronounced rather shortly, may indicate completion,
e.g. kudya pe 'finish eating, get it over'; yamba pe 'get the
first stage over ' ; endape ' go and have done with it ', or again,
light beating [like Pa] as of wind, pepa pe ' blow softly '.

(b) With a prolonged e (pay), it suggests quiet, rest, silence,
e. g. kalipe f it is still '.

CH. vii VERBS AND NOUNS 45

(c) With a short e (peh, or as in ' met '), it describes opening
out, show, e. g. nsalu pe, calico spread out.

Pfa, Pfe, Pfi, are used in describing water spirting, bark
being stripped from the tree, a weapon piercing, respectively.

Pi is appropriate to such actions as a sharp blow, stirring
porridge with a stick, firing a gun, extinguishing fire, clapping
hands, sucking with the lips, covering so as to conceal an
object, and others.

Po is variously descriptive of crackling or splitting, clapping
the hands, sound of a pop-gun, a fall or breakage, closing or
shutting, ending and completion.

Pu is used after words expressing shutting together, in-
haling a smell, striking a blow, wind blowing, water pouring.

Psa is used of squeezing, grazing slightly, dragging a net,
sweeping ; Pya, of darkness closing in, of something thin or
slight ; Psi, of pressing, and of sucking ; Pwe, of breaking,
bursting ; Pwi, of splitting, cracking, also of sewing ; Pyu, of
whistling, and also to describe red colour.

All the following interjections find appropriate use in
describing the act of striking according to the instrument used,
object struck, force of the blow, &c. :

Ga, Ba, Be, Ndwa, Nde, Mbi, Pi, T.i, Di, Pwa, Pwe, Wi,
Pu, Pi, Go, Gwa.

These illustrations should be taken in connexion with the
general description of the possibilities of the interjection
previously given.

PART III. WORD-GROWTH

CHAPTER VIII

VERBS

So far, the life of Bantu speech has been traced in its three
earliest stages : first the simple interjection, owing its sugges-
tiveness partly to sound-variation, partly to the whole circum-
stances surrounding and tending to define each particular
utterance ; secondly the interjection further defined in a degree
by juxtaposition with another preceding it and gradually
limited to a semi-subordinate or adverbial use in relation to
the first ; thirdly, the union of these two interjectional germs
resulting in the birth of a word, consisting essentially of
a radical and a formative element, analogous in a certain
sense to soul and body, a word which is grammatically a verb
or noun. This third stage will now be taken as a new basis,
and further steps traced by which advance in social evolution
presumably met its own needs of increasing definition of
meaning by further differentiation of the earliest forms of
words. As this is meant to be an Introduction to Bantu speech
generally, and not a formal account of its grammar, what
appears to be the natural order will be followed which means
the treatment of the verb first, and not, as is usual in Grammars,
of the noun.

It has been noted that the same simple monosonant or
monosyllabic elements (Table B) supply the material of each
of the three stages already dealt with. The same is true of
the further stage now to be considered. The first step in verb-
differentiation is effected on the general principle that any
of the monosyllables of Table B which end in -a, can be
combined as a suffix with any others in the same Table to
form a new series of verbs. Though the final -a of the new
compound verb properly belongs to the whole verb as such, it

VERBS 47

will be convenient to regard the new final syllable as a suffix
itself ending with -a.

The gain of this combination is obvious. The simple mono-
syllabic verb of vague and several but suggestive significances,
and ripe (as it were) for a further advance in definiteness, is able
to provide itself with a considerable number of new forms,
each fitted to give definite expression to one or other of the
phases of suggestiveness contained in the simpler original, and
so far to fix and stereotype them as separate verbs. The
process is carried out with considerable regularity and com-
pleteness, and marks the first of a series of stages, in which
successive suffixes, applied in a certain order to the original
root -verb, furnish abundant means not only of embodying
particular varieties of meaning in new verb-stems, but of
exhibiting these meanings under the various (grammatical)
aspects of Voice (active, passive, and neuter), Mood (indicative
and subjunctive), and (partly) Conjugation (affirmative and
negative).

The simple combination of two monosyllables, the first being
the root and the second a suffix ending in -a, forming a new
verb, with a distinct meaning simply as a verb, will be first
illustrated.

The twenty syllables theoretically available in Table B (not
of course by any means all that are available in Bantu) are as
follow :

A, Ba, CHa, Da, Fa, Ga,Ha, Ja, Ka, La, Ma, Na, NG'a,
Pa, Ra, Sa, Ta, Va, Wa, Wa, Ya, Za, (twenty-two).

These syllables used as suffixes to all the monosyllables
in the Table would produce over 2,000 disyllabic verbs.
Allowing on the one hand for comparative rarity in Bantu of
some of the twenty-two syllables (e.g. ffa), comparative
vagueness in the sound of others (CHa tending to be pronounced
like Ja,Fa like Va, Walike Ba or Va, La like Ra),and on the
other for a preference in many for a nasalized form when used
as suffix, it may be conjectured that about 2,000 such disyllabic
verbs will be found to form the ground-work of a large pro-
portion of the total of verb-forms in any given Bantu dialect.
The connexion in meaning between the disyllabic and the

48 WORD-GROWTH PT. in

root verb may be difficult to trace, or (under the countless
influences which act and react on living speech) be wholly
lost, but in form the great majority will consist of some such
simple combination as given above.

The syllable Ta may again be taken for illustration, in
connexion with the Nyanja dialect, and the opportunity will
be used to show not only some of the disyllabic verbs formed
on it, with their present ordinary meanings, but also some of
the trisyllabic and polysyllabic ; the formation of which will be
explained later.

Ta, as an interjection, has been already mentioned as
indicating (a) striking, (b) stretching, and as a verb meaning
(a) do, (b} finish.

It is combined with the suffixes

-da, (nasalized) Ta-mba 'spread'. Hence Tamba-lala,
Tamba-tula, Tamba-suka, Tamba-sulila, Tamba-lilila, Tambwa,
Tambitsa, Tambsa, Tambsya, and other verbs.

-da, (nasalized) Ta-nda 'spread, extend'. Hence Tanda-
la, Tanda-lika, Tand-iza, Tand-izika, Tand-ula, Tand-uka,
&c.

-ga, (nasalized) Ta-nga ' make a beginning '. Hence Tanga-
lala, Tanga-ta, Tanga-tila, &c.

-ja, used for -ga in some forms of the verb. Ja-ja is
a common verb in Swahili.

-ka can be recognized in the extended compound Taka-nya
' spread ', also in Taka-lala, Taka-sa, Taka-sika y Taka-ta, &c.
Tdka is also a very common Swahili verb, ' want'

-la, is seen in Tala-la, Tala-ma.

-ma, 7#-7#0, whence Tama-nda'showtrustm', Tama-nga 1 run'.

-pa, Tapa 'put in parts, divide, spread'. Hence, Tapa-sa,
Tapa-ta, Tap-ika, Tap-ilana, also Tapsya, Tampsya.

-sa, seen in Tas-ika, Tasa-ma, Tasa-mika, Tasa-lila, Tasa-
lula, also Tansa.

-fa, Ta-ta ' pant, struggle ', and (nasalized) Ta-nta ' cross
(stretch, extend) over '. Hence Tanta-lika, Tanta-niza, &c.

-wa, Ta-wa ' run away '. Hence Tawa-lika.

-ya, Ta-ya ' throw away '.

-za, (nasalized) Ta-nza ( stretch '.

CH. VIII

VERBS

49

The above is a mere selection from words at present used in
one dialect not by any means exhaustively examined. Yet it
will be seen that thirteen out of twenty-two monosyllables in -a
are used as suffixes to form disyllabic verbs. Of those not
found here, i. e. -cha, -fa, -ha, -na, -va, other dialects might
supply examples. And the various series of suffixes inci-
dentally noted, (two, three, four, or five together in different
arrangements), could be added to indefinitely.

It seems unnecessary to multiply illustrations, though Pa
would furnish an interesting series of such disyllabic verbs as

Pa-mba, Pa-cha, Pa-nda, Pa-nga or Pa-nja, Pa-ka, Pa-la,
Pa-ma, Pa-na, Pa-pa, Pa-sa, Pa- fa, Pa-wa, Pa-ya, Pa-nza.

And Ka another:

Ka-mba, Ka-nda, Ka-nga, Ka-ka, Ka-la, Ka-ma, Ka-na,
Ka-pa, Ka-sa, Ka-ta, Ka-wa, Ka-nta, Ka-za.

As form rather than meaning is now in view, a table will
sufficiently suggest how each vowel in Table B plays its part
in the production of disyllabic verbs. Gaps in the Table
merely represent a limited experience, not that difficulty
would be found in filling them from a wider knowledge of
Bantu.

P, T, L.

-nda

Pa-nda

Pe-nda

Pi-nda

Po-nda

Pu-nda

Ta-nda

Te-nda

Ti-nda

To-nda

Tu-nda

La-nda

Le-nda

L i-nda

Lo-nda

Lu-nda

-nga
-ka

Pa-nga
Ta-nga
La-nga
Pa-ka

Pe-nga
Te-nga
Le-nga
Pe-ka

Pi-nga
Ti-nga
Li-nga
Pi-ka

To-nga
Lo-nga
Po-ka

Pu-nga
Tu-nga
Lu-nga
Pu-ka

Ta-ka

Te-ka

Ti-ka

To-ka

Tu-ka

La-ka

Le-ka

Li-ka

Lo-ka

Lu-ka

-la

Pa-la

Pe-la

Pi-la

Po-la

Ptt-la

La-la

Le-la

Li-la

Lo-la

Lu-la

-sa

Pa-sa

Pe-sa

Pi-sa

Po-sa

Pu-sa

-ma

Pa-ma

Pe-ma

Pi-ma

Pu-ma

La-ma

Le-ma

Li-ma

Lu-ma

-fa

Pa-ta

Pe-ta

Pi-ta

Po-ta

Pu-ta

Ta-ta

Te-ta

To-ta

Tu-ta

-mba

Pa-mba

Pe-mba

Po-mba

Pu-mba

Ta-mba

Te-mba

Ti-mba

To-mba

Tu-mba

La-mba

Le-mba

Li-mba

Lo-mba

Lu-mba

1214

D

50 WORD-GROWTH

Here again illustrations might be multiplied indefinitely.
For the present purpose it is enough if the above clearly
indicate a stage and a principle in the growth of the Bantu
verb. The disyllabic verbs are in fact so numerous and
important as quite to overshadow the simpler verbs on which
they appear to be founded, and in general it is quite sufficient
to trace verb-forms back to one of these as a Primary Verb
or Root.

This chapter may well be closed with a quotation from
Scott's Manganja Dictionary, part of an incidental note on
MT-, and following a short table of disyllabic verbs like the
preceding. 'The same analysis can be made of the whole
language : and the general grip of the language resolves itself
into (a] the general idea contained in the simple L or D, (b] the
various modifications (often slight) introduced by the different
vowels, La, Le, Li, Lo, Lu, (c) the extension of these into such
pronounceable corrections as La-mba. Le-mba, Le-nda, Lo-nda,
&c., and (d) the modifications of these to produce the various
syntactical ideas of normal grammar.'

CHAPTER IX

VERB-SUFFIXES

FROM the preceding chapter the question naturally arises :
Has each of the various suffixes mentioned a specific meaning
in connexion with the root to which it is suffixed ? and if so,
can it be accounted for ?

Here an admission must be made and a line drawn. It
appears that Bantu research has not so far produced any
comprehensive collection or comparison of verbs with similar
suffixes, e. g. of disyllabic verbs ending with -nda, -mba, -nga,
-la, -fa, &c., on which to base a general conclusion as to the
idea conveyed by the suffix as such. Probably many more
dictionaries will have to be compiled and grammars written,
before a satisfactory collection is possible. Much less has it
been possible to connect such meaning of the suffix with other
meanings of the same monosyllable as interjection or verb.
Yet the general identity of these monosyllables, as sounds,
in their various uses is obvious, and holds out a prospect of
fruitful investigation in the future.

But a line may be drawn among the suffixes, important
though by no means in all cases clear. They may be divided
into two classes. Some suffixes once attached to a monosyllabic
or even disyllabic root-verb appear to become an organic and
inseparable part of it, fixing upon it a meaning often widely
divergent from that of the original verb, and constituting the
whole form itself a new and independent root in relation to all
subsequent grammatical changes. They become as it were a
woody part of the word-growth, retaining essential usefulness as
part of it, with a certain loss of vitality and power of removal or
alteration. Other suffixes have a large amount of independent
vitality. They can be used practically at pleasure. Each
supplies a special aspect of the root-idea, and is suffixed to it
for this purpose. Each can be removed, and another put in its
place, to alter the aspect, or they can be used together in

D 2

5* WORD-GROWTH PT. ill

various combinations or groups. They resemble the green and
growing part of a vegetable organism. The two classes to some
extent run into each other, but may usefully be considered
apart.

To the former class of Fixed Suffixes belong for the most
part

-(m}ba, -(m)pa, -(n}da, -(ri)ta, -(n)ga, -(n}ja, -la and -ra,
-(n}cha, -fa, -ma, -sa, -va, -wa.

* Practically these may be regarded as merely formatives,
that is, elements whose original, living, organic, and particular
meaning has become merged in a generalized, semi-mechanical,
and in a sense subordinate function. United with a monosyl-
labic root, they form a compound which is itself a new and
permanent root-verb for all the purposes of dictionary and
grammar. The suffix singles out and crystallizes (it may be
supposed) one of the possible varieties of meaning suggested
by the root, how, it is difficult to say, but possibly by giving
prominence to one of those elements in the environment of living
speech which have been so often alluded to, as able to supply
sufficient practical definition to the vague but rich suggestive-
ness of the root itself. The monosyllables Fa, Pa, Ta, which
as separate verbs describe acting, making, placing, doing, and
Sa, Za y often used of movement, might well lend themselves as
suffixes to such a use. Only in the case of Ma, used as a suffix,
does a fairly definite idea seem traceable in many of the com-
pound verbs including it, viz. relative fixity, of condition, state,
position an idea which appears in the common verb ima, also
sima, simama, and others. Such verbs are sometimes recog-
nized as forming a class, and called Stative. Another group,
sometimes called 'Extensives' (in contrast with 'Intensives'
described at p. 56), is distinguished by an extended suffix, or
rather combination of suffixes, and contains the idea of action
on a large scale, sweeping, wholesale, comprehensive in nature or
effects. Such ' extended ' suffixes are -aila, -aula, -olola, -ulula,
and others.

It remains to be noticed here that the Fixed Suffixes are
themselves found combined in some verbs, producing trisyllabic
and still longer forms, e.g. Ta-ma-nga, Ka-ma-ta, A-mba-sa,

CH. ix VERB-SUFFIXES 53

Ta-mba-la-la, and so on. But in general verbs so ending are
regarded as base-words, not directly resolvable into simpler
elements, and assumed as such in dictionaries, to which
recourse may be had for further details.

Suffixes which are movable, or partly fixed and partly
movable, are the following

-sa (also -za, -sha, -zhd).

-la (-ila, -ela, -uld)

-na

-ya

-ka (-ika, -eka, -iika)

-wa

All these suffixes are of great importance, and constitute
one of the most living elements in Bantu speech as it at present
exists. Within limits hereafter outlined, they can be used or
not at pleasure as suffixes of any of such verb-stems as
previously described, from monosyllables upwards. In each
case they have a specific effect on the meaning of the verb,
and they are used not only singly but in many combinations,
freely formed at each speaker's will and only limited by his
powers of utilizing their meaning to express his own. In fact,
this is one of the main features which give the Bantu verb its
extraordinary richness, actual and potential. Any verb-stem,
with a suitable meaning, can as a rule be made the base of
some twenty or thirty others, all reflecting the root idea in
various lights, sometimes curiously limited by usage to a par-
ticular aspect and limited significance, mostly quite free and
unrestrained in growth, and each again bearing the whole
luxuriant super-growth of voices, moods, tenses, and person-
forms, to the utmost limits of its powers of logical extension.
To follow out a single root-verb in all its possible derivatives
and their grammatical developments in all their forms, though
strictly germane to the purpose of this Introduction, would
require a disproportionate mass of illustrative detail, which can
be found elsewhere. (See Scott's Mang'anja Dictionary,
Stapleton's Comparative Grammar and Vocabulary of Congo
Dialects, and the introductions to Madan's English-Swahili and
Swahili- English Dictionaries. )

54 WORD-GROWTH PT. m

Here some brief explanations may be offered of each of the
six movable verb-suffixes just enumerated. Only it must be
remembered that these are peculiarly living parts of present
daily word-formation in Bantu, and that therefore remarks
can hardly be made upon them which will not seem to require
modification, when viewed in the light of any particular
dialectal usage.

wa

As final syllable of any verb-form, -wa in general denotes the
' Passive Voice '. As such, it may be suffixed to any verb-stem
whatever (the final -a of the stem being dropped or changed to
-i or -e in the new combination, see below), and the whole
structure of a passive voice, conjugations, moods, tenses, &c.,
be raised upon it, as well as other verb-stems, and the many
nouns, which can be formed on verb-stems at every stage of
evolution. Thus in Swahili, Pend-a, active, ' love,' Pend-wa,
passive, ' be loved/

As to origin, Wa is a syllable widely used in Bantu as a verb
meaning ' be ', and it is obvious to compare the English passive
' be love-d ' with a possible Bantu Passive of similar elements,
pend-iva, ' love(d) be.'

ka

Here ample allowance must be made for differences between
dialects and within a dialect, but one of the most general and
regular uses of -ka as final suffix is to denote a 'Neuter Voice',
side by side with the passive, in capacities for use in further
verb- formation (as indicated above), similar in meaning and
translation, but without implied reference to an agent. Thus

Pend~a, active, 'love.'

Pend-wa, passive, ' be loved ' (by some one).

Pend-eka, neuter, ' be loved ' (a loved object, in general).
(As in case of -wa, the connecting vowel is i or e.)

The remaining suffixes are regularly used to form the four
classes of derived verb-stems, commonly distinguished as
applicative, causative, reciprocal, and reversive. Only a
general description of these classes is given here.

CH. ix VERB-SUFFIXES 55

la

(1) With the vowel i or e preceding it, i. e. as -ila or -ela,
this suffix has the effect of exhibiting the root-idea as in rela-
tion to another, and in almost any of the relations commonly
indicated in English by a preposition following the verb. It
may be briefly described as attaching such words as ' to, at,
from, for, by, with, against, about ' to the root-idea, and gener-
ally as a substitute for prepositions in Bantu. Vague and un-
satisfactory as this may seem, especially for literary purposes,
and judged by the standards of written speech, this vagueness
can in many ways be lessened by the use of various compound
prepositions and prepositional phrases, but in practice gives
little inconvenience, where livmg speech has its natural setting
in life and circumstances.

(2) With the vowel u prefixed, a wholly different effect is
produced. The verb becomes what is called reversive. The
suffix -ula has some of the force of the English prefixes un-,
and dis-, giving an idea of reversal, change to the opposite, or
still more generally of analysis, disintegration, dispersal, anni-
hilation. In this suffix, the vowel -u appears as the more im-
portant and distinctive element, since the -la is often exchanged
for -ka, with the effect of -uka producing a neuter instead of
an active voice. Alike in 'ila, -ela, and -ula, the consonant
L shows a characteristic weakness and evanescence as a sound
between two vowels, and in some dialects disappears, leaving,
as in Swahali, -ia, -ea t -ua as carrying all the force of the fuller
forms. This relative unimportance of L and importance of the
vowel e, i, or o may indicate that -la as a movable suffix is not
the same in origin as fixed -la. -ola, as a suffix, is commonly
a contraction of a-u-la.

sa

Other forms of this suffix are -sha, -za, -sha, -tsa, -dza, and,
like -la, are preceded by i or e, i. e. -tsa, -isha, &c. The general
effect of their use, as suffixes to any verb-stem, is to denote
an exhibition of force, energy, power, or will in connexion with
the root-idea of the verb. The description is necessarily com-
prehensive in terms, because it has to include two quite

56 WORD-GROWTH PT. in

distinct fields of meaning covered by the same suffix, which
are usually called intensive and causative. The fact is, the
idea of force or will conveyed by the suffix may be conceived
either as intensifying the idea itself of the root, acting as it
were within it, or as external, and taking effect upon it.
Thus in Swahili, /*# ' strike \pig-isha either (a) ' strike hard '
(intensive) or (b) ' cause to strike ' (causative). Many curious
and delicate shades of meaning can be conveyed by this suffix,
e. g. many kinds of causation, not only direct, such as force,
compulsion, but also indirect, such as persuasion, suggestion,
permission, or merely acquiescence in an effect or consequence.
In general, its significance is best represented in English by an
appropriate adverb, just as that of the suffix -la by a pre-
position.

ya

This suffix is often used as a causative in the sense just
explained, and may be classed with -sa. And it may be noted
that ya, sa, za, as monosyllabic verbs, commonly denote
energy in motion, ' come, go.'

na

This suffix, unlike -la and -sa, is always preceded by -, and
-ana supplies the reciprocal class of derivative verb-stems.
Its effect is to attach to the root-idea that of mutuality, inter-
action, action and reaction, combined action, or association.
Thus in Swahili, pendana ' love each other ', pigana ' exchange
blows, fight', liana 'all cry together', fuatana 'follow with,
accompany, go together '.

A list of the possible or even common combinations of
the above, or illustrations of their use in refinements of ex-
pressiveness, is out of the question here. For instance, the
suffix ita may itself be repeated twice or even three times
with suggestion of 'to' (an object), 'with' (an instrument),
' for ' (a purpose), i. e. -ila, -ilila, -ililila may be combined
with -ana, as -ilana> or -anila y or with -iska, ilisha, ilanisha,
and this with -wa, forming a passive -ilanishwa, and each of
these may conceivably be the starting-point of other similar

CH. IX VERB-SUFFIXES 57

series, or sets of series, and so on without limit. There is
literally no getting to the end in this direction of the potential
capabilities of Bantu expression, and that in respect of any of
the verb-stems in any stage of evolution from the monosyl-
lable. Verbs may be found again and again in many dialects
ending with such combinations as -aila, -aula, -auka, -ola, -oka,
-ulula, -ululula, -ulumula, -mbata, -mbatula, -ukula, -ukusula,
-olola, -ololoka, -olomula, -okomeka, -ongola, -nzula^ -famula,
-uhmga, -ndama, -mbala, -umata, -nyangula, -ngalala, -nyenta^
-nyanga, -manga, -nganta, &c. All these and many others on
examination appear to be merely agglomerations of formative
elements, surviving (for whatever cause) in connexion with (it
may be) particular roots, because hitting off some special
aspect of the idea in a highly suggestive or picturesque way,
and the curious result sometimes follows, that the formative
element seems to run away from the radical, embodying in
itself sufficient expression of an idea, so that the original base
or root-syllable becomes of little account, and may be repre-
sented by different letters, such as /, t, or k, without affecting
the meaning. To quote Scott's Dictionary once more a note
on the word mbuna ' Here again are seen the great features
of the language, namely (i) the conscious touch each word has
with its root-idea, and (2) the modal forms that breathe through
and mould these ideas. Verbs become but modal forms of
root-ideas ; and the most common combinations simply show,
by beautiful gradation to less and less common forms, that the
whole material of the language consists of certain nucleated ideas
in a scarce-breathed consonantal resource (setting ?), moulded
into speech by a rich and full modalisation. This modifi-
cation in the verb is final, in consonants (nouns ?) it is chiefly
initial.' (There are signs of lack of final revision in the text.)
Before leaving the subject of verb-suffixes, the final vowel
-a requires further remark. It has been already characterized
as the universal final sound of all true Bantu verbs with very
few exceptions. In Swahili the many verbs ending with other
vowels are of foreign, mostly Arabic, origin. Strictly, however,
the use of final -a is limited to those parts and forms of the
verb conveying an objective meaning, or fact as fact, e.g. in

58 WORD-GROWTH

the indicative, imperative and (so called) infinitive moods.
But when the meaning is subjective, and the idea of the root is
to be represented as conceived in the mind, as purpose, or con-
tingency, as thought of rather than realized, i. e. in the sub-
junctive (or subjective) mood, then final -a is changed to -e
very generally in Bantu. Final 4 is also found in some nega-
tive tenses of the verb in many dialects, and this may be
clearly attributed to the presence of a common and forcible
particle of negation, i or iai, or to si, both of which may
further be recognized in a common verb sia, used to imply
negation. In this case the negative particle first followed the
full verb-form, and afterwards was absorbed and took the
place of the final -a. Whether the final -e of the subjunctive
(subjective) mood represents the fusion of a and the negative
sign i, subjective idea being indicated by elements expressing
not objectively real, can only be at present matter of con-
jecture.

One other widely used verb-form, ending in -e, or -ile, will
be noticed in connexion with the subject of Tense (p. 62).

CHAPTER X

VERB-PREFIXES

So far verb-suffixes, or formative syllables following the
root, have been considered. These have been shown to convey
the distinctions described in grammars by the terms derivate
stems, voices, moods (in part), negative and affirmative con-
jugations (in part). Single and combined, they provide
material for a larger number of different fixed verb-stems
based on a monosyllabic root, and for a rich crop of derivative
verb-stems formed at pleasure from any verb-stem of suitable
meaning. All stems, fixed and derivative, may have active,
passive, and neuter forms, each with its full equipment of
conjugation, moods, and tenses. Drawn out in full gram-
matical detail, the luxuriant growth of forms in a single Bantu
verb can best recall once more language descriptive of tropical
vegetation, an equatorial forest of huge trunk-roots throwing
up stems, with many crossing and interlacing branches, the
wealth of leaves often almost concealing their supports and
hiding the trunk from view, while all parts contribute to
develop the final beauty of flower and fruit, the sounds of
common daily human talk.

This is, however, in some degree anticipation. Many a lan-
guage may in a measure admit of being illustrated by the
same metaphors of plant-life and growth. Yet in Bantu, life
seems peculiarly present and at work in the free moulding of
materials, and the wide choice open in several directions for
the conscious as well as instructive self-expression of the
human spirit, and this not only in verbs but in nouns, and by
use of prefixes as well as the suffixes so far described.

To pass, however, from the suffixes to the prefixes of the verb,
is to pass into a somewhat different atmosphere. Any one
familiar with a Bantu dialect will feel that there is a difference,
which may have its origin in a different stage or age of speech-
development, and that the prefix is the later of the two. The

60 WORD-GROWTH PT. in

genius of the languages seems to expend itself first on the
elaboration of the embodiment of the root-idea, and afterwards
to provide for its adaptation to the need of expressing its
relation to time and subject. It is as if the master-workman
had laboured at perfecting the main block of his weapon, its
forging, tempering, shape, and finish, and left the mechanism
of discharge, lock and stock, to other if not less able hands.
The suffixes of a verb-stem seem welded into a solid homo-
geneous, yet flexible whole, while the prefixes which represent
relation of tense and person are by comparison separate,
detachable, and interchangeable parts. This difference may be
illustrated by the fact, that any verb-stem once complete with
root and suffixes, one or more, but with final vowel a, admits at
once of being used as a noun as well as a verb. It is only
a matter of prefix. MM- placed before any verb-stem
connects the verb-idea with a person, Ku- forms with it the
infinitive mood, i.e. a general verbal noun. Thus M-penda
means (in Swahili) ' one who loves ', M-pendeya ' one who
pleases, a flatterer ', M-pendelewa ' a favourite ', while Ku-
penda means ' loving ', Kn-pendeya ' giving pleasure ', and so
on. On the other hand, other prefixes, used in a certain order,
adapt the verb-stems to express relation to time, subject, and
object, and these must now be briefly reviewed.

Their character of comparative separateness and detachability
as contrasted with the suffixes, is attested by the fact that
from the first in some South African dialects the verb-prefixes
have been written as separate and distinct monosyllables, not
as groups of monosyllables, much less as forming one word
with the verb-stem which they qualify as to tense and person.
Whether this system of spelling is better, in practice or theory,
than the other of writing each group of prefixes, root, and
suffixes as a single verb-form, is still an open question, but the
coexistence and extended use of the two cause much diffi-
culty in the study of Bantu.

Verb-prefixes are of two classes, one marking distinctions of
Tense, the other of Person, or more exactly of Subject and
Object. The initial prefix of every finite verb-form, i. e. in
the indicative and subjunctive mood, combined or not with

CH. x VERB-PREFIXES 61

a particle of negation, relates to the Subject. The tense-prefix,
single or compound, is strictly a pre-radical prefix, and, when
used, follows the subject-prefix. The object-prefix, when
used, follows the tense-prefix.

Tense-Prefixes

No exhaustive list of monosyllables used as tense-signs can
be given here. In the dialects here chiefly kept in view, most
of the monosyllables in the A column of Table B are so used,
and some others, viz.

A, Ba, CHi, Da, Fa, (n)Ga, Ka, Ma, Na, La (and Ra\
Sa, Ta, Ya, Za. Also Ku, Ki, Ko (for Ka-u), Me, Le, Li,
Lo (for La-u).

Combinations are also used, e.g. a-li, ka-na, nga-li, ka-ku,
ta-ku, fa-ku, za-ku.

It is, however, probably misleading to connect any of the
tense-prefixes directly with the interjectional monosyllables
which they resemble in form. It is more likely that they all
represent the later stage of actual words. Some are obviously
simple verbs of the monosyllabic type, such as li ' be ', fa ' do ',
ka ' go ', ma ' remain ', sa < come ', ta ' do ', ya ' go ', za ' come ',
lending themselves easily to subordinate or ' auxiliary ' use in
connexion with other verbs. Others may be of similar origin
or be residua of longer verb-stems worn down by use. Others,
again, are common as noun-prefixes, and recall the capacity of
verb-stems to be used with them as nouns, e.g. ki, chi, ku.

As tense-signs, it is difficult to attach definite meanings to
some of them, except within the limits of a particular dialect.
In fact, apart from mere difference in pronunciation and
phonetic disguise of identical words, there are few points in
which the individuality of a dialect as such more clearly
asserts itself than in the particular selection of t tense-signs
which it habitually employs, and the particular sense it
attaches to each. Few dialects seem to agree altogether in
this rather important point. Some, otherwise closely alike,
differ in this.

Again, the primary distinctions of time, past, present, and
future, so important in modern written languages, are by no

62 WORD-GROWTH PT. in

means clearly marked as a rule in Bantu. Swahili is a char-
acteristic exception, practical usefulness having limited its
tenses mostly to single and clear meanings. But in Bantu
generally distinctions are conveyed of a different kind, modal
rather than temporal, or modal as well as temporal, aspects of
the root-idea (whether act, state, or process) analogous to light
and shade, colour, perspective ; and such distinctions as simple
act or fact apart from time ; as general, particular, habitual ; as
continuous, momentary, or repeated ; as near or distant, past
or future ; as actual or conditional, and so on. With such
resources ' the swift and subtle instinct of the native narrators,
selecting, changing, and mixing tense with tense, unfolds to
a hearer alive to every tone, look, and gesture, a picture, or
rather panorama, which written language could hardly repro-
duce' (Wisa Handbook}. On the other hand, a simple
' Aorist ' tense serves to describe a past war or future campaign
equally well with a present battle, in terms of living speech.

By including all the different ways in which auxiliaries and
other elements may be combined, lists of sixty or seventy tenses
in some dialects have been compiled.

The general uniformity of structure in Bantu must not be
allowed to obscure the fact of exceptions. In at least one case
a tense distinction is widely expressed by a change in the final
syllable of the whole verb-form, final -a becoming -He, -ele,
or only -e. This form will be found discussed in grammars.

The person-prefixes of a verb serve to define its subject and
object. They effect concord or grammatical agreement
between them and the verb, and are syllables closely connected
with the class prefixes of nouns, and with pronouns. They
will therefore be mentioned again in a later chapter.

As outcome and recapitulation of what has been said, the
following general analysis of Bantu verb-forms may be given
here.

i. Prefixes

(a) defining subject, in respect of class, number, and person
(see next chapter).

(b) Tense-sign.

(c) defining object, when required.

CH. x - VERB-PREFIXES 63

2. Root, conveying the verb-idea.

3. Suffixes

(a) of stem, including fixed stems, i.e. new root-verbs,
and derivative stems.

(b) of Voice (passive, neuter).

(c) of Mood (indicative and subjunctive).
Take for example, (Swahili) idimwambia,

u-li-mw-A MB-i-a

u (subject-prefix) ' you ', li (prefix of past tense) ' did ', mw
(object-prefix) ' him ', amb (root) ' say ', i (suffix forming
applicative stem) ' to ', a (indicating active voice, indicative
mood), i. e. ' you said to him ' (' you-did-him-say-to ').

CHAPTER XI

NOUNS

NOUNS in Bantu are for the most part traceable, like verbs,
to single letters or syllables as their base or root, and, as in
verbs, these bases are in origin presumably interjectional, in the
sense already explained (Chapter V). Obviously, often the same
bases are the root of both verbs and nouns. Nouns, however,
actually monosyllabic in root are, like verbs, comparatively few,
and those mostly expressive of common objects, e. g. (m)tu
'a person', (m)ti 'a tree', (m)to 'a river', (li)so 'eye', (u)ta 'a bow'.
(The bracketed part does not belong to the root of the word.)
The majority of nouns as well as of verbs are found in the
next stage, combining two or more germ-elements, and
in later stages and fuller combinations the parallelism con-
tinues, nouns being formed at every stage from the developing
verb-stem.

Apart, however, from this practical identity as to base and
stem, the noun and verb are very clearly and definitely dis-
tinguished. To become a verb- word, a root has to be followed
by the sound -a, as its final vowel or suffix. To become a noun-
word, the root-syllable may end with any vowel, not necessarily
involving any new or fixed element in addition to the original
interjectional monosyllable, but it must be preceded in all cases
(with only apparent exceptions) by one of a limited number of
formative syllables as prefix. These syllables are the well-
known class-prefixes or classifiers, which figure so largely in all
accounts of the Bantu Family of Speech.

Except to notice that in nouns with monosyllabic roots, -a (the
final vowel of all verbs) is less common than -e, -i, -o, and -, the
final vowel of nouns may be passed over for the present. In
such nouns they do not seem to convey any special differences
of significance beyond that of the interjectional sound, as such.

CH. xi NOUNS 65

In noun-roots of a later stage, involving (as in verbs) other
formative elements, the terminal vowel or syllable does in some
cases carry a distinct aspect of the root-meaning, or a wholly
different one. Thus (in Swahili), m-pend-a ' one who loves ',
m-pen(d)-si ' one who is loved ', m-pend-o ' love ' ; or in Nyanja
m-pand-a 'a branch, or part', m-pand-e 'shell ornament',
m-pand-i ' a destitute person', m-pand-o ' a seat ', m-pand-u ' a front
tooth '. Sometimes agent and instrument (for instance) are thus
distinguished. The possibility of using five different final vowels,
multiplied by the great number of possible verb-stems springing
from a single root, and these again multiplied by the possi-
bility of using different classifiers with each of the many nouns
which can be formed from them, is a further striking illustration
of the great potential richness of Bantu speech. But this
subject must be left to be followed out in detail in the
various grammars.

Here the classifiers, or distinctive noun-prefixes, claim special*
attention, and first as to their general significance as a system.
Probably the systems of declension, and of division by gender,
in other languages would repay comparison. But the present
object is not comparison, but to characterize the Bantu system
as it now exists. To an individual of the Bantu-speaking race,
nothing becomes an intelligible object of thought without being
ipso facto conceived in relations or under limitations referring
it to a general class, and this for the most part instinctively
and (so to speak) automatically, but also in the case of novel
words and ideas by selective preference. The whole range of
knowable objects and ideas thus falling into classes, the
number of such classes might be expected to be large, whether
in respect of intrinsic natural differences of things or of possible
points of view from which they are regarded. As a matter of
fact, the number to a Bantu is seldom more than eight. The
simple syllabarium of Table B supplies plenty of material for
extending indefinitely the number of groups by the use of dis-
tinctive prefixes. On the other hand, a relatively small and
in many ways curiously limited selection is made from that
Table, sixteen or seventeen monosyllables out of the whole.

The fact may possibly point to some inherent weakness,

1214 E

66 WORD-GROWTH PT. in

deficiency or limitation, in the Bantu mental equipment, a
natural incapacity for appreciating phenomena in their infinite
variety, or to a peculiar faculty for grasping certain broad
aspects of phenomena, which are to the Bantu all-comprehen-
sive, and yet not as a fact recognized as dominating and crucial
by other races. It may stamp the Bantu mind as of a different,
not to say lower, quality as compared with others, or as possessed
of an instinctive gift, which may prove a clue to real discovery
in human nature. It is at any rate a striking and character-
istic peculiarity.

The nature of the Bantu categories themselves is by no
means clear, at least as to original or inherent meaning. The
system has nothing whatever to do with gender or distinctions
of sex, real or imagined. No Bantu prefix distinguishes sex
as such. A superficial similarity can at once be noticed between
the final vowels of Bantu nouns, -a, -e, -z, -o, -u, and the stem-
vowels which assign mensa, dominus, rex, &c., to their respective
declensions. But the Bantu vowel has no such definite function,
as has been stated above. Nor do the English noun-termina-
tions, -ness, -ship, -dom, and others suggest more than an ana-
logous classification. The prefix, not the suffix, is the all-
important classifying formative in Bantu nouns. But though
the systems of classification may have no common distinctive
features in formation, it is quite possible that the Bantu
classifiers may throw light on a principle underlying both. At
any rate the Bantu prefixes represent a relatively early stage
in language-growth. Their present forms are not likely in any
case to be the earliest of all, but many are singularly distinct
and seemingly well-preserved, and (under many limitations,
which involve separate and detailed study) convey still distinc-
tions of meaning, which are living ones, and involve actual if
instinctive selection for use at the present day. The same
noun-stem maybe referred to four or five classes, with a difference
greater or less of meaning. Various noun-stems may be
transferred from one class to another, and so be invested with
a new aspect. The same nouns may even be in different
classes in different dialects, and yet have the same meaning as
to objects denoted. But this by no means disproves the fact,

CH. xi NOUNS 67

that a general survey is still able to detect common distinctive
characteristics for the most part attaching to objects included
in a given class, more clearly in some than others, and that in
nouns as well as verbs the human spirit in Bantu has materials
which it is free to deal with, mould, and fit to the form which
its instinct or reason selects as best for its purpose of expression.
In this respect also Bantu is speech still alive.

E 2

CHAPTER XII
NOUN-PREFIXES

A PHENOMENON so characteristic and interesting as that of
the Bantu Class-prefixes of nouns seems to call for somewhat
full and detailed treatment, though at the same time much will
seem to any student of Bantu to be omitted or inadequately
represented.

The following are the prefix-syllables in forms found in
dialects of the Central Lakes region, with a few common
variants, given first in alphabetical order for comparison with
Table B, then in other arrangements according to use and
meaning :

In (N) Mi

Ka Mu (M)

Ki (CHi) Tu (Ti)

Ku U ( ifru, Bu)

Li (Di> I) Vi (Ft)

Lu Wa ( Wa, Ba, Va> A)

Ma Zi

Pa is sometimes added, but has a different use, and (with
Ku, Mti) is noticed in Chapter XV.

They may next be divided as Singular and Plural, the leading
forms only being given :

Singular. Ka Plural. Ma

Ki Wa

Li Mi

U Vi

Ku Zi

Lu Tu
Mu

In is used both as singular and plural.

NOUN-PREFIXES 69

The classes are formed on the basis of various combinations
of singular and plural prefixes which for the most part are as
follows :

Sing. Plur.
Class i. Mil Wa, e. g. mu-ntu ' person ', wa-ntu.

2. Mu Mi, e. g. mu-ji ' village ', mi-ji.

(In

3. In I y ' e. g. in-kuku ' fowl '.

4. Ki Vi, e. g. ki-ntu ' thing ', vi-ntu.

5. Ka Tu, e. g. ka-nwa ' mouth ', tu-nwa.

6. Li Ma, e. g. li-se ' hoe ', ma-se.

7. Lu \2i e- l u -k wa ' bark ', ma-kwa.

8. U Ma, e. g. u-ta ' bow ', ma-fa.

9. jfiTw .#ftf, e. g. ku-twi ' ear ', ma-twi.

The order and numbering of the Classes are arbitrary, and
they are conveniently distinguished as the Mu- Wa class,
Mii-Mi class, and so on.

It is hardly to be expected that the earliest forms of these
prefixes, still less the original source (interjectional or other) of
forms and meanings, should be within the reach of investigation,
present or future, and to suggest etymological theories seems
peculiarly futile till the bases of Bantu speech are more fully
understood. But a certain amount of light may be thrown on
both meanings and forms.

As to form, the marked Bantu predilection for the nasals
M, N, and the vowels most closely connected with them, U, /,
seems reflected in the class-prefixes, and indicates a high value
attached to them as vehicles of expression. A, I, U are the
only vowels used in them. Mu, Mi, Ma, In, Ki (Chi), and U
are the most stable, widely used, and perhaps best-preserved ;
Ka, Tu, and those including the weaker consonants L, V, Z,
i. e. Li, Vi, Zi, less so. Ku is used with few except verbal
nouns, i. e. as a sign of the infinitive mood. The use of Ma
and Zi as plural prefixes of several classes (excepting the two
first) points to a decline from an earlier standard of distinct-
ness, as between class and class.

The prefixes are found in both longer and (in some cases)
shorter forms. Each prefix is used in some dialects in a di-

70 WORD-GROWTH PT. in

syllabic form, its own vowel being placed before it, e. g. u-mu,
a-wa, t-mi, i-ki, i-li, u-lu, a-ka, u-u, &c., a change which may
well represent an earlier actual reduplication of the prefix,
i. e. mu-mu, wa-wa, mi-mi, &c. The fuller form seems to add
a faint demonstrative meaning, like that of English ' a, the ',
in addition to its use as defining class. It will be shown later
that the class-prefixes supply the base of the pronominal
demonstratives generally, and the full reduplicated forms,
mu-mu, u-u, ka-ka, li-li, &c., are actually used as such.

Some of the class-prefixes are also found in a shorter form.
In general, the class-prefix of a noun is also the prefix of every
adjective, pronominal or other, and of every verb-form by
which it is qualified. It is in this repetition or reflection of
the noun-prefix as the prefix of parts of speech grammatically
related to the noun, that the characteristic Bantu system of
concord consists. In most cases, the noun-prefix and the
concord-prefix are the same. But where a nasal is used in
the noun-prefix, it disappears in the concord-prefix. The
corresponding prefixes are therefore :

Class-prefix Concord-prefix

mu u (with some exceptions)

mi i

ma a (or ya}

in i (sing.), si (plur.)

The loss of so important an element as the nasal is remark-
able, but may be due to gradual loss of living significance in
the sound itself and then to phonetic decay.

II

As to the distinctive meaning, at present if not always
attached to the various class-prefixes, certain clear lines may
be drawn. But a review of words found (in one dialect or