This is a reproduction of a library book that was digitized

by Google as part of an ongoing effort to preserve the

information in books and make it universally accessible.

https://books.google.com
 This is an authorized facsimile
printed by microfilm /xerography on acid - free paper
in 1982 by
UNIVERSITY MICROFILMS INTERNATIONAL
Ann Arbor, Michigan, U.S.A.
London , England
 SEMANTICS :
STUDIES IN THE SCIENCE OF MEANING

e l f r e d
da
BY
,
MICHEL BRÉAL
PROFESSOR OF COMPARATIVE GRAMMXR AT THE COLLEGE DE FRANCE

TRANSLATED BY

MRS. HENRY CUST

WITH A PREPACE BY

J. P. POSTGATE
PROFESSOR OF COMPARATIVE PHILOLOGY AT UNIVERSITY COLLEGL LONDON

NEW YORK

HENRY HOLT & COMPANY

MDCCCC
 Urry
il .) 67.
y, Snient List
Breil, Michel Jile's life ,1: 0.9.1915.
Semantics: studies in the science of reaning, by Michel
i Breul ... Tr. by Mrs. Henry ( 'uit, wi" !: o prica by J. P.
l'ostgate ... New York , 11.10lt i comny , 1.900 ,
Iw1, 341 3. 20fr*.
51220 €
G.2127
12-17
.
9--12183
I.llornego ol ( MNGITAN 107.11.118
1112041 , NIN .
 SEMANTICS :
STUDIES IN THE SCIENCE OF MEANING
Undergraduate
Library

Р
741
.8833

PRINTED IN ENGLAND .

This Edition is for sale in the United

States of America only, and is 1
not to be imported into countries
signatory to the Berno Treaty.
 Ondergraduate
Library

sition scatt
12-12

TRANSLATOR'S NOTE

I WISH to express my most sincere thanks to Mr.

Charles Whibley for his constant advice and assistance .
I am also indebted to Professor Postgate for kindly
quyuç

reading through the proof-sheets.

NINA CUST.
90-+:
l
 PREFACE
:

THE illustrious author of the Essai de Sémantique

needs no introduction to the politer circles of ours or any
other cultivated community. But the same cannot be
said — and a certain feeling of humiliation must attend
the acknowledgment of the subject to which he has
made his brilliant contribution. That is in no position
to disdain the humblest efforts of the most insignificant
contributor.
My interest in these studies is now little short of a
.

quarter of a century old : it goes back to the year 1877,

when I chose the science of meaning for the subject of
a “ Fellowship dissertation " at Trinity College. The
investigation proved to be embarrassed by difficulties,
the chief of which was the dearth of suitable materials,
that the ardour of youth had not foreseen ; results were
not commensurate with hopes, and its prosecution had
almost perforce to be abandoned for studies which
promised a more immediate outcome. But the old
desire still pricked in the memory, and at last, in an
inaugural address at University College, London, I had
the opportunity to call the attention of such as would
 1

viii PREFACE

listen to the claims of a subject whose investigation was

now more possible and very much more urgent."
Of that address this introduction may be called the
immediate offspring. Its aims are primarily, to press
upon English and English -speaking students the instant
claims of this branch of linguistic inquiry; secondarily and
incidentally, to offer to future inquirers such help, whether
in the shape of negative criticism , or it may be of positive
suggestions, as space and circumstance may allow . What
I say will, I trust, be of general application, but if I keep
an eye on the requirements of the students of the two
classical languages, let this not be ascribed altogether
to that natural bias which leads us to draw by preference
from topics which fall within our daily observation .
The exponents of these languages have still the chief
control of our literary culture and linguistic investigation.
This supremacy is not unchallenged, and it may be that
it will fall ; but while it lasts it is within the mark of
equity to say that so great an influence demands from
its possessors an equal enlightenment. I will conclude
this personal explanation with a brief reference to a .

point on which I am particularly anxious to be clear.

In the development of all sciences general statements
are necessarily employed before they are in strictness
justified. The convenience of provisional formulæ is
great as an anchor for our ideas and a mark for our
critics. Their provisional character must never be left
1 This address, which is reprinted at the end of the present volume,
was delivered in October 1896. M. Bréal's Essai de Sémantique
appeared in July 1897.
 PREFACE ix

out of sight ; and so whenever in the following pages I

may assert that such and such is the case, this is to be
understood as the equivalent of a statement that such
and such are the hypotheses that appear best to agree
with the relevant facts which have come within my own
observation .
There is no need to go far for proofs that the import
of the forms of speech is had in insufficient Confusion of
regard. The venerable theory that gram- gender
sex ,
and
matical gender was properly a mark of sex is
still widely held both here and on the Continent, and
not unnaturally applied to the languages of Greece and
Rome. Abroad , however, the practical influence of this
theory is counteracted by the silent negative which the
languages of the Continent present to its application.
But in England and in America the only classification
of nouns recognized by language is that according to
sex, and with this therefore is gender identified .
Mark Twain may serve to show the popular concep
tion . In his burlesque of the German genders he trades
" How blindly the sexual association clings to the English suffix
-ess is exemplified in the naïve definition quoted by Prof. B. I.
Wheeler, “ Origin of Grammatical Gender,” Journal of Germanic
Philology , vol. ii. p. 542, “ A fort is a place to keep men in, a
fortress to keep women in 1" Victor Hugo's biting chanson in the
Châtiments, p. 384 (Paris, 1882), furnishes an odd parallel if, as
seems not unl.kely, the French suffix suggested the comparison
there. Speaking of the first Napoleon, he says
" Berlin , Vienne, étaient ses maîtresses ;
Il les forçait,
Leste, et prenant les forteresses
Par le corset. "
X PREFACE

upon the assumption that gender is meaningless unless

it signifies sex . " In German a young lady has no sex ,
while a turnip has. Think what overwrought reverence
that shows for the turnip , and what callous disrespect for
the girl. See how it looks in print. I translate this
from a conversation in one of the best German Sunday
school books
Gretchen . Wilhelm , where is the turnip ?
Wilhelm . She has gone to the kitchen.
Gretchen. Where is the accomplished and beautiful
English maiden ?
Wilhelm . It has gone to the opera.”
We need not be surprised if this jesting is incompre
hensible to a German, but we may pity the straits to
which the jester is reduced who cannot expose the
uselessness of the gender distinctions without making
them out to be something which they are not. Let us
turn from a conscious humourist to unconscious ones.
Adjectives, a well- known grammarian tells us, in
language at the best misleading, should be regarded as
substantives of wide general application, e. g. bonus " a
good he," bona “ a good she," bonum " a good thing."
But decem aurcos tibi misi docs not mean “ I have sent
you ten gold mcn ," but " I have sent you ten gold
pieces " ; nor does ecce, duae longae mean “Look, two
long women !” but “ Look, two long galleys !"
Aristophanes in his Birds, 1313 sq., makes the Chorus
say that soon some will be calling Cloud -Cuckoo -town
πολυάνωρ : ταχύ δή πολυάνορα την πόλιν καλεί τις ανθρωπων .
A Tramp Abroad, Appendix D, p. 543 (ed. 1880 ).
out of s
may as
underst
and suc
with the
observa
There
of the
regard
matical
still wid
not unna
Rome.
theory is
language
But in
of nouns
sex, and
Mark 1
tion. In

1 How bli
-ess is exem
Wheeler, "
Philology, vo
fortress to kee
Châtiments , pe
seems not un
there. Speakin
xii PREFACE

-os," he is holding up to derision the theory that the

function of this inflexion was to denote a male. The
methods of the ancient and modern humourist are thus
seen to afford a curious contrast. Both, it is true,
scout an unfamiliar conception by an appeal to popular
feeling ; but it is just what the American thinks self
evident that the Athenian thinks absurd .
The Roman view was similar. Even in the case of
animals the scxualizing, if I may so call it, of the endings
-us and -a was by no means complete. In old Latin
lupus was a wolf ; lupus femina (Ennius) " a she-wolf" ;
lupa " a wolf-woman," i.e. a prostitute. Lupa in the
sense of a “ she -wolf ” was an innovation of the Augustan
times. No doubt porcus was often used for a male
and porca for a female swine ; but this difference was not
rooted in the terminations. Otherwise Virgil could not
have used porca of the male, which was always sacrificed
when a treaty was struck.?
· The stereotyped appellations masculine and feminine must be
avoided if we would reproduce the meaning of the original. And
similarly in ν. 683 [τών ονομάτων] αττ ' άρρεν) εστιν άττα δ'αυτών θήλεα,
i.e. “ which proper names are males' and which are females'? "
· Aen ., 8. 641 , " caesa iungebant foedera porca ," with Quintilian's
very significant comment (8. 3, 19), “ quaedam non tam ratione
--- porca : fecit elegans fictio
quam sensu iudicantur ut illud caesa-
nominis ; quod si porcus fuisset, uile erat." Quintilian's meaning
is that Virgil, desiring to avoid the hackneyed porcus, invented the
new term porca, which his readers were to understand in the
same sense, their perception guiding them to the right meaning.
Compare Hor., Carm ., 3. 23, 4, auida porca. Statius, Siluae,
2. I , 9, uses orbati leones for lionesses reft of their cubs ; and
Valerius Flaccus, 6. 347, has leo even in the singular for a
lioness. And assuredly Homer felt no such difficulty as his
 PREFACE xü

It is a still idler task to seek for these distinctions outside

the nomenclature of the animal kingdom . At Catullus, 62.
54 , " at si forte eadem est ulmo coniuncta marita ," the
MS. authority is divided between marito and marita, and
more than one of the editors who read marito have
coquetted with the fancy that the elm, being the
husband, should have the nobler gender. The philolo
gist will, of course, allow that Catullus may have written
marito (which would be a substantive, " as a husband"),
becausc its inflexion agreed with the inflexion of ulmo
and the assonance so produced amused himself and his
readers with a pleasing impression of concinnity. But the
notion that his choice was determined by an association
of sex and gender the phrase of Quintilian, “ maritam
ulmum" ( Inst., 8. 3, 8), is enough to disprove. The dul
ness, if we like to call it so , of the Roman feeling in this
matter is shown by Manilius, 5. 238, “ et te, Bacche,
tuasi nubentem iunget ad ulmos,” when the writer, in
wedding the vine to the elm, never troubles himself over
the circumstance that his synonym for the bride is the
name of a male god.
But we need not wonder at the vitality of the error in
students and teachers of literature when even professed

commentators have raised on Il. 21. 483 (addressed to Artemis ),

drel adorta yuvaiflv | Zeùs Onkev ; or on 17. 133, where a new
(masc.) ights περί οίσι τέκεσσι ; or 18. 318, where a λις ήυγένειος
(masc.) pursues the hunter who has robbed her. The scholiast in
A notes here το δε λέαινα νεώτερον όνομα..
· The better manuscripts here have tuos. And of course ulmus,
like some other names of trees, may have occasionally been
masculine. But nothing could be more futile than to endeavour
to make the gender significant in a passage like the present.
xiv PREFACE

philologists are still under its spell. Prof. B. I. Wheeler,

quem honoris causa nomino, in the article which I have
already cited, while recognizing the strength of Brug
mann's position , - " that the formal gender in our Indo
European languages for thousands of years was not con
nected with the idea of the masculine or feminine, is shown
by quite unmistakable evidence," 1 - proposes (with Jacobi)
to explain the growth of the feminine termination (-ă, a,
-n) as duc to the influence of the " shc- form " (så = Gk. )
of the pronoun. " Thc --i forin was introduced into the
adjectives ( verbal noun -adjcctives) of the os - ending to aid
the precision of denotation when an object of female sex
was referred to by such noun -adjective ; thus să leugós
yielded to så leuqá or leuga " (i devrós to i, devký or
Aevxh ). The assumption of the attracting iwwer of the
pronoun (article) is a legitimate one, as we may see
from the passage of Aristophanes already quoted, where
the innovator in language proposch to act liv kápooTTON
right by assimilating the noun to thc article, not vice versa ;
though it may be doubted how far we are permitted to
carry it back. But there is not the slightest evidence
that I can see for assuming that să (or sī) meant either
she alone or she predominant, not she and it in
differently, as the German sie docs now, and the English
she did once, and if so, the assignment of certain
terminations to express sex is still uncxplained .'
Brugmann, On the Nature and Origin of Noun genders in the
Indo-Eurupean Languages, I quote from the English translation .
New York, 1897, p. II .
• This is not the place to discuss the whole complicated question.
But I may add that it is against Wheeler's theory that the assimi.
lation began with verbal noun -adjectives (his example Neuads
 PREFACE ху

I will conclude with some illustrations from modern

composition. In an elegant version of a poem of Shelley
the lines

“ No sister flower would be forgiven

If it disdained its brother,"
are translated
“ Num rosa silvicolis tibi dis placitura videtur,
Contemnat fratrem quac soror asphodelum ? "
The point of selecting flowers of different genders to
harmonize with the different sex of soror and fratrem is
easily caught by the modern reader ; but unless I am
altogether mistaken, it would have been quite lost on
an ancient. In another modern rendering that I have
sccn, a disconsolatc maiden 19 made, in accordance with
modern sentiment, to bewall her woes to a turda. But
a hen thrush would have suggested nothing to a Roman
unless, it may be, the kitchen ! 1'crsius, 6. 24, " tenues
turdarum nossc salluas," 1 of the cpicure whonc delicate
does not seem to be a very good one, as there is nothing very
distinctively verbal about it) that the most characteristic class of
verbal adjectives, e.g. touds, resisted the assimilation down to
classical times. The explanation of the origin of the neuter gender
as developed out of the object accusative seems, on the other hand ,
very plausible. Its acceptance would remove a difficulty out of
the way of supporters of the new thcory of gender propounded by
Dr. J. G. Frazer in the Fortnightly Review of January 1900 ( pp.
78 999.), viz. that masculine and feminino names of things go
back to a time when langunge was modified according as the
speaker was a man or a woman.
1 It does not follow that these versions are necessarily faulty
because they are not true to the classic norm . The modern who
aims at making his version conform in all respects to ancient
canons sets himself an impossible task.
o
xvi PREFACE

palate can distinguish between the flavour of hen- and

cock -thrushes.
Yet another question is raised by the passage of the
Personifica Birds which we have already discussed . TÓALS
tions and sex.Didávwp stands for no person but for a person
ification ; and the question arises, what is the composition
of these make -believes, and how far does the notion of
sex enter into it ? For the ancients I think the answer
must be " but siightly,” even to the crude materialism of
the Roman. “ Mens bona, si qua dea es tua me in
sacraria dono ," writes Propertius of one of these hypo
statized abstractions, and the doubt is significant. When
the Fetial began his solemn formula of reclamation ,
" audi Iuppiter ! audite fines (cuiuscumque gentis sunt
nominat) ! audiat fas !” ( Livy, i. 32, 6), the personality
in " righteousness" is clearly appealed to in the adjur
ation ; but all beyond is undefined. Whether these
beings clothe themselves to the consciousness of language
with attributes more specifically human depends not so
much upon language as upon art. “ Segnius irritant
animum demissa per aurem | quam quae sunt oculis
subiecta fidelibus."
Our subject has now brought us close to the confines
Meaning of what is perhaps the most bewildering depart
and
mytho ment of human inquiry-the chaotic realm of
logy.
mythology. The mythologist may have, as he
should have, at his command all the resources of archæo
logy, history, and geography ; but he will often go sadly
wrong unless he also holds the linguistic clue, and can
penetrate with clear vision to the facts that lie beyond
the dædal fabric of speech. That a strict and rigorous
 PREFACE xvii

investigation of literary and linguistic data may throw

light upon what otherwise would be tangled and obscure,
may perhaps be shown by the investigation of a single
point in the myth of Cerberus.
It is well known that on the number of this creature's
heads there is no agreement among authorities. The
In general he is allotted a comparatively modest Cerberus.
headsof
number, varying from one, as on an ancient
Argive bowl figured in 0. Immisch's article in Roscher's
Lexikon, p. 1121 , to two, as on old Attic vases, and
three, his normal allowance both in literature and art.
But, on the other hand, we find him with the monstrous
totals of fifty ( Hesiod) and a hundred (Pindar). This
contradiction becomes acute when we find the same
writer apparently representing him now with three heads
and now with a hundred . In our texts of Horace we
read at Carm ., III. II , 15 sqq.
“ Cessit immanis tibi blandienti
Ianitor aulae
Cerberus, quamuis furiale centum
Muniant angues caput eius atque
Spiritus taeter saniesque manet
Ore trilingui.” 1
but at Carm ., II . 13, 34
“ Demittit atras belua centiceps
Aures et intorti capillis
Eumenidum recreantur angues.”
The confrontation of these two passages raises at once
· The words eius atque are corrupt, as the appearance of eius
in an Horatian lyric shows ; and Bentley proposed for them
exeatque. The real correction is not certain, but this does not
matter to our present argument.
xviii PREFACE

a question whose resolution belongs to the science of

meaning. Could a well- known mythological figure like
that of Cerberus have presented itself to the conception
of the same people, now with three and now with a
hundred heads ? Is such a fluctuation conceptually
possible ? The difficulty is only evaded by pronouncing
the Sapphic stanza to be spurious, as Immisch, following
many Horatian scholars, does. The hundred -headedness
of the Hound of Hell is too widely certified to be cut
away by the critical knife. So other expedients are tried.
The testimony of Hesiod, Theogony, 311, Képßepov
ώμη στην, 'Αίδεω κύνα, χαλκεόφωνον | πεντηκοντακάρηνον, is
whittled away by the explanation, that the number
ought not to be pressed , ουκ αριθμητικώς αλλ' αντί του
TOAUKépalos, which is quoted from a comment of a
scholiast at Pindar, Pyth. 1 , 31 , upon the hundred heads
of Typhoeus. The expression is asserted to be no repre
sentation of the popular belief and ascribed to poetical
licence . Poetical perversity, we ought rather to call it,
which takes a popular concept and confounds it by
turning three into fifty.”
1 " By ' fifty -headed ,' which perhaps expresses an indefinite
number, later poets gave the epithet oplæpavos' ” (Paley on the
above cited place of Hesiod's Theogony).— “ Ein dichterischerAus
druck - keine Vorstellung des Volksglaubens, was für diese älteste
Stelle namentlicher Erwähnung des K. wohl zu bewahren und schon
hier scharf zu betonen ist. ” ( Immisch, Roscher's Lexikon, 1.c.)
• It must be pointed out that Hesiod's words assume that a
“ fifty -headed ” Cerberus was no stranger to his audience ; other.
wise he would have expressed himself in this sense- " You think,
my friends, that Cerberus has three heads, but I must tell you that
he has fifty.” As it is conceivable an objector might say that he
 PREFACE

investigation of literary and linguistic data

light upon what otherwise would be tangled
may perhaps be shown by the investigatio
point in the myth of Cerberus.
It is well known that on the number of
heads there is no agreement among autho
In general he is allotted a comparatively
number, varying from one, as on an
Argive bowl figured in 0. Immisch's artic
Lexikon , p. 1121 , to two, as on old A
three, his normal allowance both in lite
But, on the other hand, we find him wid
totals of fifty (Hesiod) and a hundred
contradiction becomes acute when we
writer apparently representing him now
and now with a hundred. In our text
read at Carm ., III. II , 15 sqq.
“ Cessit immanis tibi blandienti
Ianitor aulae
Cerberus, quamuis furiale centum
Muniant angues caput eius atque
Spiritus taeter saniesque manet
Ore trilingui.” 1
but at Carm ., II . 13, 34–
“ Demittit atras belua centiceps
Aures et intorti capillis
Eumenidum recreantur angues ."
The confrontation of these two passages rai
1 The words eius atque are corrupt, as the appear
in an Horatian lyric shows ; and Bentley propos
exeatque. The real correction is not certain, but the
matter to our present argument.
mecaneng.Cincidua well-instal
afofitehsasamensopie, now e3

imestice ? The dialy க்கை

maniritasian scholars,desde
ofte hand ofHellis too
22EUDU critical knife. So

μπτη, Αίεεω πίπα , χαλαιά

winteed aaway by the
abright not to be pressed
montados, which is
stohtohastatPindar, Pyti
fypneeus. Theexpres
sentation of the popular
bieenee . Poetical per
with takes a pob Herc. Fr. 1390
turning those into
By fifty-heade
number, later poets
above cited place of
Druck - keine Vors
Stellenamentliche Detoo manyainstoe .
hierscharf zu bele refore all my stanas
It must be
fifty -headed
wisehewould
myfriends, the
he has fifty."
 PREFACE xix

Previous discussions have been lax enough, it is true,

but still they have recognized, albeit dimly, the real
issue " Is the numerical discrepancy essential or un
essential ? " And to this we will accordingly now
address ourselves.
There is no gulf fixed between mythological concepts
and other concepts. My conception of any serpent which
I have never seen but believe to exist, does not differ
fundamentally from my conception of the sea -serpent
in whose existence I disbelieve completely. Now to all
concepts there is no quality so vital as that of clearness
or vividness of conception. For this it is necessary that
there should be no uncertainty about its essential parts.
I shall accordingly start with the two propositions that
to the ancients the dog Cerberus was a clear concept,
and that of this the dog's head (or heads) was an
essential part.
Of these two propositions no proof need be furnished.
But it may be urged that the number of heads was un
essential, in other words, that the dog might be con
ceived now with three heads, now with fifty or a hundred,
without injury to the clearness of the general conception.
Now if the Greeks, like some savage tribes, had been
unable to count above two, and regarded all beyond
as confused plurality, the position would be tenable. A
concept which thus varied between three and a hundred,
would not indeed be a clear concept, but it would be as
clear as the circumstances allowed . But this is not so .
does, and this is why merrnkortardprvov appears at the beginning of
the line, it may be rejoined that it could occupy no other position.
XX PREFACE

Between three and a hundred or fifty there was for the

ancients, as for us, a conceptual break . Accordingly I
infer that under these different statements lies more
than a mere numerical discrepancy. There is something
further to consider. Cerberus was a very terrible monster.
So is a crcature with threc dog's heads ; therc is a head
to menace you at every turn . But increase their number,
and he loses his terrors and becomes simply foolish and
ineffective. We should pause then before we make
Horace import into a solemn picture of the world below
the figure of a quadruped with a bush of heads
simultaneously dropping two hundred cars.
The monstrous beings to whom a multitude of heads
is assigned appear to be the following - Typhoeus (one
hundred according to Hesiod and Pindar) ; the giants
Ægæon ( Briareus), Gyas, and Cottos ( fifty heads and
one hundred hands each according to Hesiod ); Scylla
(twelve according to Homer) ; Hydra (nine according to
Alcæus, fifty according to Simonides, Palæphatus and
Virgil (Aen ., 6.576 ),one hundred according to Diodorus,
Ovid, Euripides (Herc. Fur., 1190 ), and others, while,
according to Euripides, ib. 419, it is uupiókpavos). Here
too there is obvious numerical discrepancy, but it is un
essential. The numbers given exceed the limit of ready
visualizing, they all include too many units to be counted
at a glance ; and therefore all may stand as signs of
indefinite multitude.
Now what have these monsters in common ? It is
their snaky character. Not only Hydra, who has “ a
hundred snaky leads” (Eur., Phoen . 1135 , with the
with free II si ze u rere any ani
textie seas tre Sonia vel mers se

tre dans vos a uns so I tas tavo corg

niza Icess and mus * in
misica 12 nary - isiens vas ei va
Sta . CCESS Hur mere İ HT AT RSU
at Is Scarei V is a mini sef
Incze L * I . Isong 12 TARDUS
nuar, zis is at I was a minor of the sia
Sie am fie as veret taiirne vien
tee sise - ergen Vi vas diei.
IIS SAAT HE S stovi v mure ta ze
nőiczech i Caes Ses sie zu and
Esida Sectos serene HEST
By Caimacius 37 , L. stas re mercere in
cenneron va him in mere state age 12
maa3 si erdeice pa veg . DU mirada
cecacies, ac in re se nec aš cie AS
: . , Listeccanes compares totes
in iz te ec ict cac maen :sns dut but
ceste , ur. UT IS TSus haarses rsstus ar
iuy is mirat ! Juur- slevene i Radial
Cung Lus it IP , Karm is estở kstaki và sau
மது பேரம்பாp EDLtr turpis
xujalons include repús. kan arte cuced that
te escean is ccccaring as victry ever Cece to
Hercies caçtıre of Cerberus Lotusthe
xxii PREFACE

standing epithet of the dog ). But assurance is riveted by

Knights, 1030, compared with 1017 (and also Peace, 312 ),
where Cleon is called explicitly κύνα Κέρβερον. Round
the head of this metaphorical Cerberus play a hundred
heads of cursed parasites. The verb chosen, édexuôvro,
makes the picture unmistakable : it is the one most proper
of the darting and flickering of the serpent's tongue.?
This, then, is the solution of the mystery, and Cerberus
has his dog's head or heads never reckoned at more
than three, and his snake's heads, which are too
numerous to be counted at a glance. The “ belua cen
ticeps " of Horace, Carm . 2. 13, 34, is the same Cerberus
whose " furiale caput" " centum muniant angues " in
the supposed interpolation of Carin . 2. II . The very
epithet atras suggests the snake, and as if to leave
no loophole for doubt, there is a final touch to the
picture, " Eumenidum recreantur angues." In the sus
pected passage the triple dog's head appears ; the
monster's mouths are open, and show the saliva drip
ping from all his three tongues. Horace is in perfect
accord with his contemporary, Virgil, Aen., 6. 417,
"Cerberus hinc ingens latratu regna trifauci | per
sonat,” the three open barking throats with (later on)
the erected snaky heads, " horrere uidens iam colla

! | translate the singular kepalhr. But we must not press it to

mean that the coinedian conceived of Cerberus as single-headed.
See Seneca, Herc. (quoted below ), where trina capita is immediately
followed by sordidum caput. The head of Cerberus is whatever
his ruff or collar of snakes surrounds.
• Bentley proposed gdwooal here and in Peace, 756, where the
passage is repeated. But no change is needed.
 PREFACE xxiji

colubris." And to turn to art, on the Argive bowl already

mentioned is a graphic representation of the monster,
where he is represented as a dog with active snaky
heads and necks erected from various parts of his body.”
Some of my readers may not unnaturally desire to
get rid of the discrepancy about the number of the dog's
heads as well. The single and the double head seem to
be found only in art, and so lie outside our immediate
" I cannot claim for my interpretation of the second place in
Horace that it is new, though I hit upon it independently. It is, in
fact, as old as the Scholiasts, Acron and Porphyrius, or Belun centi.
ceps (i. 13, 34), “ Cerberum dixit propter multitudinem anguium qui
in capite eius eminebant," and has been advocated more or less
timidly by one or two scholars. But this very fact was another
reason for a re -examination of the case, as it showed how little
weight has evidence when it has to combat false theory and erring
method. I subjoin some further references' to passages in ancient
literature which vouch for the conception of Cerberus which I
have been maintaining. Lygdamus (Tibullus, iii.), 4, 87 sq.,
nec canis anguina redimitus terga caterua | cui tres sunt linguae
tergeminunque caput." (Lygdamus, who frequently imitates
Horace, is very probably expanding him here.) Seneca, Herc., 787
sqq. , “ Stygius canis ' qui trina uasto capita (latera, R. Peiper )
concutiens sono | regnum tuetur, sordidum tabo caput | lambunt
colubrae, uiperis horrent iubae, | -ecce latratu graui | loca muta
terret, sibilat totos minax | serpens per armos.” Val. Fl., 3. 227 sq.,
" ast illum fluviis et nocte remersum (so I would read for remensa )
| Eumenidum canis et sparsae iuba reppulit hydrae ” ( i.e. Cerberus
with his hydra -like mane of snakes).
* The hundred serpent heads may be found in another Augustan
poet, if an ingenious conjecture of A. Palmer's be right. In Tibullus,
i. 3, 71 sq ., he proposed " tunc niger in porta per centum Cerberus
ora | stridet et aeratas excubat ante fores” for “ serpentum — ore."
Stridet, “ hisses,” does not require the addition of serpentum to
make its reference clear : cf. Sil. It., 6. 1976, “ tempestas oritur,
mixtam stridore procellam | Cerbereo torquens,” “ with the hissing
of Cerberus's snakes.”
xxiv PREFACE

province. Immisch says that "considering the difficulty

of the representation for primitive art we must not
consider the fluctuation decisive, but that to ascribe it
only to technical grounds is impossible." A difference
in the representation of Cerberus at different periods or
in different schools of art of course involves no self
contradiction in the conception, and the double head
seems to belong to old Attic, and the triple head
( according to a conjecture of Furtwängler) to old Ionic
art. But still there is something to account for. Let us
hazard a conjecture. May not the double head of the
KÓW "Aidov, the ianitor Orci, whose duty it was to keep
the thrcshold of the lower world from being trespassed on
from either side, bc compared to the double aspect of the
god of the doorway, upon which his triple head is a later
refinement to symbolize the tplodos or forking of the ways,
the one leading to Elysium and the other to Tartaros ? 1
Before leaving the subject of mythology, I should like
to cite another case in which the Greek mythologer's
treatment has been much too lax, the alleged confusion
of the Water -nymphs and the Tree -nymphs, which I
have discussed in American Journal of Philology, xvii.
pp. 30-44, xviii. pp. 75-6.
The study of ineaning will revise not a few of our
Science grammatical notions ; and in no department
ofMean
ing and
of linguistics perhaps is this revision more needed .
formal Hore, it is true, thc logician has lost much of his
grammar .
former power ; but the influencc of the pedagoguc
" Plato, Gorgias, p. 524 A, dy aſi Tıbben it ha pépetov tà dBus in Medy
εις μακάρων νήσους ή δ' εις Τάρταρον .
 PREFACE XXV

is almost wholly unimpaired. Nothing is so hard to

shake off as the old associations of the school-room ;
and pedagogy, not content with its early dominance,
invades the territory of science, whose discoveries it seeks
to discredit on the ground that they cannot be taught 11
nil sine magno uita labore dedit mortalibus, and it is
unfortunately true that what is most troublesome to
teach is often most instructive to learn . The practice of
acquiring a foreign tongue through the medium of one's
native one is no doubt a convenient one, but it leaves its
mark on all after study. Because que, te are usually
translated and in English, a very distinguished scholar
writes that " accuracy of scholarship is checked from
the outset when a boy turns up his dictionary and finds
one of the meanings given for que is or.” It would
seem, however, that real and intelligent appreciation
of the ancient language is just beginning when a
boy discovers that que cannot be translated mechanic
ally by and, and that a Roman who was contem
plating two alternatives was, unlike ourselves, free to
vicw them either simultaneously ( que) or successively
(ue). We may dislike the Latin use as shocking our
" I have elsewhere protested against this noxious fallacy
(Classical Review , 1899, p. 68). So I will only say here, better teach
nothing than what is not true. But we are not reduced to this
dilemma. The practice may be taught and the theory reserved.
Our classical teaching concerns itself too little with the living plant
and too much with the desiccatcd specimen .
' I refer to the use of que in passages like Virgil, Georgi, 3. 120
(of a horse), “ quamuis saepe fuga uersos ille egerit hostes et
patriam Epirum referat fortisque Mycenas,” the horse could not
have been born both in Epiros and Mycenæ. For some inscrutable
XXVI PREFACE

native ineradicable notions of or and and; but to refer

it to the English translation is the merest irrelevance.
The greater simplicity of English syntax, which is
often only another way of saying its greater ambiguity,
is well known to all Greek and Latin students. But
occasionally even English has discriminations which
they cannot render. The distinction, apparently doomed
to disappear, between “ if it is necessary " and " if it be
necessary ” is such a one ; and it is noticeable how often
it proves a stumbling-block to student and commentator.
Few can have failed to observe how potently these
imperfect abstractions, which we call grammatical rules,
work in the brain. And as if this were not enougli, we
must needs corjure up the sheeted phantoms of normal
standard and of purity of style to dull still further our
appreciation of the living speech. It is almost painful
to read in the pages of an eminent grammarian how
often the usages of Livy and Sallust are peu correctes. A
reason this particular cavil is regarded in more than one quarter as a
mark of high scholarship. On a similar use of nal compare Professor
W. M. Ramsay's protest in the Classical Review , 1898, p. 337 .
' It is, in fact, nothing but an instance of the popular view of
language cropping up to the surface. How ineradicable is the
notion that there is no finality or authority in the expressions of
ideas outside those of our mother-tongue, has never been more
clearly shown than by the anecdotes in F. Polle's entertaining little
book , Wie denkt das Volk über die Sprache, ch. ii., e.g. the memor
able utterance of a German from among the common folk_“ Im
Deutschen heisst das Brot Brot und ist auch Brot ; im Französi
schen heisst es pain und ist auch weiter nichts als Brot. " We
might thus paraphrase for the English reader—' In English bread is
called bread , and it is bread. In French it is called pain, and it is
only bread after all.'
Cf. Sweet, New English Grammar, SS 2272, 2274 .
 PREFACE Xxvii

distinguished living philologist, whom no one could accuse

of slavish adherence to convention, calls the sentence
" The captain with three of his men were taken prisoners,"
ungrammatical.” As a direction to a school-boy the
word might do ; but scientific grammar should have
avoided it as passing an irrelevant imputation upon an
idiom found in more than one cultivated speech.'
“ Grammar " pushes language very hard when, as
happens not so very rarely, the only available expression
is stigmatized as “ ungrammatical.” 2
I find the following stanza in a poem called “ The
Haunted Czar," in Harriet E. H. King's Ballads of the
North
“ Did not God make us, I and thou ?
Have pity even for His sake.
My hair is blanched upon my brow ,
At every rose's fall I shake. "
What is the unfortunate writer to do ? I and thou is
ungrammatical ,' and me and thee would be ridiculous.
Of course we do not deny that the reproach which the
word conveys may be deservedly applied, as when no
reason is obscrvable for the deviation from common usage
or only an excuse of metre or rhyme. But linguistic science
concerns herself primarily with what is, and not with what
should be, said, and at the very first sound of solecistic ,'
ungrammatical,' and the like she must be awake and on
her guard. They are the creakings of Procrustes' bed .
1 In Latin, Livy, 26. 46, 8, “ in quam Mago cum omnibus armatis
refugerant," and elsewhere, even in Cicero ; and (though not so
frequently) in Greek, Thuc., 3. 109, Anuoo évns perd Tær ovotpatryan
'Axapvdrws onévdovral. It is also found in Middle High German.
: 0. Riemann, Syntaxe Latine, passim.
xxviii PREFACE

If we seck for the chief differences in the expression

Differenceof
in ancient
thought and emotion between ancient and
and modern speech,we shall find them to reside not so
modern
employ, much in the mental vocabulary, if I may call it
epirrhemes. so, but in the mental syntax. The contrasts
which ancient and modern rhemes present are great
and startling ; but there is a more fundamental diversity
in their treatment of the epirrhemes. It breaks into light
directly we attempt to translate either from Greek or
Latin. In their expression of relations between ideas
these ancient languages use media holding in solution
elements which we know only as precipitates, and whose
true character will continue to escape us, unless, in despite
of that use which is second nature, we resolutely refuse
to apply to them the rc - agents of our own mental con
stitution. The neuter nominative and accusative were
once, we know, but a single case ; and who can tell
exactly when they ceased to be such ?? Greek grammar
is often at a loss to say in which light the ambiguous
form should be regarded. In Electra, 74, Tò pòv medcoow
βάντι φρουρήσαι χρέος , 1s το σόν .. χρέος, subject to μελίσθω,
or object to ppovpnoa ? If Sophocles rose from the dead,
could he tell us ? A modern reader, however, will not be
| For the explanation of these necessary novelties in terminology
see below , p. 329.
* This appears to be the reason why the neuter of the participle,
used impersonally, chooses the acc. nom. rather than the genitive,
which is preferred in personal constructions, and it appears to go
some way towards accounting for some striking discrepancies
in usage, e.g. 86£av raūta ( Plato, Xenophon ) by the side of
86farta taüta (Andocides, Xenophon), and Euripides, aiai débortai,
πρέσβυ, τλήμονες φυγαί.
 PREFACE xxix

happy till he knows. How many times are we jarred

by finding, as we think, antithesis when we expected
correspondence, as with the accusatives in Sophocles,
Edipus Coloneus, 1685 sqq., tws yap ñ niv ' anlay yâv ň
πόντιον κλύδωνα αλώμεναι δύσοιστον έξομεν τροφάν ; « for
how shall we find our bitter livelihood, roaming to some
far land, or on the waves of the sea " (Jebb's translation );
or with the passive inflexions in Ovid, Trist., 3. 7, 47,
" ingenio tamen ipse meo comitorque fruorque,” and
Martial, 7. 5 , 5 sq., “tuoque | terretur uultu barbarus
et fruitur," where the only way in which we can attempt
to reproduce the formal correspondence is to turn into
a passive a verb which was active as early as Plautus.
Some may perhaps think this ' poetical freedom .'
Then what of prose ? Open Cicero anywhere, and you
will not read far before coming upon passages like the
following :“ si haec non gesta audiretis sed picta uideretis,
tamen appareret uter esset insidiator, uter nihil cogitaret
mali," pro Milone, $ 54. Observe the formal antithesis in
the two participles gesta - picta, and notice how Cicero
glides unconsciously from one sense of the participle to
another. He continues, " cum alter ueheretur in racda
paenulatus, una sederet uxor - quid horum non impeditis
simum, uestitus an uehiculum an comes ? quid minus
promptum ad pugnam , cum paenula inretitus, raeda
impeditus, uxore paene constrictus esset.” Observe
1 We might translate, ' If you only saw these things as painted
and did not hear them as they were done.' The craft of the orator
in turning the ambiguities of language to account, and suggesting,
without appearing to do so, that he is giving a statement of fact is
worth attention .
XXX PREFACE

how the participle -adjective impeditus is first applied

to the causes of the “ impedition " ( impeditissimum of
uehiculum , etc.), and immediately after to the person
who is its object. Note again the slide in the quid
from the actual external objects (uestitus, etc. ) to the
circumstances, " quid minus promptum ad pugnam cum.
And then consider how all this is to be rendered .
Compare the indeterminateness of the subject in
Livy, 26. 19 , 8 sq., of the miraculous stories current
about Scipio Africanus— “ Multa eiusdem generis alia
uera, alia adsimulata, admirationis humanae in eo iuuene
excesserant modum, quibus freta tunc ciuitas aetati haud
maturae tantam rerum molem tantumque imperium
permisit,” where the multa scem to change their char
acter with excesserant, being now merged in the feel
ing which Scipio excited among his fellow citizens,
instead of being the external ground for that feeling, as
before and after in quibus freta - civitas permisit). To
say that the problems of interpretation which underlie
passages like the above are beyond our solution is
true, but far from the full truth. Their very statement
eludes us.
Of the ever-widening rift between ancient and modern
Dead
forms of language, the disappearance of a
figures
of speech ."
number of the figures of speech is another
proof. Enallage, Hypallage, Prolepsis, Zeugma,
" If Livy had been able to use the passive of excedere he might
have expressed his meaning by writing ‘multa-fecerunt ut ad
mirationis humanac in eo iuuene excederetur modus.' 'Although
he was so young, much had combined to make him the object of
more admiration than falls to the share of man.'
‫کم‬
 PREFACE xxxi

Hendiadys, the and KOLVOÜ construction, Hyperbaton,

Hysteron proteron , are all dead or dying.
The much -quoted “he pitched into the room and then
into the lawyer " is grotesque, because language
Zeugma.
no longer permits of such elastic couplings ;
but “ domum simul aduocatumque inuasit ” would have
seemed to a Roman, and should seem to all who would
appreciate Latin, the most natural thing in the world.
The cause of the repetition of the pronoun in French
phrases like ‘ je te remercie et te serre la main ," is, as
Prof. Bréal tells us (p. 51, infra ), to be sought in the
Latin ; but its continued vitality is due to the sharpen
ing of the sense, that different relations of objects must
be differently, or at least separately, expressed.
Nothing is stranger to the modern feeling than the
derangement of epithets or parts of epithets Enallage.
with which ancient poetry swarms ; e.g. Öykov
ονόματος μητρώον, « the weight of the name of mother
(Sophocles) ; Tateixeis Fódovs, “ the seven exits from
the walls ” (Æschylus) ; " Tyrrhenusque tubae clangor ”
( Virgil) .
Allied to this is a use of the vocative of adjectives
for which Plato furnishes instructive examples. Irrational
When in his lively dialogue he uses à Davudore Vocatives.
pardple and the like, this by no means always imports
that the persons so addressed have anything surprising
or blissful about them. It may merely mean that the
· Even mature scholarship is misled by these " enallages.” “The
interpolation in Hor., Odes, 3. 11, 17 sqq., most strangely ascribes
to him a single head with an os trilingue," says the author of the
article Kerberos in Roscher's Lexicon . See above, p. xxii.
xxxii PREFACE

emotion excited by some other object attaches itself in

its expression to the vocative. Thus και θαυμάσιε Κράτυλε
Cratylus 439 only indicates the speaker's, Socrates',
“ intense interest in the mystery of the ideas ” (Campbell,
Theaetetus, Appendix B). In Theaetetus, p. 151 C , Troll
γάρ ήδη , ώ θαυμάσιε, πρός με ούτω διετέθησαν ώστε ατεχνώς
δάκνειν έτοιμοι είναι, the address simply expresses the
astonishment of Socrates at having awakened so much
hostility. You may be surprised to learn ,' we should
say, but with much wcaker effect that some have actually
been ready to bite me ! ' “ It can hardly be rendered
cxcept by a note of admiration " ( Campbell, ib.). Con
trast thc cold blooded way in which Ovid, Ars. Am. 1 .
145, ' cuius cqui ueniant facito studiose requiras,' and
ib. 2. 254, ‘ iunge tuas ambitiose manus,' uses the voca
tive to obtain the short vowel endings of which Latin
poetry felt so much the want. But for the quantity,
we could hardly help taking studiose, ambitiose to be
adverbs.
In a stage of language when shifts like these are
"Attrac- possible, it is not surprising that the “ forms."
tion . " often appear to master the " form ," and cause
a contradiction to arise between what is said and
what is signified, as in the widely extended phenomena
of Attraction . Here where the sense was clear the

| For the tendency of emotional expression to diffuse itself irra .

tionally compare Polle, op. cit. pp. 25 sqq.
? I mean real attraction, not cases like " in quo numero," which
is often said to be for " in quorum numero," though the genitive is
a later exactness .
 PREFACE Xxxiii

words were allowed a free course , with “ irrationality "

as the result."
Thus in Sophocles, Ed. Col., 1321-2, lapdevorraios : ..
επώνυμος της πρόσθεν αδμήτης χρόνω | μητρός λοχευθείς,
“ named from the birth of one who so long a virgin
became a mother at last," the expression , unintelligible
in itself, may be disentangled by observing firstly, that
for the verbal and semi-abstract phrase, énovuuos toll
doxevoivat, Sophocles substitutes the nominal and con
crete one, της - λοχευθείσης (a predicative use familiar to
Latin and not so strange to Greck as some suppose) ;
and secondly, that the feminine participle (which was
metrically inconvenient) became the masculine, with a
shift of meaning from middlc (hoxevomat, "bring forth ")
to passive. Here the“ irrationality " .is glaring ; but it is
just as real in Lucan, 8. 315 sqq., " sed cuncta reuoluens
| uitae fata meae, semper uenerabilis illa | orbis parte
fui.” Only we glide over it, as we do not realize the
value of inflexions which we do not possess, and render
reuoluens by the caseless incoherent English participle.
Lucan's mcaning, “ Turning over all the experiences of
my life (I find) I have always been respected in that
part of the world ," is clear ; but it is not in his words,
which can hardly be analyzed, though I should con
jecture they are the bastard offspring of a union between

· The confusion which attraction works when allied ideas (e.g.

those of age and time) are not clearly discriminated, may be
studied in the Latin expressions for old, older, etc., which are
collected in Madvig's Bemerkungen, pp. 85, 86, or Roby's Latin
Grammar, II . $ 1273 .
xxxiv PREFACE

reuoluenti and fui, as the past of sum in the sense of

videor .'
These and many other phenomena become intelligible
Hysteron when we treat the ancient sentence as a whole.
proteron. A whole it was in a very different sense to that
of our modern sentence, with its fixity of order and its
intolerance of anything like complexity of construction .
Our modern critics are unduly sensitive on the subject
of unusual order, especially in verse. Not so long ago a
heroic attempt was made to expunge hysteron proteron ?
from the roll of grammatical figures. The definition at
tacked was that given in B. H. Kennedy's Public School
Latin Grammar : " hysteron proteron is, when of two
things that which naturally comes first is mentioned
last, as moriamur et in media arma ruamus." The
attack succeeded in so far as it exposed the inappropri
ateness of the adverb naturally, for which usually would
have been better, and directed attention to the preference
of poets for the side-by -side construction of clauses ;
(in other words hendiadys in sentences : there is no
difference in principle between breuia et Syrtes,' the
shallow Syrtes, Aen. 1. 115 , and 'progeniem sed enim
Troiano a sanguine duci audierat Tyrias olim quae
A very similar incoherence is found in Pliny, N. H., 32, § 144
(in an enumeration of marine animals), “fibri quorum generis
lutras nusquam mari accepimus mergi, tantum marina dicentes."
The words in italics have nothing to do in time or in sense with
"lutras nusquam meri accepimus mergi ; ” they give the major
premiss, “ We treat only of marine animals,' to which “ lutras
mergi” is the minor, ' We know of no instances of marine lutrae.'
The conclusion, “ we omit the lutrac," is left out.
• Classical Review , 1894, p. 203 sq.
 PREFACE xxxy

uerteret arces, I hincpopulum late regem belloque superbum

| uenturum cxcidio Libyae,' Aen . 1. 23 sqq., where the
description of the Roman people is divided between the
two sentences.) But it left unexplained why, when you had
got your co - ordinate clauses, they were placed in an un
usual order, and it wholly neglected the influence of metre.
All readers of English poetry are aware of the paucity
of the rhymes in our language, and when they reach
the end of a verse, know better than to inquire too
closely into the perfect appropriateness of the rhyming
words. They give, and the rhymer takes, as of right,
a certain liberty in this part of the verse.'
It was quantity that was the trouble of the ancient
muses, and its yoke was especially heavy on the Roman.
Even in the frcer rhythm of oratory its claims were
exigent. Ciccro finds it necessary to point out to his
contemporarics that flagrant trajections of words are
not to be used, " quo melius aut cadat aut uoluatur
oratio ” ( Orator, § 229 ). The Roman orator who used
these trajections was no more ignorant than ourselves
that thoughts were to be set out in the order in which
they should be presented to the mind, and he deviated
from this order more or less consciously. And the same
is true of the Roman poets. Compare Ibis, 125 sq .,
luctatusque diu cruciatos deserat artus / spiritus et
longa torqueat ante mora,” where the compromise,
“ deserat-torqueat ante," has the effect of a half
apology to the reader.
1 “ Without rhyme or reason " runs the popular phrase ; and the
distinction is charged with unconscious satire.
xxxvi PREFACE

The particular kind of trajection which transports a

Hyper. word from its proper clause to one with which
balon . it has no concern, e.g. Horace, Serm ., 1. 5, 71 ,
"scdulus hospes | pacne macros arsit dum turdos uersat
in igni,” is still more shocking to the modern sense. The
language even of recognition shows how difficult is
their appreciation. Surprise has been expressed that
the ancients, lacking our system of punctuation, could
understand such an arrangement of words at all. The
wonder is natural but misplaced, nor is anything gained
by sowing caltrops over the Latin page after the
following fashion : Ovid, Fasti, 3. 383 sq., “ Mamurius,
morum fabracne exactior artis, difficile est, illud, dicere,
clausit opus." ? The ancients, and perhaps we may say
cspccially the Romans, had a control over the framework
of language of which we have but a faint conception :
this is shown not only by the claborateness of the
'period ,' and the subtlety by which an antithesis was
indicated by a mere juxtaposition of the contrasted
sentences, but by the capacity, sometimes we may think
abused, of holding a disciplined attention in suspense
until, with the addition of the last element, be it word
or sentence, the circuit is completed, and intelligence
flashes over all the connecting wires.?
' I regret that I have sometimes myself adopted this ugly and
confusing practice, as at Catullus, 66. 18, where, when we have
chopped our line to pieces thus— “ non, ita me diui, uera gemunt
iuerint,” we have still to fit them together again.
' I do not of course deny that modern languages can hold
a construction in suspense. But, speaking generally, modern
languages plan their sentences from the beginning, not from the
cnd ; ancient oncs from both .
 PREFACE Xxxvii

The sentences already quoted illustrates this for

words. Horace, Epistles, I. 15, 1 sqq ., is an example of
a sentence which, with two parentheses of cloven and six
lines respectively, extends over twenty -five hexameters
without falling into anacoluthon.
The same thing is evidenced by what we may call the
" inverted &TÒ KOLVOû construction," where a word Inverted
κοινού
-most commonly an epithet-has to be sup- από construc
tion .
plied from the following context. The beginner
finds no difficulty in understanding that in bonae messes et
uites the goodness is to be applied twice ; nor again in
bonae messes et uina, for he easily overlooks the differ
ence in gender. But it is a real effort to him, and often
to his teacher, to learn that in “ messes et bona uina date "
(Tibullus, 1 , 1 , 24 ) the meaning is exactly the same.
To turn to another trifle, as somc may think it, the
perplexing doubt which we often feel as to " Com
when to use a capital first letter for a certain mon "and
“ proper
class of nouns in Greek and Latin is not a names in
one .
mere chimera ; it corresponds to a real indeter
minateness in the ancient rheme. It arises both in the
case of “ common " names which have been specialized
in a " proper" meaning, and of “ proper " names which
have assumed a " common " one ; it affects Venus, the
goddess of physical charm, once a neuter substantive
like genus ( compare venus- tus with onus- tus), Amor,
Spes, and other abstracts imperfectly deified, and Bacchus
and Ceres, godheads reduced to the rank of substances.
1 For the wavering between Spes and spes cf. Tibullus, 2. 6 ,
19-28 ; Ovid, ex Ponto, 1. 6, 27 sqq.
xxxviii PREFACE

Ovid thus describes the mythical transformation of

Ardea
the town of Ardea, Met., 14, 573 sqq.
and ardea .
" cadit Ardea Turno
sospite dicta potcns quam postquam Dardanus ignis
abstulit et tepida latucrunt tecta fauilla,
congerie e media tum primum cognita praepes
subuolat et cineres plausis euerberat alis.
et sonus et macies et pallor et omnia captam
quae deceant urbem, nomen quoque mansit in illa
urbis, ET IPSA SUIS DEPLANGITUR ARDEA PENNIS ."
To blind faith no doubt thic grotesque miracle itself
might still be possible ; but the picture of the last line
it is beyond the capacity of the modern imagination to
visualize. Ardea and ardea can no longer be merged in
onc : in thought no more than in typography.'
By the side of this cohesion of " proper ” and
Namc “ common names may be observed a similar
and

Thing. phenomenon in the conception of name and

66
thing." In Plautus, Pseudolus, 35 , the slave who is
reading a letter written on the customary waxen tablets
says to his master, pointing to the name Phoenicium
1 When, to venture on a translation, Ovid says “ HERON beats
its wings in mourning for itself," he is not indulging in a bold
metaphor like Dante's in Inferno, c. 26 init. “ Godi, Firenze, poi che
sei si grande l'che per mare e per terra batli l'ali," " Rejoice,
Florence, since thou art so great that over land and sea thou
beatest thy wings," nor crudely, as we should say, confusing the
literal with the metaphorical, as at ex Ponto, 2. 5. 38, when he speaks
of a man's mind (pectora ) being ' fair ' - ' fairer than milk or un.
trodden snow ' ; but he is forcing us in a professed picture of fact to
dwell upon an impossible transformation .
 PREFACE xxxix

which is before him, " Tuam amicam uideo, Calidore."

Calidorus looks round with the question, “ Ubi east,
opsecro ? " To which Pseudolus replies, “ Eccam in
tabulis porrectam : in cera cubat.” With us the joke
falls flat, because the bonds of association between the
written name and the living person are but weak and
slight ; but not so with the contemporaries of Plautus.
To them nothing was more natural than the close
association of name and thing. The genuine old Latin
for “ my name is Mercurius " is " nomen Mercurio est
mihi." Towards the end of the Republic the nomina
tive was creeping in from the Greek.'
Students of the Latin poet Propertius will remember
a peculiarity which arrests even a very cursory A peculiar
attention, and which may perhaps be explained " use
synofo
from a similar cause. It is best scen from cx. nyms ."
amples, Prop., i. 22, 6-8, “ sit mihi praecipue puluis
Etrusca dolor ; tu proiecta mei perpessa es membra pro
pinqui, tu nullo miseri contegis ossa solo." In the last
line the dust of Etruria (puluis Etrusca ) is blamed not
for not covering the body (that were an intelligible figure ),
but for not covering it with any solum . Here what is
substantially one idea ( that of the soil) masquerades as
two, and appcars in the same sentence, now as the agent,
now as the instrument with which the action is per
formed . Other writers have the same thing, though
less often . We read in the Panegyric of Messalla, in
cluded among the works of Tibullus, 4. I , 29 sq., ( nam
quamquam antiquae gentis superant tibi laudes, non
1
Compare Polle, op. cit., pp. 122 sqq.
 PREFACE

tua maiorum contenta est gloria fama." A man may be

said to be satisfied with the glory of his ancestry ; but
how his reputation can be no one can say.' Silius
Italicus, 6. 176 sq. , " uastoque e gurgite fusa | tempesta
oritur mixtam stridore procellam | Cerbereo torquens.”
No conceptual line can be drawn between the “ tempestas
quae torquet ” and the " procella quae torquetur.” The
same tendency is seen in the use of adjectives. In Prop.,
3. 16, 30, " cingat Bassaricas Lydia mitra comas," the
poet is addressing Bacchus, and he uses an epithet
formed from Bassareus, one of the god's attributes,
instead of tuas. Lucan , speaking of the dream of
Pompey before the battle of Pharsalia, has, 7. 9, " nam
Pompeiani uisus sibi sede theatri | innumeram effigiem
Romanae cernere plebis,” where even ‘ in the Pompeian
theatre ' would strike us as unnatural ; and so Judge
Ridley translates, " he heard Innumerable Romans
shout his name | Within his theatre.”
It is not allowable to refer these and similar phenomena
to a hankering after variety, such as we may observe
now -a -days in tawdry historians and the writers of the
sporting press, who, whenever they have to repeat the
mention of one of their personages, must revive their
jaded rcader's attention by a new periphrasis. This rest
lessness is absent in the ancient writers, to whom it never
occurred to avoid a word on the sole ground that they
had just made use of it.
" If it be thought that gloria means “ their glory," I would refer
to Prop., 2. 7, 17, " hinc etenim meruit tantum mea gloria nomen , "
wherethis explanation is inadmissible.
 PREFACE xli

No ; here too we must recognize that we are con

templating phenomena of a stage in language which
we have left behind, and that the feeling of mental dis
comfort which they arouse in us is one of the results of
a linguistic evolution which has developed a once im
perfect differentiation of principal and accessory con
cepts, and thus delimitated more strictly the confines of
Rheme and Epirrheme. That the modern interdict marks
intellectual advance can hardly be denied. Expressions
like " Bacchus bound the Bassaric tresses," or "Apollo
strung thc Phocbcan lyrc," obscure the fact that the
relation of the objcct to the subjcct (that of possessed
to possessor) is in both cases the same, and thus con
fuse the perception, while the employment of two rhemes
( Bacchus - Bassareus) where one is enough , weakens the
effect of both , and paves the way for the disappearance
of one of them."
To sum up , may we not say that language has to a
certain extent lost the extreme ductility which arrested
the attention of Cicero , ' est oratio mollis et tenera et ita
flexibilis ut sequatur quocumque torqueas,' Orator & 52,
cf. De Or. 3. 176, or, to change the metaphor that, while
ancient compositions in their blending hucs and often
hazy forms recall thc cffccts of the brush and the
palcttc, modern ones suggest rather thc dcfinitc colours
and sharper outlines of work in mosaic ?

* In the introduction to my Select Elegies of Propertius I adopted

the term “ disjunctiveness ” for these phenomena. From the
point of view of expression they are indeed “disjunctive," but from
that of thought they are rather conjunctive.
 PREFACE
The rule that synonyms, or quasi-synonyms, must not
Contrastedbe used
in contrasted relations is a rule of modern
theme speech ; but the law on which I shall now touch
belongs, it would scem , to language in cvery stage.
It is that consciously contrasted rhemes must differ in
outward form .

I will illustrate this from a tendency which I have

observed in my own pronunciation of the word ass. The
a of the small group of vocables to which ass belongs
(grass pass, etc.) is, as we all know, rendered differently
by different English speakers, some giving the sound of
cat, others that which is customarily heard in father.
The first was the pronunciation to which I was brought
up, and this I naturally use in speaking of the animal,
but when I have occasion to use the word metaphorically
I feel impelled to give its a the second value. A friend
to whom I mentioned this told me that he too felt the
impulse to differentiate the mcaning through the sound,
but, as his familiar a was the second one, his literal ass
would be my metaphorical one. If we were obliged to
use the word in its different senses in the same context,
the distinction which now seems natural would become
inevitable .
This may seem obvious, but when an ctyinologist of
the standing of Professor Skcat has lent countenance to
a different view , it is not supcrſuous to state it. In his
' The German schwein must have been similarly differentiated
by the swineherd whose jealousy for the honour of his hogs has
furnishedDr. Polle with one of hisamusing stories. “ You think a
pig is a pig ? A pig is not a pig (ein Schwein ist kein Schwein),
but a very clean beast ! ” (Op. cit. p. 52).
 PREFACE xliji

Etymological Dictionary of the English Language he

writes “ Quean, a contemptible woman, a hussy [E.). In
Shak ., Merry Wives, IV. ii. 180. Absolutely the same
word as queen, the orig. sense being ' woman .' The
difference in spelling is unoriginal, but may have marked
some variation of pronunciation. The best passage to
illustrate this word is in P. Plowman, c. ix. 46, where the
author says that in the grave all are alike ; you cannot
there tell a knight from a knave or a queen from a quean.
The MS. have queyne, queene, quene, in the former case,
and queene, quene in the latter ; i.e. they make no dis
tinction, none being possible.” Some may think it a
pity to destroy the fancy which has amused many
writers and readers of etymology that the name of the
highest lady in the land is " absolutely the same " as that
of “ a contemptible woman, a hussy." But the two
words were for all that absolutely distinct, and until quear
died out of our language so they must have remained ;
and this is proved by their occurrence in opposition not
only in the passage of Piers Plowman , but elsewhere,
€.g. Ford, Perkin Warbeck, ii. 3—
“ I never was ambitious
Of using congecs to my daughtcr.qucen.
A queen / perhaps a qucun / "
" It is not surprising thne Dr. K. Schmidt says " Aumillig ist das
Fortbestchen von engl . qucen neben qucan," Die Griunde des Bedev .
tungs wundel (Berlin, 1894 progr.), p. 41 .
• The two words are rightly separated by Mr. Mayhew in his Old
English Phonology, $ 480, p. 138 (which see), as I found after the
above was written ; and Professor Skeat tells me that he nuw
regards them as separate. Of course they are from the same root,
that of the Greek γυνή ..
xliv PREFACE

and Heywood, A Woman Killed with Kindness, iii, 2,

171 sq .
The application of this principle throws valuable light
upon dark or doubtful places in the history of language.
When by the loss of the rough breathing and the confu
sion of n and v, nucis was no longer distinguishable from
úveis,'the instinct of the Greek.language sct itself to revive
the perishing difference between meum and tuum . And
hence it comes that the modern Greek, when he wishes
to put you in the plural, does not employ jueîs, but ocês
or docis. The same source furnishes, irrefragable evidence
that in the time of Hesiod o had a perfectly distinct
sound from , with which the modern Greek confuses it.
For in Works and Days, 242 , we read uby' enýyaye zîna
Κρονίων | λιμόν ομού και λοιμόν. The same is shown for
the time of Thucydides by a well-known passage (2. 54,
3), where he tells us that at the time of the Great Plague
at Athens the elder men remembered and quoted an old
Song (φάσκοντες ... πάλαι άδεσθαι), ήξει Δωριακός πόλεμος
και λοιμός άμ' αυτό, and that there was a dispute whether
the word was loquo's, “ pestilence," or demo's “famine " ;
but that the version with lowo's prevailed as then the
more appropriate one.

1 This is known to those who have travelled in the country.

quoto from a private letter of Professor Ernest Gardner, Into
Director of the British School at Athens— " In spoken Grcek oris
or dovîs is uscd for you plural, corresponding to ou or dol in the
singular ; the accusative is sing. oh or dolva, plural oâs doâs. bueis is
practically restricted to stilted pedantic conversation ."
See Mr. C. H. Monro's note in the Proceedings of the Cambridge
Philological Society for 1895, p. 12 .
 xlv
PREFACE

One of the most hopeful signs for the future of our

science is the stricter treatment of Analogy. At
Analogy
the first uprising of die neueste Sprachforschung
in the cause of rigid Phonctic law, the cry of meaning
was unheard, and any analogy had to serve if it could
be pressed into thc servicc ; but most of these incfficicnt
allics have now been sent about their business. Much
good work has been done in the elucidation of a difficult
subject, yet not av little remains to do. The respective
influences of sound and sense in these analogical con
nections have to be more carefully defined . Cutler is
mentally associated with cut ; why not butler with butt ??
Cut has nothing to do with cutler, butt and butler have
the same origin. The answer is that cut is a rheme of
character ; butt (or but8) is not. Cut brings before the
mind the distinct picture of an unmistakable physical
action. But is now a rheme and now an cpirrheme (the
conjunction but) ; and which of its many suggestions the
mind should take it is impossible without further light
to determinc.
In these inter-actions between rheme and rheme * both
E.g. the suggestion that the difference in the vowel of Greek
Kabvis and that of the Lat. clunis was due to a popular connection
between muovis and Kovéa . Brugmann , Zum Houtigen Stand der
Sprachwissenschaft, pp. 70, 89 (1885), and Grundriss, ed. 1,
p. 219 notc ; omitted altogether in the second edition of the
Grundriss ( 1898).
Butler, from Fr. boutelier, Lat. buticularius from the buticula,
diminutive of butis, a cask or butt.
The spelling is of course perfectly immaterial.
4 I am not sure how far Prof. Wheeler could accept my qualifi.
cation or interpretation of his "fundamental principle oflanguage
history. ... The psychological grouping from which the phenomena
xlvi PREFACE

sound and meaning are involved ; but speaking gener

Sound ally, we may say it is sound that takes the
dominant.initiative , dragging meaning in its train. These
changes always suppose some comparative weakness in
the rheic that is affected ; it is those accordingly w se
mcaning is most fluctuating or attenuated, which are
least strongly rooted in associations of their own, that
are most exposed to attack. The dominance of the
sound in the transforming of strange or foreign words
is very noticeable. Learning Greek before German, as
I did, I felt an almost irresistible impulse to attach the
sense of adelvós, “ glorious," to the German klein. I am
told that those who learn German first feel that kielvós
ought to mean little. In such cases we may hope that
the modification may regard the interests of sense ; but
there is no guarantee that it will. The English trunk
of an elephant is a most inappropriate transformation
of his trump or trumpet, i Fr. trompe.' The Roman
who saw in the same beast's tpoßoorls his fly- flapper,
promuscis, was a little more reasonable. The trans
formations which have not become current are just as
interesting for us and just as important. One of our
college servants, referring to some orange plants which
were recovering from a severe frost, observed that they
had been re-suffocated . A housekeeper in one of the
ofanalogy result is never a grouping on the basis solely of meaning,
nor the basis solely of form ; both are involved in every case .”
Origin of Grammatical Gender, 1.c., p. 533 note.
· The popular transformation (it would be a compliment to call
it an etymology) has, however, produced no confusion, as we do not
use trunk of the body of quadrupeds. 2 Resuscitated.
 PREFACE xlvii

Canons' houses at Ely referred to a new Canon as one

of them chalybeates.'? An evening class of Cambridge
girls always rendered the song " The two little orphans
of parents bereft" as "The two little oftens," etc. The
havoc which gardeners make in botanical names is well
known. Mr. P. Giles tells me of one who re -christened
the rose Gloire de Dijon as Glory to John.
As most of us know, times of mental stress cause
strange perturbations in our use of language. A can
didate in a recent examination wrote down the name
of the Roman demagogue as Serpentinus. His wool
gathering mind identified the first two syllables of
Saturninus with Satan, and the author of evil in his
traditional shape thus wriggled into the word. In
another examination a gentleman of obviously sporting
tastes turned the ' Menai Straits ' into the Conway
Stakes '!
It is equally clear that in certain classes of rhemes
which, it will be one of the first tasks of our Meaning
study to determine -- it is the meaning that is dominant.
responsible for the changes. Some of our notions are,
it would seem , natural couples, and of these either seems
to involve its fellow , be this its twin or its anti-type.
i Celibates.
* This is an interesting case . The pronunciation of a word by
the same speaker is not constant. It varies as he speaks slowly or
rapidly, carefully or carelessly ; and so arise Lento -forms and
Allegro - forms as they have been named . With these girls the
Allegro -form , i.e. the most familiar form of orphan and often, was
the same (orfn ), and this led to the transfer of the Lento -form of
often to orphan.
d
xlviii PREFACE

I have before me a piece of Latin prose where I find

that I have written magis when I meant minus, and
Latin MSS. show that copyists make the same inter
change.' M. Bréal discusses the volatile negative, and
my cldest boy when beginning to talk used "any more
for “ no more.” My friend and colleague, Prof. Arthur
Platt, tells me how an acquaintance of his, who could
not find a match for his pipe when he wanted it, said
“ I seldom forget to come without one ! ” — a curiously
complex instance of confuson . Even Goethe has
“Warum tanzen Bübchen mit Mädchen so gern ?
Ungleich dem Gleichen bleibet nicht fern ."
The inter - relations of concrete and abstract names
“ . Abu. is a subject of considerable linguistic and philo
stract
and " con sophical importance. In chapter xiii. M. Bréal
.
crete
rhemes.
gives numerous examples of “ concretion ” in the
meaning of abstracts. That names of actions and
statcs have passed, and still pass, into thc names of con
crcte objects is undeniable : but the causes at work are
by no mcans perfectly clcar. It is not surprising that,
when once the concrete application had made its way
into the usages of a word , it should overpower the
abstract. Language is built upon the senses, and they
exert upon it an attraction as powerful and constant as
gravity. But it is not easy to see what is the first stage
or the ground of the change. Why, when a concrete sub
stantive is required, does the popular mind turn to the
abstract ? Why do we speak of our “ possessions ” side
2
Te.g. at Manilius, iv. 110 . Infra, pp. 200 sqq.
3 Polle, op. cit., p. 60 .
 PREFACE xlix

by side with “ possession is nine points of the law" ?

Why did the double sense develop in the Latin man
cipium ? Was not seruus enough to express a slave that
thcy must matcrialize seruitium ? May it bc that to the 2001
popular consciousness the abstract sums up, as it were,
the potentiality of the rheme, and so that when a new
word was wanted, this naturally suggests itself, not
because it is abstract, but because it is undetermined ?
Certainly it is not easy to see how a people with the
concrete and almost earthy imagination of the Romans,
a people who built temples to Honor, Concordia ,
Virtus, and other creations of the mind, as though
they were material beings, at the same time used the
abstract with a freedom which we cannot reproduce,
unless it was for them the centre or, we might say, the
source of the potency of a rhemc.?
The classification of rhemes is one of the first ncccs
sitics to our science. Its importance to scientific The clas.
grammar has lately been recognized by two ofthemes.
1 I mean usages like those of a noteworthy passage of Cicero, de
nat. deorum , 2. $ 98, “ Adde huc fontium gelidas perennitates,
liquores perlucidos amnium, riparum uestitus uiridissimos, spelun
1
carum concauas altitudines, saxorum asperitates, impendentium
montium altitudines immensitatesque camporum adde etiam recon
ditas auri argentique uenas infinitamque uim marmoris .” Nouns,
which were not originally distinct, but have become so, are treated
in the same way : cor (the heart, we should say ' brains ') is used
for sapientia, and so , Persius can say cor iubet hoc Enni ”
( = Ennius cordatus), “ so bids sage Ennius.”
% I hope M. Bréal will not consider my language too metaphorical.
As we cannot escape from metaphor, I think we may as well
employ it. The risks of deception will be less the more freely we
do so.
1 PREFACE

inquirers. Prof. Delbrück prefaces his comparative

syntax of the Indo -European verb with a distribution of
roots according to the way in which the element of time
enters into the conception ; and Prof. Elmer has given a
not very dissimilar division of verbal roots for the purpose
of clearing the ground for his discrimination of the
usages of the Latin subjunctive present and perfect.
The utility which such divisions possess would be much
diminished if it blinded us to the possibility of illumining
the dark places of grammar by means of distributions
based upon different principles.
There is a singular phenomenon to be noticed in the
Middle
formation of the Greek verb to which Dr. Ruther
futurefrom
active ford has deservedly called attention. “ All verbs,"
presentsin
Greek .
he writes,2 " which refer primarily to a physical
process, and do not merely state the fact that
such and such an action is going on, are either deponent
throughout or deponents in the future tense " (my italics);
and later, " The verbs which reject the active endings of
the future in favour of the middle endings, at the same
time that they retain the active inflexions in their other
tenses, are all words expressing the exercise of the
senses or denoting some functional state or process."
Such words are oρώ fut. όψομαι, πίνω fut. πίομαι, νέω
fut. νέυσομαι, βοώ fut. βοήσομαι. Dr. Rutherford writes
of these formations, “ The reason for this anomaly in form
it is useless to discuss as it is impossible to discover."
· Studies in Latin Moods and Tenses, Cornell Studies in Clas
sical Philology, No. vi.
· New Phrynichus, p. 383.
 PREFACE li

In this Gibbonian sentence the eminent Grecian in

tended merely to decline and not to discourage research ;
and it should spur the votarics of our subject to provide
him with the solution that he sceks. Let us see if we
can do anything towards it. The problem involves a
double question. ( 1 ) Why should the future more than
other tenses show a preference for the middle or reflexive
forms ? ( 2) Why should this preference be confined to
verbs which have a particular meaning ? The answer
may perhaps be found without plunging deeply into
metaphysics. In proportion as the concept underlying a
rheme is felt to involve a reference to present conscious
ness, will the mind experience a difficulty in referring
it to the future, or, to use a metaphor, would we grasp
a concept that is to be so referred, it must be provided
with a handle.
All the verbs furthermore which show the peculiarity
in question involve such reference in a greater or less
degree. Take “ hearing ” for example. “ I hear” gives us
no trouble to grasp, nor, with the aid of recollection, “ I
heard ," " I have heard " ; but throw the concept into the
future “ I shall hcar,” and it seems to become elusive
at once . In this instance the reference to a present
consciousness may be called " pure," but this is not
always the case. I cannot obscrve other people's hear
ing ; but I can observe other people's jumping. And
hence into the Greek rheme, πηδώ, future πηδήσομαι,
entered not only a reference to the Greek's own muscular
sensations which accompanied the act of leaping, but
also a reference to other impressions received through
lii PREFACE

a different sense, the sense of sight, which he instinc

tively felt were very different from the former.
Now the use of the forms of the middle provided Greek
with the handle required. These apparently differed
only from the active in the greater stress laid upon the
concern of the subject in the action. Thus åkoúw was
simply I hear. The effect of akoúrouat may be given
by I will hear : 1, where thc enforcement of the idea
of the personal subject represents, we may say, the
mental effort required to project my hearing into the
future .
The fugitive, elusive nature of these concepts is shown
in another way. Modern English, when it has to express
present time, docs not use the (old ) present tense, as it
rains, but the periphrastic form as it is raining, they
are playing, and even it is lightening. But to this rule 1
verbs expressing the activities of the senses or the mind
are exceptions. Thus I hear, I smell, I think, but not,
except in occasional usc and with a special meaning, I
am smelling, I am thinking:2
It may be noted in passing that Delbrück's view
( Syntax, ii. p. 71 ) that indicative presents formed from
“ punktuell” roots should have, and once had, a future
sense ( as clue “ I shall go” actually has), and lost it
through the influence of the presents formed from this
other class of roots, receives no support from English,
1 In such cases there will be wide differences in the concepts, as
in the case of a blind man, a professional athlete, and a mere
spectator.
? I am glad to see this peculiarity duly recorded in Dr. Sweet's
Historical Grammar of English , vol. ii. § 2218.
 PREFACE liji

which replaces in this case also the form I go = Ger. ich

gehe ( “ I go, sir, and went not,” Authorised Version ) by
I am going. Qu. When do you go (or are you going) ?
Ans. I am going ( so usually) to -morrow , next week, in a
month's time .
The use of different rhemes when we should have
expected only different epirrhemes, or, in other words,
the general question of anomalous formations ( irregular
comparison, declension, conjugation, etc.) in Indo
European languages, has been dealt with by Prof. H.
Osthoff of Heidelberg in a recently published pamphlet.?
He regards these irregularities (an appellation however
against which he protests) as remains of the individual
izing tendency of carly specch in opposition to the
classifying methods which subscquently prevailed. The
explanation is unquestionably correct as far as it goes,
and it is true also, as he says, that " the objects of the
conceptual world have more sharpness and individuality
thc ncarer they are to the perceptions and thought of the
spcakcr.” But these generalities take us only a little
way on the path. What we want to know is in what
rhemes this ancient individualizing tendency persists,
and why it persists in them .
1 Delbrück divides roots into “ punktuell ” and “ nichtpunktuell.”
This definition of “ punktuell," for which there is no corresponding
English term , is as follows: “ Punktuell ist eine Aktion wenn durch
sie ausgesagt wird dass die Handlung mit ihrem Eintritt zugleich
vollendet ist.” He explains that he regards the vast majority of
all roots as “ punktuell.” To the minority, e.g. es “ be," he gives
in default of a better the name of “ nicht-punktuell.”
· Vom Suppletivwesen der Indo- Germanischen Sprachen, von H.
Osthoff, Heidelberg, 1900.
liv PREFACE

Prof. Osthoff refers, inter alia, to the interesting differ

ence which is represented in the English hart, hind.
With these exceptions the language of our Teutonic fore
fathers , formed the feminine of all wild animals from
the masculine. Prof. Osthoff rejects the explanation
that the anomaly is due to the fact that the male deer
is notably distinguished from the female by its possession
of horns ; and prefers to seek its cause in the German
love of the chasc. Supposing, and I confess I sce no
reason for this, that the two explanations are incom
patible, the only way of resolving the dispute would be
to examine the languages of other hunting nations with
this special point in view.
The preceding discussion suggests at once that psy
chology, and through that metaphysics, will
The gain advantage from the study of Semantics.
Scicnce
of
Meaning As I have touched upon this elsewhere, I will
losophy. only here say, that when philologists have
and Phi-
stripped from a rheme all that it derived through
its outward form , whatever remains is material for the
psychologist. Philosophy has long been aware of this ;
but its use of the stores of language in the past has
been but sparse and casual. With a systematized study
of meaning, we may hope that this will change.
From another point of view it should be found . of
service to the philosopher. No branch of inquiry has
suffered more from language than philosophy. For none
is strictness and consistency of language more necessary,
or, apparently, inore difficult to secure. In the latest
1 Old High German , hirus, masc. , hinta, fem .
 PREFACE lv

treatment of this subject by Dr. F. Tönnies, in a triad of

articles in Mind on Philosophical Terminology, it is said
of "the obscurity and confusion in psychological and
philosophical terminology " :
“ We must refer those who doubt it to a comparison
of the best-known works, those which excel each in its
speciality, of European and American origin. Almost
each of these operates in these regions with different
concepts, or at any rate with concepts which arc differ
ently determined. But cven within each particular work,
if we examine carefully, we shall not find a consistent
terminology ; but shall often find that the sense in which
an expression was introduced, even the definition which
was given with it, has been neglected and apparently
forgotten in the course of the exposition, so that the
reader who would rest upon it as upon a staff, feels it
break in his hand. " If we regard philosophy so far
as it comes into contact with the whole of life, we
shall feel with pain the misunderstandings which
stream from the uncertainty and confusion of language
( Eucken ).' ” 1
A most deplorable state of things. But who is to
blame ? If the present century has taught us anything,
it has surely taught us that the one way by which we
can reach an understanding of the present is through
the study of the past. If this cannot be denied, where
are the exact and patient studies by which the develop
Dr. F. Tönnies, in the Welby Prize Essay, published in Mind,
1899, p. 467. The translators into English of Wundt's Psychology
make the same complaint.
lvi . PREFACE

ment of the concepts in question is traced through

various surroundings and successive centuries, by which
their changes are unravelled or their stability demon
strated ? I ask the philosophers: Where ?
This general survey of the subject would be still more
imperfect than it is did it omit all reference to the
methods which our science should employ, and the
practical means by which its advancement may be best
secured . Its initial and ever -recurring problem concerns
the constitution of the rheme, and may be briefly put as
follows. A certain combination of articulate sounds is
used by an individual (you, I, or he) to recall to himself,
or another, a certain mental tálos or affection (idca) :
what is the mental affection so recalled , and why is this
or that combination used to recall it ? As we have said
already, no inquiry into the present of a rheme can dis
penso with the help of theinquiry into its past ; but though
The
ultimately the two inquiries are not separable,
methods they may be separately pursued. If we would
science. investigate the present contents of a rheme,
three methods appear to be at once available ;
the obvious but treacherous procedure of introspection,
the observing of subjects who are unconscious of this
observation, and the scrutiny of the cvidence furnished
in contemporary documents. To these may be added as
a fourth the study, from a linguistic point of view,' of
· Aphasia, so far as I know, appears to have been little studied
from a linguistic point of view. A contribution , however, has been
made in Meringer and . Meyer's Versprechen und Verlesen, Stuttgart,
1895
 PREFACE Ivii

the phenomena of partial aphasia , including everything,

from the confusion which accompanies a local lesion of
the brain , to those weaknesses which we know by the
familiar names of slips of the tongue or of the pen .
A metaphor may serve to make this clear. So long as
all works well, we travel from sense to sound, as in the
dark, unconscious of our route ; but once let the " points "
be deranged, and the unexpected terminus reveals an
unsuspected junction. If we wish to have a single
word to include all those branches, we might extend the
meaning of a term already in cxistence, and call it the
" Synonymik " of the present; but I should prefer the
new name Rhematology. There will also be a Rhema
tology of the past, whose methods will only differ from
the former's by the fact that inasmuch as its subject
matter is non -contemporary, the first two inodes of
investigation will not be available. The loss of the
second one can only be imperfectly replaced by a collec
tion and comparison of analogous observations from
other times and places.
The investigation of the why of rhemes is a study of
their history and cannot be too severely historical. All
competent lexicography, of which there is unfortunately
still too little, now recognizes this ; but Prof. Bréal,
in his Etymological Dictionary of Latin , has the dis
tinction of being one of the first to apply it systematically
to the etymology of an ancient tongue. In the historical
i I should limit the word to the special study of separate rhemes,
preferring Semantics as the general name of our science.
* Dictionnaire Etymologique Latin, par Michel Bréal et Anatole
Bailly, Paris, 1885 .
lviii PREFACE

treatment of the rhemes of a language there are three

cautions which must be ever borne in mind. We must
be careful not to limit history to chronology, and con
fuse the contemporary with the coeval. Progress in
language is never uniform , and the caprice of a literary
record may happen to show one and the same date for a
usage which was fresh from thc mintage of the moment,
and for one which is the last remnant of a forgotten arid
unintelligible past." We must distinguish carcfully be
tween the products of the unconscious and of the
conscious activities of mind. The difference between
popular and literary, or scientific, creations in language
is no less than the difference between instinct and inten
tion. And, lastly, in the study of this, as of every other
development, we must observe the trend. The progress
of a language may be compared in some respects to the
progress of a spiral down its conc. The same phases
will be repeated but not the same moments , and it will
· The proper recognition of the stratification of language is not
without importance in the settlement of literary problems. By a
singular fatality what is perhaps the most convincing argument for
the genuincness of Sophoclcs, Ajax, 841 sq., has bcen used to
demonstrate its spuriousncss. In order to obtain for the curse of
Ajax the utmost solcmnity possible, Sophocles has clothed it in an
ancient Ionic and epic form, as is shown by τώς, φιλίστων, όλοίατο.
This cffect could hardly have been obtained in any other way. Of
PINIOros, to which chief exception has been taken, it may be observed,
that though not found elsewhere in extant literature, it is sufficiently
supported by the pidime of the Odyssey, and by its use as a proper
name. As regards the ending -aro, we may note that it is not
without significance that tragedy confines its use to the optative,
and, as my friend Prof. Ridgeway pointed out to me a good many
years ago, to the optative of uncontracted verbs.
 PREFACE lix

never free itself from the shaping conditions of its native

medium . 1
It is easy to see how vast are the territories that apper
tain to our subject, and they are almost wholly unex
plored . The divinations of individuals, however brilliant,
can do but little. They but flash illumination into the
scene for a moment and are gone, with the darkness
blacker than before. Only by the methods to which
the progress of all other sciences is due, can this make
substantial advance. Facts must be collected from every
quarter, wide and at the same time discriminating induc
tions formed , and the whole field held and controlled by
the rigorous method and the chastened imagination of
science. The amount of work to be done is beyond
question great ; but it can be readily distributed, and
its interest is not less indubitable. But co -operation is
required. A society should be formed whose task should
be to gather and arrange materials, to furnish direction
and advice to individual workers, and to put before the
world the fruits of their research. May the foundation
of such a society be not far off ; for its need is very
pressing
" I nccd not here repeat what I have said elsewhere upon the
necessity of studying either outlying dialects, savage languages
and the speech of children. See Appendix, pp. 333 sq. [Only since
this preface was in type has Wundt's important Völkerpsychologie
( 1900) come into my hands. J. P. P.]

J. P. POSTGATE.
 CONTENTS

PAGE
PURPOSE OF THE BOOK

PART I
THE INTELLECTUAL LAWS OF LANGUAGE

CHAPTER I
THE LAW OF SPECIALISATION .
II
Definition of the word law - False idea prevalent on the
subject of languages known as synthetic and analytic
-Specialisation of function one of the characteristics
of analytic languages.

CHAPTER II
THE LAW OF DIFFERENTIATION 0
37
Proofs of the existence of Differentiation - Limits of the
principle of Differentiation .

CHAPTER III
IRRADIATION
39
What is to be understood by this word - Irradiation can
create grammatical inflections.
Ixii CONTENTS

CHAPTER IV
PAGK
SURVIVAL OF INFLECTIONS . . . . .
50
What this is — Examples drawn from French Grammar
Archaism .

CHAPTER V
FALSE PERCEPTIONS .
56
False inflections of the plural - False inllections of cases
-Apophonia.

CHAPTER VI

60
ANALOGY

False idea of Analogy - Cases in which Language allows

itself to be guided by Analogy- ( a) To avoid some
difficulty- ( ) To obtain greater clearness -- ( ) To
eniphasise either an antithesis or a similarity- (d) To
conform to some ancient or recent rule - Conclusions
on Analogy. 01

CHAPTER VII
NEW ACQUISITIONS 78
Need of indicating acquisitions together with losses
The infinitive - The 'passive - Adverbial suffixes
Historical conclusions.

CHAPTER VIII
EXTINCTION OF USELESS FORMS .
90
Difficulty of this study - Superabundant forms produced
by grammatical mechanism - Advantages of Extinc
tion - Are there any forms irrevocably doomed to
disappear ?
 CONTENTS Ixiii

PART II
HOW THE MEANING OF WORDS IS DETERMINED

CHAPTER IX
PAGB

THE SO -CALLED TENDENCIES OF WORDS .

99
Whence comes the " pejorative tendency " ? - The
" tendency to deteriorate " --Other tendencies no less
imaginary.
CHAPTER X
RESTRICTION OF MEANING . 106

Why words are necessarily disproportionate to things

How the mind re-adjusts this disproportion.
CHAPTER XI
EXPANSION OF MEANING .
. 115
Causes of Expansion of Meaning-Facts of Expansion
are so much information gained for history - They
are a consequence of the progress of thought.

CHAPTER XII
METAPHOR 122

Importance of Metaphor for the formation of Language

Popular metaphors - Divers origins of metaphorical
expressions —They pass from one language to
another.

CHAPTER XIII
ABSTRACT WORDS AND CONCRETION OF MEANING 134
What is to be understood by Concretion of Meaning
Examples drawn from various languages.
lxiv CONTENTS

CHAPTER XIV
PAGE

POLYSEMIA 139
What Polysemia means–Why it is a sign of civilisation
- Why it causes no confusion - A new acceptation
is equivalent to a new word — Indirect Polysemia.

CHAPTER XV
A SPECIAL CAUSE OF POLYSEMIA 146
Why an expression can be mutilated, without losing any
thing of its signification - Abridgment, the cause of
irregularities in the development of meanings - Ex
pressions known as " pregnant."

CHAPTER XVI
COMPOUND NOUNS .
155
Importance of meaning - The order of the terms — Why
Latin forms less compounds than Greek-Limits of
Composition in Greck-Sanscrit compounds - Com
pounds have never more than two terms.

.CHAPTER XVII
ARTICULATED GROUPS 166

Examples of Articulated Groups - Their use.

CHAPTER XVIII
HOW NAMES ARE GIVEN TO THINGS .
171

Names given to things are necessarily incomplete and

inaccurate - Opinions of the philosophers of Greece
and of India - Advantages of phonetic change
Proper names.
 CONTENTS Lxv

PART III

HOW SYNTAX IS FORMED

CHAPTER XIX
PAGE
THE PARTS OF SPEECH 181
What are we to understand by the parts of speech ?—How
they exist in the mind -- Are they innate or acquired ?
- Do they all date from the same period ?

CHAPTER XX
TRANSITIVE FORCE .
189
Whence arises our idea of a Transitive Force resident in
certain words - Verbs which change their mcaning
in becoming transitive-Transitive Force is that
which gives unity and cohesion to a phrase - The
ancient grammatical mechanism is despoiled of its
original value.

CHAPTER XXI
CONTAGION 200

Examples of Contagion - Negative words in French

The English but - The active past participle — The
Latin conjunction si

CHAPTER XXII

ON CERTAIN GRAMMATICAL INSTRUMENTS . 205

The relative pronoun — The article — The verb sub
stantive - Auxiliary verbs.
Ixvi CONTENTS

CHAPTER XXIII
PAOR
THE ORDER OF WORDS 212

Why strictness of construction is in inverse proportion to

richness of grammar - Whence comes the Order of
French construction - Advantages of a fixed Order
- Comparison with the modern languages of India.

CHAPTER XXIV
THE LOGIC OF LANGUAGE . . 219
The nature of the Logic of Language - How the popular
mind proceeds.

CHAPTER XXV
THE SUBJECTIVE ELEMENT .
• 329
What we are to understand by the Subjective Element
Its connection with speech-The Subjective Element
is the most ancient part of Language.

CHAPTER XXVI
LANGUAGE THE EDUCATOR OF THE HUMAN RACE .
239
Part played by Language in the operations of the mind
Wherein lies the superiority of the Indo - European
languages — The place which the Science of Language
should hold among sciences.

WHAT IS MEANT BY PURITY OF LANGUAGE ? 255

THE HISTORY OF WORDS 279

APPENDIX (THE SCIENCE OF MEANING) . .

311
INDEX .
337
 SEMANTICS
PURPOSE OF THE BOOK

THERE is a constant succession of books on the

subject of comparative grammar, for the use both of
students and of the general public ; yet it does not
seem that we are offcred what we really need.
Language is full of lessons for those who know how to
question it. Through all the centuries humanity has
deposited in Language the acquisitions of material
and moral life. But it must be approached from the
side on which it appeals to the mind. If we limit
ourselves to the changes of vowels and consonants, the
study is reduced to the proportions of a merely secondary
branch of acoustics and physiology ; if we think it
enough to enumerate the losses undergone by the
machinery of grammar, we give the impression of a
building that is falling into ruins ; if we confine ourselves
to vague theories on the origin of Language, we merely
add an unprofitable chapter to the history of systems.
B
2 SEMANTICS

There is something else to do. We may extract from

the Science of Language whatever food for reflection it
suggests. We may also find therein some rules for our
own tongue, since each of us contributes his share to
the evolution of human speech . It is that which should
be brought to light, and it is that which I have tried to
illustrate in this book.
Not long since the Science of Language would have
thought it derogatory to admit that it might serve some
practical end. It existed, so it said, for itself alone ;
and it no more took heed of the profit that mankind
might draw from it, than the astronomer, calculating
the orbit of celestial bodies, regards the prevision of the
tides. Should my colleagues consider that my method
will debase our science, I can only answer that I do
not belicve that their high aims are justified . Their
aims, indeed, are not suitable to the study of a human
product such as Language. For Language is a pro
duct, begun and continued with a practical goal in
view, from which, in consequence, the conception of
utility cannot be absent for a moment. More : I believe
that to cxaggerate these researches is to deprive them
of their worth . The Science of Language expresses man
to himself : it shows him how he has constructed and
perfected the most necessary instrument of civilisation,
despite obstacles of every kind, despite inevitable delays,
despite even momentary retrogressions. I. is the right
of this science also to explain by what means the
instrument, which has been confided to us, and for
which we are responsible, is preserved or modified. It
 PURPOSE OF THE BOOK 3

must surely, then, amaze the thinking reader to be told

that man counts for nothing, and that words — both in
form and meaning - live a life peculiar to themselves.
The danger of our studies has lain and still lies in
the abuse of abstractions and of metaphors. We have
seen languages. treated as living beings : we have been
told that words are born, fight, propagate, and die.
There would be no harm in this way of speaking, if
there were not people who interpret it literally. But
since such people exist, it is necessary to protest with
out ceasing against a terminology which, amongst other
drawbacks, exempts us from seeking the true causes."
The Indo - European tongues are condemned to figur
ative speech. They can no more escape from it than
man , according to the Arab proverb, can jump out of
his shadow. The structure of the phrase compels them
to it : there is a perpetual temptation to vivify that
which is lifeless, to transmute into action that which
is merely a state. Dry grammar even is defenceless
against imagery : what is it but a beginning of myth,
when we say that pépw borrows its tenses from évéykw, or
" In writing this, I am thinking of a whole series of books and
articles, both foreign and French. The French reader will especi.
ally recall the little book of Arsène Darmesteter, La Vie des Mots.
There is no doubt that the author has unduly emphasised and
pushed too deep the comparison, so that at times he really appears
to believe in his own metaphors, a pardonable fault when one
thinks of the influence of style. I was a friend of both the
Darmesteters, those Açvins of French philology, throughout their
lives, I have done homage to their memory, and I should deeply
regret saying anything that might reflect upon it. (See at the end
of this book my article on La Vie des Mots .)
 4 SEMANTICS

that “ clou ” takes an s in the plural? But of all men

philologists should be most on their guard against this
snare .

It is not only primitive man , the child of nature, who

regards himself as thc mcasure and model of all things,
and fills the sky and air with beings like unto him
self. Science not exempt from this crror. Take
the genealogical table of language as described and
even drawn in many books. Is it not a product of the
| purest anthropomorphism
there has ? What
been on the difference a deal
between of writing
mother tongues
and daughter tongues ! But tongues have no daughters,
neither do they give birth to dialects. To speak of
proto -Hellenic or proto -Aryan is to borrow habits of
thought from another order of ideas, to force upon the
Science of Language the hypothesis of zoology. The
same folly is secn in that pro -ethnic Indo -European
language which many tireless philologists construct and
re- construct, just as, to account for their different races,
the Greeks imagined Ancestors - Aeolus, Dorus, Ion, and
Achaeus, sons or grandsons of Hellen ."
There arc few books which, in small compass, contain
so many paradoxes as the little volume in which
Schleicher gives his ideas on the origin and development
of languages. Though , being a botanist and Darwinian,
he usually keeps his mind clear and methodical, he
betrays in this work habits of thought appropriate to
" I would draw the attention of my readers to the recent work
of M. Victor Henry, who combats the same error from a different
point of view - Antinomies linguistiques.
 PURPOSE OF THE BOOK 3

some disciple of the mystics. For instance, he places

the epoch of the perfection of languages in the remote
past, before all history. As soon as a people makes
its entry into history (he says ), and begins to have a
literature , decadence, irreparable decadence, appears.
Language, in fact, is developed inversely to the progress
of mind. A noteworthy example of the power which
first impressions and ideas received in childhood may
exercise ! 1
Leaving aside the phonetic changes which belong to
physiological grammar, I propose to study the intel
lectual causes which have influenced the transformation
of our languages. To make this research methodical, I
have ranged the facts under a certain number of laws,
though readers will see later on what I mean by law ,
an expression that must not be taken in the imperative
sense. These do not belong to the blind laws without
exceptions, among which, if we may believe some of
our colleagues, are the laws of phonetics. · I have taken
care, on the contrary, to mark the limits of each. I
have shown that the history of Language, by the side of
changes pursued with a rare logic, displays also many a
tentative essay, sketched out , and left half-finished .
An onward march in a straight line, with neither turn
nor winding, would be a new experience in human
affairs. On the contrary, human achicvements appear
to us as laborious things, thwarted without ceasing by
the survivals of a past that cannot be annullcd, by
| Schleicher was originally destined to the ecclesiastical pro
fession . He afterwards became a Hegelian.
6 SEMANTICS

collateral enterprises conce:ived in a different spirit, or

even by the unlooked for results of the attempts actually
in progress.
I have at last decided to publish this book, which
I have hitherto abandoned as often as I have begun
Extracts 1 from it have appeared at various times in the
form of essays, but again and again, repelled by the
difficulty of the subject, I have vowed never to return
to the book itself. And yet this long incubation has
not been without use. I am sure that I sce more
clearly to-day into the development of Language than
I did thirty years ago. My progress has lain in setting
aside all secondary causes, and in appealing directly
to the only true causes, which are human intelligence
and will .
To permit will to intervene in the history of Language
seems almost a heresy, so carefully has it been banished
and excluded for forty years. But though men were
justified in renouncing the puerilities of the ancient
science, they contented themselves with an unduly
simple psychology when they plunged into the opposite
extreme. Between the actions of a consciously de
liberate will, and the purely instinctive phenomenon,
there is room for many intermediary states ; and our
philologists would have profited ill by the lessons of
contemporary philosophy did they still impose upon us
a choice between the two horns of this dilemma. You
1
In my Mélanges de mythologie et de linguistique, in the
Annuaire de l'Association des études grecques, in the Mémoires de
la Société de linguistique, in the Journal des Savants, etc.
 PURPOSE OF THE BOOK 7

must close your eyes to evidence not to see that a

will, dim but persistent, presides over the changes of
Language.
How should this will be represented ?
I think that it should be represented under the form
of thousands, of millions, of billions of furtive attempts,
for the most part unfortunate, sometimes attended by a
partial success, attempts which, thus guided, thus cor
rected, thus made perfect, attain to definiteness in some
specified direction. The goal of Language is to be
understood. The child, for many months, exercises his
tongue in uttering vowels, in articulating consonants.
How many failures precede the clear pronunciation of
a single syllable ! Grammatical innovations are of a
kindred nature, with the difference, that in them a
whole people collaborates. There must, in truth , be
a thousand unskilful, incorrect, obscure constructions
before that one is found which is an expression of
thought, not perfect of such there are none) but at
least sufficient. In this long labour there is nothing
that cannot be traced to the will.1
Such is the study to which I invite my readers. They
must not expect to find any very complicated facts of
nature. As always where the spirit of the people is in
question, we are surprised by the simplicity of the
I " A breath,” says Herder somewhere, “ becomes the painting of
the world, the picture of our ideas and of our feelings ! ” This is a
presentment of things from the standpoint of a philosopher in love
with mystery. There was more truth in the picture drawn by
Lucretius. Centuries have been needed and countless efforts for
this breath to produce one clearly formulated thought !
8 SEMANTICS

means, a simplicity in strong contrast to the extent and

importance of the results obtained.
I have designedly drawn my examples from the most
generally known languages: it will be as easy to increase
the number, as to produce others from less widely ex
plored regions. Since the laws which I have tried to
indicate are chiefly of the psychological order, I do
not doubt that they will hold good outside the Indo
European family. My object has been to trace out a
few great lines, to mark a few divisions, and to sketch
a provisional plan in a domain which has not yet been
exploited, and demands the combined labour of several
generations of philologists. I therefore beg the reader
to look upon this book as a mere Introduction to
the science that I propose to name Semantics."
1 Enuavtiki) réxvn, the Science of Significations, from the verb
onpaivw, to " signify," in opposition to Phonetics, the science of
sounds.
 PART 1

THE INTELLECTUAL LAWS OF LANGUAGE

 CHAPTER I
THE LAW OF SPECIALISATION

Definition of the word law - False idea prevalent on the subject

of languages known as synthetic and analytic - Specialisation of
function one of the characteristics of analytic languages.
We define law , using the word in the philosophic
sense, as the constant relation discoverable in a series 1
of phenomena. One or two examples will make this
clearer.
If all the changes that take place in the government
and habits of a people tend towards centralisation, we
say that centralisation is the law of the government and
habits of that people. If the literature and art of an
epoch are distinguished for the qualities of order and
moderation , we say that order and moderation are the
law of the art and literature of that epoch. In the same
way , if the grammar of a language tends consistently
towards simplification , we can say that simplification is
the law of the grammar of that language. And, to
come to our point, if certain modifications of thought,
expressed primarily by all words, are little by little
restricted to a small number of words, or even to a
II
12 SEMANTICS

single word , which takes upon itself alone the whole

function, we say that Specialisation is the law that has
presided over these changes. There can be no question
of a preconcerted law, still less of a law imposed in the
name of a higher authority,

Everybody is aware of the distinction between lan

guages known as synthetic and languages known as
analytic. Repetition has reduced it to a commonplace.
Everybody also can tell more or less fully wherein the
difference consists. But as to how this evolution was
brought about, and by what causes, the vaguest and
most inaccurate ideas still prevail.
No one has expressed better than J. J. Ampère
the connection between the Latin and the Romance
languages as popularly set forth. The book is justly
criticised, but none the less represents the ideas of the
majority even at the present day.
" The ancient grammatical synthesis,” he writes, “ in
virtue of which a dying language was organised, is
destroyed ; the grammatical infections are lost ; the cases
of the nouns, the tenses of the verbs, are no longer
sufficiently distinguished. How can this confusion be
simplified ? The relations conveyed by the grammatical
signs now confused or abolished are expressed by
separate words ; prepositions are made to supply the
place of the terminations which denoted the cases of
substantives ; auxiliaries are substituted for those that
marked the tenses of verbs, Genders are indicated by
articles, and persons by pronouns. . . . In all languages
 THE LAW OF SPECIALISATION 13

men have employed the same remedy for the same evil ;
in the same distress they have had recourse to the same
expedient."» 1

So, it is to restore ruins, to remedy an evil, to emerge

from confusion, that the new processes have been
invented. Such a statement of thc casc (and the same
idca still exists, I repcat, in the minds of most philolo
gists, even of those who most severely criticised the
book ) confuses the true succession of facts, and renders
unintelligible the history of languages. In reality there
was no restoration of ruins, for the terminations that
were set aside had long ago become useless. The
ancient languages were never in distress. The time has
come to write a simple and a true history in the place of
these improbabilities, and at the beginning of this history
should be placed The Law of Specialisation.
One tendency of the mind, explained by the need for
clearness, is the substitution of invariable independent
exponents for exponents that are variable and depend
ent. This is a tendency consistent with the general
aim of Language, which is to make itself understood at
slightest cost, I mean with the least possible amount of
trouble. But as the conditions in which Language is
placed do not admit of creation ex nihilo, this effort is
realised but slowly, by the means and at the expense of
that which previously existed.

A first and tangible example is furnished by the

comparative and the superlative.
Histoire de la langue franquise, and edition, p./3, 10.
 14 SEMANTICS

In ancient languages the adjective expresses degree

by means of suffixes. These suffixes were originally
numerous and diverse. Thus the comparative could be
marked by the syllables -ro ( superus, inferus), -tero
( interus, exterus), -ior ( purior, largior). The superlative
could be marked by the syllables -mo (summus, infimus),
-timo ( intimus, extimus), -issimo (dulcissimus). Latin, as
we know it, has already abandoned this diversity,
keeping for each degree one suffix alone ( -ior, issimus ).
Here is the first step towards simplicity.
If from Latin we pass to French, we still find a few
comparatives in the ancient manner, an inheritance from
the Latin : graignor, forçor, hauçor, juvenor, gencior.'
There are also some superlatives : pesme ( pessimus),
proisme ( proximus). But this mechanism, already
deprived of its true mcaning, rapidly disappears, not ,
as has been said, in consequence of phonetic change (for
these words were perfectly practicable), but by the action
of the law of Specialisation. One single word assumes
in French the function of all thcsc comparatives and
superlatives. So also in the other Rornance languages.
In French , plus ; in Italian, piis ; in Spanish, mas ; in
Portuguese, mais ; in Roumanian, mai.
But what must be noted is that this privileged word
which succeeds to all the comparatives of former days
is itself a comparative. Plus represents the ancient
Latin ploins ( = Greek Telov) ; the Spanish mas, the
.
Portuguese mais represent magis. It is therefore the
last survival of a species that is extinct, and extinct not
Comparative ofgrand,fort, haut, jeune, gent.
 THE LAW OF SPECIALISATION IS

without intention, and it fills by itself alone the place of

all the others. The only exceptions are some compara
tives such as meilleur, pire, moindre. These possess the
advantage of relative brevity and are of such frequent
usage that the new process has not supplanted them.
We can see from this first example in what the law
of Specialisation consists. Among all words of a certain
kind, distinguished by a certain grammatical imprint,
there is always one which is little by little drawn apart
from its fellows. It becomes the pre-eminent exponent
of the grammatical conception of which it bears the stamp.
But at the same time it loses its individual value, and
is no more than a grammatical instrument, one of the
wheels of the phrase. When the French say un temps
plus long, unejournée plus courte, thc word plus serves to
determine the adjective by which it is followed ; but by
itself it has no more existence from a semantic point of
view than the inflection -ior. In the same way we can
guess the reason why the law of Specialisation demands
the help of centuries before it can operate. Words have
too much individual significance to adapt themselves '
at once to playing the part of auxiliary. It is only a '
long use of words in various connections that slowly
prepares the mind to remove from them their superfluous
value.

1 That does not prevent the continued usage of the word plus, in
the sense of recov, in its fullest and widest signification. Ex.
" En voulez-vous plus ? -- Qui peut le plus peut le moins." We
shall find later on numerous examples of this segmentation of
meanings. It is curious to note that pronunciation has up to a
certain - raint differentiated these two plus. 1 ?
16 SEMANTICS

It is not therefore, as is often said, the collapse of inflec

tions, that has brought about the use, as a sort of make
shift, of plusand magis ; that use dates from a time when
inflections were still commonly employed. There are
even examples of the cumulative use of the two pro
cesses : Plautus writes, Magis dulcius, magis facilius,
mollior magis. Thcsc cxamples show the compara
tive idea beginning already to find its definite domicile
in one particular adverb, although the mechanism -ior,
-issimus is still in full forcc.

We come now to the substitution of prepositions for

the ancient declensions.
It is well known that at first every substantive marked
the relations of dependence, of interiority, of instrument,
etc., by modifying its final syllables. But this method
of expression was at once complicated and insufficient.
It was complicated , because substantives, not being
all constructed alike, appeared under different forms in
onc and the same case ( genitive : domini, rose , arboris,
etc.). It was insufficient because the cases of the
declensions were too few in number to express all the
relations that the mind was capable of conceiving.1
This was the reason why adverbs were placed by the
side of these cases to define them. But the habit of
placing the same adverb by the side of the same case
could not fail in the long run to produce upon the minds
· The cases of the declensions indicated indeed the place whither
onc gocs, whenco one comes, in which one is. But there was no
inflection to express " across," " upon ," " with," “ round ," etc.
 THE LAW OF SPECIALISATION 17

of men an effect of which we shall presently have other

examples : between the inflection and the particle of
place or time they supposed the existence of some special
connection, some relation of cause to effect. Instead of
considering the adverb as a mere determinant of the
case, popular intelligence saw in it the actual cause of
the case ; a well-known paralogism that philosophy
designates by the formula cum hoc, ergo propter hoc. But
when a paralogism is universal, it is undoubtedly not
far from giving the impression of a truth . In a matter
of Language, that which the people believes itself
to feel passes into the condition of reality. Adverbs
of place and time like από, περί, επί, πρός , μετά , παρά,
from having been the accompaniment of the genitive,
dative, or accusative, became the cause of these cases :
from having been adverbs, they became prepositions.
The minds of men endowed them with a Transitive
Force .
In the Homeric tongue, the transformation is already
three- quarters accomplished . It is wholly accomplished
* In the Syntax of Delbrück will be found many examples of this
change of character, ancient adverbs becoming prepositions. But
I differ in opinion from the author of the Grundriss on the order
and connection of the facts.
2 In this fragment of a phrase : Bledupws and Jáspvor jey (2
palpebris lacrimam demisit )úno accompanies rather than governs
the genitive. It is the same with me and the dative : oloty & à Zivg
Bijke kako pópor (quibus Jupiter impossit malam sortem ). Or with
the accusative and mepé : woow more pápa muxrog circipiros lorepurwras
(insulam quam circum pontos innitus muii ). It might also be
easily imagined that, in these examples, the particle of place
determines the verbs.
18 SEMANTICS

in the most ancient documents by which the Latin

language has been preserved for us. In the Vedic texts,
on the contrary, we find words which have since become
well-known prepositions, still in the condition of adverbs :
per, ob, ad, sub, super, ab.
The existence of the declension is threatened from
the moment that a language possesses prepositions.
What, indeed, is the use of these cases that add nothing
to meaning ? Is the preposition not enough ? The pre
position is not only perfectly sufficient, but also fulfils a
more important function , for it marks precisely and
explicitly the relations indicated by the inflection in a
vaguc and general manner. It is morcover more con
vcnicnt to use, for it is always the samc, always casily
rccognisablc. Yct, as nothing comes about quickly
when time-honoured habits, common to large masscs of
men, are in question, the inflections do not disappear at
one blow, once and for all. They begin by growing
vague. They are used loosely, they are confused one
with another .
The first symptoms of this transformation go back
much farther than is usually believed. The passage
from Suetonius is often quoted, in which, speaking of
the habits of Augustus, he reports that the Emperor, for
the sake of greater clearness, did not hesitate to add
prepositions to nouns, and conjunctions to verbs. The
passage is curious in itself. But the last words especi
ally must be noted ( præpositiones) quæ detractæ affer
unt aliquid obscuritatis, ctsi gratiam augent.' It was
1
Life' of Octavius Augustus, 86.
 THE LAW OF SPECIALISATION 19

considered elegant and well- bred to dispense with

prepositions and conjunctions. This was the ancient
Latin language. But the Emperor, who as we know
liked to affect rustic habits, adopted the new custom .
of this rustic speech we have another contemporary
witness. This is the dedication and regulations of a
temple of Sabina, in the year 57 B.C. These regula
tions provide for the event of donations being made to
the temple : Si pecunia ad id templum data erit. ...
Quod ad eam ædem donum datum erit. ... Instead of
the dative, we here find the modern construction : " To !
this temple.” We must note that this is an official
document, at once legal and religious. Official language
prefers to be archaic, if nothing is thereby lost in matter
of precision : but from the moment that precision is at
stake, it does not shrink from neologism .
Not long after Augustus, we already witness the
decadence of case - inflections. At Pompeii, we find :
cum discentes, “ with his pupils ; ” cum collegas, " with
his colleagues." In an inscription of Misenum of the
year 159 A.D. is written : per multo tempore. In another
of almost the same period : ex litteras. The Latin of
Africa, ever since the epoch of Hadrian , displays this
kind of defect. An engineer of Lambes , though know
ing his language well, makes a mistake on a like point:
and writes : a rigorem , sine curam .'
If we come down to two centuries later, we find the
Corpus Inscriptionneme Latinarum , ix. 3513 .
* C.I.L., viii. 19,570 ; X. 3344 .
* Boissier, Journal des Savants, 1896, p . 503.
70 SEMANTICS

use of inflections more and more uncertain , that of

prepositions more and more frequent. In the Pilgrimage
of Silvia ( fourth century ) are found expressions like
these : Fundamenta de habitationibus ipsorum .
fallere vos super hanc rem non possum valde instruc
tus de scripturis. . . . And even : Lecto omnia de libro
Moysi, “ having read all of the book of Moses. " TO
gether with the Latin prepositions, we find the Greek
preposition kard . Cata singulos hymnosfit oratio ."
M. Max Bonnet, in his book on Le Latin de Grégoire
de Tours, draws attention to the fact that Grégoire is
mistaken in his usage of cases when preceded by a prepo
sition . Not that he is unaware of the Latin declension,
and of the value of cach casc . But when he employs
one of the propositions cum , de, ad, per, in, sub, it is a
matter of indifference to him whether he uscs with it the
accusative or the ablative.
It was not, therefore, through ignorance, through
the destruction of forms, or through the impossibility of
agreement, that, once the declension was in ruins, the
" It is well known that this preposition passed afterwards into
the Romance languages : Spanish, cada uno ; Italian, caduno ; old
French, chaün, cheün.
P. 522. Speaking of the confusion of cases, M. Bonnet says :
“ It is allowable to doubt whether economy of forms has had much
to do with the matter. It must not be forgotten, for instance, that
if the accusative singular is generally distinguished from the abla
tive only by an m, which was probably pronounced with difficulty,
it is quite different with the plural and singular neuter in the third
declension. Here the inflections as and is, os and is, es and ibus,
es and ebus, us and ibus, us and ore, en and ine, etc.—had kept
their sounds perfectly distinct. So much help was not needed for
the distir.ction of cases.
 THE LAW OF
SPEČIALISATION
people had recourse in despair to another method of
21

representing the same ideas. No : it is at the summit

of Roman hierarchy, in the most splendid moment of its
literature, that we find the first examples. The language
of ordinary affairs must have been the first to welcome
the innovation, and thus prepared the way for a new
grammatical system .
The most important fact in the history of our
languages, that which permanently characteriscs the
transition from synthesis to analysis, comes, therefore,
under the heading of the principle of Specialisation.
Yet there is one use of the cases to which prepositions
afforded no help : that is their use to distinguish subject
and object. And it is accordingly the distinction of
nominative and accusative which has lasted longest. We
shall return to this when treating of construction .
In proportion as the oldadverbs changed into
prepositions, the custom prevailed of placing them
regularly before the substantive : I may be permitted to
make on this subject an observation which I think is
important. But that it sounds odd to speak in such a
way, I should say that our modern languages have
never had a happier chance, and have never escaped a
greater danger than when Latin had the sense to change
into prepositions little words such as in , ad, per, cum ,
which up to that time had been habitually joined to
their object in the form of postpositions. Forms such
as mecum , tecum , vobiscum , semper, paulisper, quoad, still
bear witness to this condition, which has been aban
doned by Latin, but from which the brothers of Latin,
 SEMANTICS

Umbrian and Oscan, have never succeeded in emerging.

In Umbrian, for example, not only cum , but also in,
ad, per, and all ancient expressions of that kind,
have remained postpositions. “To the altar, towards
the altar, upon the altar” are expressed by asacuin ,
asamen , asamad, and, owing to carelessness of pronunci
ation, by asaco, asame, asama. " To the limit, towards
the limit, on the limit " are expressed by terinnuco,
termnume, termnuma. And so on. As early as the first
century before the Christian era, we see by the defects
which make their appearance that confusion is setting
in. Between the substantive and the little word by
which it is followed vicious connections are formed . If
Latin had not been diverted from this path, its declen
sion would have taken a very different turn. Instead of
becoming impoverished, it would have been enriched,
for new cases would have been formed. Instead of
ending in the Romance languages, Latin would have
ended in some such idiom as Basquc.
Owing to a just recognition of the demands of clear
ness, modern languages have become more and more
rigorous on this point. They have insisted that nothing
.

should come between the preposition and its " object " ;
and while Latin still tolerates certain intercalations,
French admits of no exceptions to this rule.1

But in no language are the effects of the principle of

Specialisation seen so clearly as in English English
We cannot accordingly approve of the new fashion, established
within the last few years, on the subject of the preposition avec.
 THE LAW OF SPECIALISATION 23 .

has not abandoned its genitive : but it has put the

exponent of the genitive to so bold a use that it
obtains therefrom the same services as if it were an
independent word. Having adopted a simple s as a
uniform inflection for all substantives, it has made this s
moveable, so that it may be placed after two or more
substantives. The Queen of Great Britain's Navy. Pope
and Addison's Age. In this way English has provided
for itself two varieties of genitive, the one with s, the 3
other with of, the one progressive, the other retrogressive,
a curious example which shows how, by becoming
supple, the mechanism of a language can be perfected,
and its resources widened.1
The English conjugation will give us another example
of the law of Specialisation.
Without doubt English is the most analytical of
modern languages. It has often been said that this
analytical character was due to the mingling of Anglo
Saxon with French grammar ; an explanation which, in
this form , is inaccurate. What is true is that the upper
classes of society, in making use of French for many
centuries, had abandoned English to the lower classes.
Now, as we have just seen, it is the cultured portion +
of a nation that retards the evolution of a language.
1 There is, as Mr. Jespersen remarks, a certain mathematical
elegance in the substitution of a single letter for the varied inflec
tions of Latin. But it cannot be doubted that the ancients took
pleasure in this variety : it was to them as a series of musical
chords, the resonance and sequence of which it gave them satis
faction to hear. Language has cast off this soinewhat childish
luxury .
24 SEMANTICS

When aristocracies take no interest in the national

tongue, this evolution progresses at a greater rate.
The Germanic conjugation, with its complicated rules
that present so great a difficulty to the foreigner, is not
wholly easy even for the native. Jacob Grimm counts
in German as many as twelve classes of conjugation, of
which more or less well-preserved specimens appear
equally in English, such verbs, for instance, as I give, I
gave ; I bind, I bound ; I dig , I dug ; I hold, I held, etc.
We know how modern English overcomes this diffi
culty : in the stead and place of these present and
preterite tenses with their manifold formations, it makes
use, or at least is at liberty to make use, of the present
I do, of the preterite I did, turning the verb into an
invariable word. The change began with the interroga
tive and negative turns of phrase. Then the verb do,
continuing its progress, introduced itself into merely
affirmative sentences. Suppose by a fresh step in
advance it imposes itself on affirmative phrases and so
becomes of constant and obligatory usc, English will
have substituted its auxiliary verb for all other verbs.
This auxiliary verb will then express the conceptions of
time, of person, of mood, as well as that of affirmation,
which every verb had up to that time expressed on its
own account. From now onwards the verb do is so
ready for the most various utility that it can serve as
auxiliary to itself.
But universality of usage has its reverse side. When
do accompanies another verb, it is nothing more than
a grammatical tool. By a division, that would seem
 n 2
2 g
THE LAW OF SPECIALISATION 25

extremely subtle had it been made deliberately and

from the first, English sets on the one side the concrete
expression of the act, and on the other side the concep
tions of affirmation , person , time and mood . In a
dialogue such as this : Does,he consent ? He doesn't, all
the movement of the action, all the grammatical appara
tus, is concentrated in the auxiliary.
But it is rare for the principle of Specialisation to
triumph at once. The history of languages is sown with
abortive attempts and half-successes. Centuries before
the verb do was turned into an auxiliary verb, it had
already been once employed to remedy certain diffi
culties of conjugation . It had been found easiest, in
order to form the perfect of certain verbs, to borrow the
perfect of the verb do. In Gothic the loan was most
obvious : ski-da, " I searched," sôki-dédum ," we searched . "
This is well known to be the origin of the perfect
known as “ weak ." The attempt only half succeeded.
It made a mistake in appearing in an epoch of synthesis.
The auxiliary was united to the principal verb and
formed with it an indissoluble whole, in such wise that
the Germanic conjugation, instead of being simplified,
had to reckon with an additional series of forms.
We can liken to this the fate of the future and con
ditional tenses in the Romance languages. We know
that these languages have found in the verb habere an
exponent as simple as it is convenient. Ovid writes in
his Letters from Pontus :

“ Plura quidem mandare tibi, si quæris, habebam :

Sed timeo tarda causa fuisse more ."
26 SEMANTICS
We have here the beginning of the modern conditional,
And in the following phrase taken from a sermon of St.
Augustine, we find the beginning of the future tense ; it
is the end of the world that is in question : Petant aut
non petant, venire habct. But the auxiliary having
welded itself on to the principal verb , the attempt
miscarried, at least from the point of view of the prin
ciple of Specialisation. If we go back half- a -score of
centurics, we find absolutely similar attempts in im
perfects such as amabam , in futures such as amabo, in
perfects such as amavi and as duc- si. In this case
verbs signifying “ to be” in Sanscrit bhi and as, in
Latin fuo and esse) join themselves on to the principal
verb. But, cast into the middle of a synthetic
conjugation, these auxiliaries are at once absorbed .
Finally we discover a first attempt as early as the
Indo -European period. The future (in Greek dúow, in
Sanscrit dāsjāmi) composed with the auxiliary as,
together with the other tenses composed with the same
auxiliary, are attempts which show us how often
Language has had recourse to the same means, before
realising at last the progress that it had in view ,
 CHAPTER II
THE LAW OF DIFFERENTIATION

Proofs of the exis.:ence of Differentiation - Limits of the principle

of Differentiation .

We define " Differentiation " as the intentional,ordered

process by which words, apparently synonymous, and
once synonyms, have nevertheless taken different mean
ings, and can no longer be used indiscriminately.
Has there been such Differentiation ? The majority
of philologists deny it. When they find themselves in
the presence of facts that are too obvious to be neglected,
they declare that these facts do not count : that it is a
scholar's Differentiation, and in no wise a popular one,
by which they are confronted . Thus they betray the
same lack of psychological analysis, that we 'noted at
the beginning, which only admits the intervention of
the human will, if this will has been conscious and
deliberate.
I will first draw attention to the fact that the people
is not of this opinion. It admits the existence of such
Differentiation ; it does not believe that there exist in
27
28 SEMANTICS

Language absolutely identical terms.' Feeling strongly

that Language is intended to serve for the exchange of
ideas, for the expression of feelings, for the discussion
of interests, it refuses to bclieve in a synonymy that would
be both useless and dangerous. Now as the people is
at once the depositary and author of Language, the
fact that it denies the existence of synonyms actually
brings about their speedy disappearance : either they are
differentiated, or else one of the two terms ceases to
exist.
All the discredit that has been thrown on this subject
arises from the distinctions attempted in private by self
constituted would -be doctors of Language. There are
no good distinctions, excepting those that come about
without premeditation, under the pressure ofcircumstance,
by sudden inspiration, and in face of real need, through
men concerned with the things themselves. The dis
tinctions that are made by the people are the only true
and the only good oncs. No sooner docs it see things,
than it associates words with them. We will give
some examples.
Whenever two languages or even two dialects find
themselves face to face, a process of classification takes
place, which consists in attributing degrees to synonymous
expressions. According as an idiom is considered
superior or inferior, these terms are seen to increase or
diminish in dignity. The question of the Science of
Language is at bottom a social or national question. M.
· From this arises the question so often heard : What is the
difference ... ?
 THE LAW OF DIFFERENTIATION 29

J. Gilliéron describes the effects produced through the in

vasion of a Swiss dialect by French. In proportion as a
French word is adopted, the Patois vocable, degraded and
driven back , becomes vulgar and trivial. Formerly a room
was called pailé : since the word chambre has come into है
the village, pailé means a garret. In Brittany, says the
Abbé Rousselot, gardens were formerly called courtils :
now that the word jardin is known, shade of con
tempt adheres to the rustic word. It matters little that
the two terms had the same origin . The Savoyard uses
the names of père and mère for his parents, while he
kecps for his cattle the old words påre and maré. Among
the Romans, coquina signified “ kitchen ” : the Oscan
popina, which is the same word, meant a common tavern.
It will perhaps be said that these words are naturally
differentiated by the things they designate, and that they
have never been compared. This would be to maintain
that the popular mind is incapable of percciving two
things at once. I think on the contrary that there has
been comparison, and that the popular term owes to
this comparison a downfall which would otherwise be
incomprehensible. In the matter of Language, signifi
cance is the great regulator of the memory ; in order to
fix themselves in our minds, new words need to be
associated with some word of kindred sense. The people
therefore has its synonyms , which it arranges and
subordinates according to its own ideas . In proportion
as new words are earnt, these are inserted among the

· Le patois de la Commune ue Vionnaz ( Bas- Valais), in the Biblio

thèque de l'École des hautes études, 1880.
30 SEMANTICS

words that are already known. It is not surprising that

these latter suffer a certain depreciation, a recoil. So
long as populations mix with cach other, there will be
fresh examples of Differentiation. To prevent this, you
must put a customs-duty on Language or employ the
closure.
What the people does by instinct is done with the
same spontaneity by all growing science, by all searching
analysis, by every discussion that desires to attain an
end, by every opinion that wishes to be recognised and
defined . Plato, wishing to combat the ideas of the
Ionian school, reproaches Thales with having confounded
the principles or åpxal with the elements or otoixeia..
Now, the elements are water, fire, earth , air, while the
principles are things of a more general and imperishable
nature, such as numbers. The distinction thus made by
the Greek thinker, though philosophical and profound,
is of the samc kind as the distinctions quoted above, at
Icast from the point of view of the Science of Language.
By an immcdiatc apperception, the two words, till then
synonymous, were differentiated. Are we to place facts
of this naturc outside the history of Language ? In so
doing we should run the risk of restricting it on its most
important side. The history of Language is a series
of diffcrentiations. That, and that alone, took place
at the birth of languages. That, and that alone, takes
place at the first lispings of a child ; for it is by Differ
entiation that he applies little by little to distinct
objects the syllables which he at first scatters imparti
ally upon everything that he meets.
 THE LAW OF DIFFERENTIATION 31

Let us now turn to some effects of Differentiation in

an ancient epoch of our languages.
The root man scems to have served in the beginning
to designatc vaguely all the operations of the soul; for
we find it expressing thought (mens), memory (memini,
μέμνημαι, μιμνήσκω), passion (μένος ), and even perhaps
madness (uavla ).' But a less rudimentary psychology
introduced some order into this confusion , keeping
certain words, evicting others to be replaced by synonyms,
giving in short to each one its special domain. Such a
siſting was not done by chance. Rather might we here
recapitulate, with especial force on a purely human and
historical ground, the whole argument of Fenelon .
We are in the habit of drawing a distinction between
the active courage which goes out to meet danger and
to fight with it, and the passivc courage which bears
cvil fortunc with an cquablc mind. Although capable
of existing in the same man , these are, at bottom, two
distinct qualities, as can be seen by observing the ex
tremes to which cach are led by exaggeration. Active
courage pushed too far ends in foolhardiness ; passive
courage carried beyond due measure degenerates into
apathy.
We might have expected that Language would have
registered so natural a distinction from the very first ;
but it was not so. In the language of Homer, the two
conceptions secm to be confounded with one another,
and the same verb toluow, which mcans to dare means
also to bear ; the same adjective tiņuwv, which means
" A. Meillet, De Indo- Europæa radice men , Paris, Bouillon, 1897.
32 SEMANTICS

patient means also bold . After Homer, gnomic poetry

furnishes us with other examples of this confusion
" It is necessary to bear that which the gods send to
mortals," says a proverb .
Τολμάν χρή τα διδούσι θεοί θνητοίσι βροτοίσιν .
And elsewhere— " Be enduring in misfortune, O my
soul, even when thou sufferest uncndurable things."
Τόλμα , θυμε, κακoίσιν , όμως άτλητα πεπονθώς.
It was therefore by a tardy distinction that boldness
(and boldness pushed even to the verge of rashness and
insolence) was confided to toluéw and its family, while
constancy and resignation fell to the lot of Tálas and of
τλήμων.3
No one would now dream of expressing by the same
word two idcas so different as thc plcasure of the senses
and the ideal plcasure caused by the subjective fceling
of hope. Yet there was a time when the same expres
sion served for the two ideas. The Greek tongue drew
from this root a series of words expressing hope-eATTIS,
ελπίζω, ελπομαι. Latin took from it the words that
denote pleasure - volupe, voluptas.* The idea left un
represented found, on either side, other symbols : idor
(from 18opar " to enjoy ") became in Greek the name for
I ., xx . 19 ; Od. xxiv. 162, etc.
· Theognis, v. 591 , 1,029.
In modern languages, the root tol contained in torudw, served
in German to designate patience (Ge-dul.a ). It reappears also
the Latin tolerare.
.

• The verb 'EATW began with a v orſ, as is seen by the perfect

"Eolta ( for Féfonra).
 THE LAW OF DIFFERENTIATION 33

pleasure, and in Latin spes, " respiration, relief," the

name for hope.
It is in this way that, going back into the past,
one finds on the way semantic conglomerates which
have needed centuries to distinguish . And the process
is not finished even to -day. In French the difference
between sentir (to feel) and penser ( to think) is now
marked in the verbs, but hardly appears in the sub
stantive sentiment. Similarly, the adjcctive sensible
which in French bclongs to the soul, has in English
taken the meaning of " intelligent, reasonablc." It is
well known that in Latin sentir pertains rather to
thought, as is seen by compounds such as dissentio,
consentio, and by derivatives such as sententia.
Owing to a confusion that has not yet wholly dis
appeared, ancient languages expressed by the same word
the " wicked " and the “ unhappy." The adjective
Trovnpós has the double meaning. In the childhood of
societies, the poor were objects of aversion as much as
of pity : it is in this tone that Homer speaks of beggars.
Little by little trompós abandoned this double meaning,
and was exclusively applied to the idea of perversity,
while its congener névns designated the pauper.
The nearer words approximate in form , the more do >
they invite Differentiation . Here is a sentence, at first
sight extraordinary cnough, that has been preserved for
us by Varro - Religentem esse oportet, religiosum nefas.
1 Πονηρά ιππάρια, πονηρόν όψον, δωρ. Πονηρά πράγματα. From the
. same root that has supplied wovos, " trouble " ; mevla, “ poverty " ;
rivouai, " to be a pauper." Cf. the double sense of méchant in
French .
D
 34 SEMANTICS

The two words religens and religiosus, etymologically

synonymous, are opposed to each other. The meaning
of the proverb is that religion is a good thing, but that
superstition is not. There is a certain elegance, to
which the people is by no means insensible, in thus
differentiating words that have almost the same
sound.
The needs of thought are the first agent of Differenti
ation. Thus it was that Greek and German met in
making the distinction between Mann and Mensch,
between ανήρ and άνθρωπος..
Between årp and dvopwntos there was originally no
3 difference of meaning : the one signified " man ,” the
other “ he who has the face of a man .” Homer, speak
ing of the Ethiopians who inhabit the ends of the earth ,
calls them čoyatou åvopôv. But, owing to an antithesis,
the opportunity for which could not fail to arise, they
have little by little been distinguished one from the
other, and placed in opposition. Herodotus, speaking
of the army of the Persians, says that at Thermopylæ
Xerxes could perceive ότι πολλοί μεν άνθρωποι είεν , ολίγοι
dè åvopes. The distinction became in this way familiar
to the Greeks. Xenophon, speaking of the love of
glory which constitutes the worth of life, adds that by
this men recognise each other : άνδρες και ουκέτι άνθρωποι
Móvov voulcóuevot. Nothing, either in the etymological
meaning of ανήρ, or in that of άνθρωπος, predestined
them to this opposition .
We shall come back to this point in the chapter on Analogy.
the adjective (& v@pwaos having first been an adjective) that
 THE LAW OF DIFFERENTIATION 35

When the popular mind has once devised a certain

kind of Differentiation , it is naturally tempted to
complete the series. It is well known that there are X Х
languages in which the various acts of life are not
designated in the same way if an exalted personage be
concerned, as when the ordinary man is in question.
The Cambodgians do not designate the members of the
body nor the daily operations of life by the same terms
when speaking of the king, as when speaking of a mere
individual. To express that a man cats, they used the
word si ; in speaking of a chicf, they would say pisa ;
if they spoke of a bonze or of a king, it would be soi.
In speaking to an inferior, “ me” is expressed by anh ;
to a superior, by knhom ; to a bonze, by chhan. The
followers of Zoroaster, who regarded the world as
divided between two opposing powers, had a doubie
vocabulary, according to whether they spoke of a
crcature of Ormuzd, or of a creature of Ahriman. In
these cxamples we sce Differentiation leaving an impress
more or less profound ; as one observes certain habits
of mind, barely noticeable in one man, governing the
whole life of another.
Nothing, in truth, is more natural or more necessary
than Differentiation, since our intelligence gathers in
takes, the most general signification. It is the same with Mann
and itensch. It is the same also in French with les hommes and
les humains (men and human beings).
' There is something of the same kind in English, but only in
1 rudimentary state. To mark the difference between men and
animals there are mouth and muzzle, nose and snout, etc. It is
ous that etymology has had nothing to do with this.
 S
36 SEMANTIC

words of different epochs and different surroundings,

and would be wholly given over to confusion if it did
not keep a certain order among them. What is done
by collections of synonyms is done by us all : when
terms distinguished or subordinated by common usage
are examined, it is seen that etymology rarely justifies
the differences that we attribute to them. Take for
example the words species and kind : for what reason
has a wider meaning been given to the former than to
the latter ? To branch than to class ? Takc thc words
division, brigade, regiment, battalion ; thcsc technical
terms, so cxactly subordinated cach to the others, have
no distinguishing feature that specially fits them for
this or that place. Perhaps we should find the same
thing were it possible for us to go back to the epoch
in which the series of the nouns of number was first
formed .
In passing from material to moral ideas, we shall
sce still better the effects of Differentiation. Between
esteem , respect, veneration, no gradation imposed by
etymology can be perceived. To establish certain dis
tinctions, there were needed exact and precise minds,
and a society which was well regulated and observant of
gradations : but that is no reason for Icaving these dis
tinctions outside the history of Language. Wc know
but little about the creation of Language ; but Differen
tiation is the truc demiurge thercof. Differentiation
has been that second crcation, that melior natura, of
which Ovid speaks when retracing the successive ages
of the world .
 THE LAW OF DIFFERENTIATION 37

Differentiation , however, like all the laws that we are

passing in review , has its limits.
Obviously it must first find a material in which
to work. As it does not crcate, but only attaches
itself to what is, in order to use and perfect it, the terms
to be differentiated must already exist in the language.
We could instance certain confusions, of which , for lack
of a word, even the most perfect idioms have never
succeeded in ridding themselves.
Inversely the mind docs not always manage to fer
tilise all the riches proffered to it by Language. Gram
matical mechanism, by the combination of existing
elements, can produce such a wealth of forms as to
cmbarrass the intelligence. Gcorges Curtius has counted
that the number of personal forms of the Greek verb }
amounts to two hundred and sixty -eight, a very con
siderable number, though still much inferior in quantity
to that of the Sanscrit verb which reaches to eight }
hundred and ninety -one. But Differentiation has not
been able to turn this abundance to account, though it
is already much that the Greek tongue should have
known how to differentiate its four preterites ( imperfect,
aorist , perfect, pluperfect ). Between the first and
second futures and between the first and second
perfects, the most attentive observation has failed to
discover any semantic distinction . Added to this over
production of tenses , we find an over-production of
verbs. If we take for example the root puy, " to flee,"
we find by the side of φεύγω a verb φυγγάνω, which
has the same meaning. By the side of onui we find
38 SEMANTICS

φάσκω. By the side of πίμπλημι, we find πλήθω. The

single verb signifying “ to extend ” is represented by
τεινω, τιταίνω and τανύω. We have βάινω , βίβημι, and
Báokw, all three of which signify “to walk .” The
Extinction of useless forms 1 fortunately diminishes the
weight of this dead capital.
Another limit to the principle of Differentiation is set
by the greater or less progress of civilisation . There
are shades of meaning which arise only among cul
tivated peoples. Through synonymy we learn with
what objects the thought of a nation has been chiefly
preoccupied. Distinctions are first made by a few
minds that are more subtle than others : then they
become the common property of all. Intellect, as has
been said, consists in seeing differences in similar things.
This intellect is communicated up to a certain point by
Language, for by recognising the distinctions, which the
most gifted alone perceived at first, the mental sight of
each individual becomes more piercing.
Onc question that concerns rather the philosopher
than the philologist is this : how does this Differentiation
come about in us ? Or, to put it somewhat crudely,
but intelligibly, have we in our heads a dictionary of
synonyms ? I think that in able and observant minds
this dictionary does exist, but that it only opens in case
of necessity and at the master's call. Sometimes the
right word springs forth at the first signal. At other
times it has to be waited for : then the latent dictionary
is opened, and displays successively the synonyms that
it holds in reserve, until the desired term be revealed.
1 See at the end of the first part.
 THE LAV

Differentiation ,
passing in review ,
Obviously it
to work. As it
itself to what is, is
to be differentiate
We could instand
of a word , even
succeeded in ridd
Inversely the
tilise all the riche
matical mocham
elements , can
embarrass the in
that the number
amounts to two
siderable number
to that of the
hundred and is
been able to tur
is already much
known how to d
korist, perfect,
second futures
perfects, the mo
discover anysen
production of te
Terbs. If we ta .
we find by the
has the same m
 40 SEMANTICS
From whence then has Latin taken it ? It comes
from such verbs as adolesco,floresco, senesco, etc. A man
does not grow up, flourish, or grow old in a moment;
and the idea of a slow and gradual action having thus
first been introduced into these verbs, it appeared after
x wards to be inherent in the suffix. It was irradiated
into it.
Something similar happened with the verbs known as
desideratives, such as esurio, nupturio, empturio. If they
follow the rare conjugation in -io, it is, in my opinion,
because they have modelled themselves on sitio, " to be
thirsty ." The syllable that precedes the inflection is
in spite of the difference of quantity - nothing but the
suffix -tor or -sor, which forms so many substantives in
Latin : emptor, " buyer " ; scriptor, " writer " ; esor ( for
ed -tor ), " eater . " ? The desiderative note has so thoroughly
entered into this inflection, that Cicero , writing to
Atticus, could say of Pompey with no fear of misunder
standing : sullaturit animus ejus et proscripturit.
We might here recall a discussion of the last century
that shows how casy it is to make a mistake on this
point : thc etymology - true or false - of an inflection is
more readily given than its birth and propagation are
traced . On the subject of these verbs in urire, the

" I rud , etc. In Homer, oww is added indifferently to all verbs.

See for example, Odyssey, xvii. 331, 335, xviii. 324, etc. This
same inflection is also to be found in Sanscrit, but not to any greater
degree in the inchoative sense.
i There is a difference of quantity, the suffix - tor having originally
had a long o or a short o, according to its position. Cf. in Greek porap
biropas.
 CHAPTER III

IRRADIATION

What is to be understood by this word - Irradiation can create

grammatical infections.
We designate by this term, for want of a better, a series
of facts that has not yet been named. To tell the truth ,
it has so far been barely noticed, though it is of real
importance for the psychology of Language.?
A few examples will make the subject clear.
Latin verbs in -sco, such as maturesco, marcesco, arc
commonly called “ inchoatives,” because they appear to
denote a beginning of action, or an action that comes }
about gradually. But this shade of meaning did not
belong originally to the infection -sco. It is not to
be found in nosco, “ I know " ; scisco, “ I decide ” ; pasco,
“ I nourish ," etc. Neither does it exist in cognate
languages.
1 We must however except the two American scholars, Mr.
Wheeler and Mr. Lanman, from whose works we shall quote further
on. M. Ludwig first drew attention to this question, under the
name of Adaptation.
? Cf. in Greek esplokw, " I find ," Titpuokw, “ I wound," 818pdoro,
39
 40 SEMANTICS

From whence then has Latin taken it ? It comes

from such verbs as adolesco, floresco, senesco, etc. A man
does not grow up, flourish, or grow old in a moment ;
and the idea of a slow and gradual action having thus
first been introduced into these verbs, it appeared after
x wards to be inherent in the suffix . It was irradiated
into it.
Something similar happened with the verbs known as
desideratives, such as esurio, nupturio, empturio. If they
follow the rare conjugation in -io, it is, in my opinion,
because they have modelled themselves on sitio, " to be
thirsty ." The syllable that precedes the inflection is
in spite of the difference of quantity - nothing but the
suffix -tor or -sor, which forms so many substantives in
Latin : emptor, “ buyer " ; scriptor, " writer " ; esor ( for
ed -tor )," eater." 1 The desiderative note has so thoroughly
entered into this inflection, that Cicero, writing to
Atticus, could say of Pompey with no fear of misunder
standing : sullaturit animus ejus et proscripturit.
We might here recall a discussion of the last century
that shows how casy it is to make a mistake on this
point : thc ctymology-true or false - of an inflection is
more readily given than its birth and propagation are
traced. On the subject of these verbs in urire, the

“ I run, etc. In Homer, okw is added indifferently to all verbs.

See for example, Odyssey, xvii. 331, 335, xviii. 324, etc. This
same inflection is also to be found in Sanscrit, but not to any greater
degree in the inchoative sense.
i There is a difference of quantity, the suffix - tor having originally
had a long o or a short o, according to its position. Cf. in Greek shtep
ρήτορος ..
 IRRADIATION 41

President de Brosses, in his Méchanique des Langues,

wrote : “ The Latin termination -urire is appropriated
to a keen and ardent desire to do something : micturire,
esurire, from which it appears that it was fundamentally
formed on the word urcre and on the root ur, which in
so many languages signifies fire. Thus the termination
-urire was well chosen to designate a burning desire."
Voltaire, clearer of vision, protests. Scenting one of
those theories to which the President was so prone, he
inakes objections. “Where is the idea of burning in
verbs such as scaturire, ' to well up ' ? ... This little
system is much in fault ; a fresh reason for distrusting
systems. "

There exists in Greek a group of verbs ending in law ,

that express a disease of the body or the soul : dovtiów ,
" to have toothache, ” from bots, “ tooth " ; onlnviów,
" to suffer in one's spleen," from otv, “ spleen " ;
λαρυγγιάω, « to have a sore throat," from ápuys,
" throat.” The sense of discasc scems so truly inherent
in these verbs, that the inflection has been added to
words of all kinds : Móaußdos, " lcad,” uolupodw, " to be
leaden -hued " ; 1Coos, “ stone," c0ców , “ to have the discase
of stone." Moreover, variations on the same model
have been fabricated : pulliów (in speaking of a tree ),
“ to produce only leaves " ; édeßopráw, " to need helle
bore " ; otparnyidw, “to have the disease of wishing to be
a strategist.”
The idea of disease has now entered into this inflection,
but it was by no means there originally. The starting.
42 SEMANTICS

point must be looked for in certain substantives in ca,

such as opoarula, " ophthalmia "; Melayxolla, “ melan
cholia. " 1 It was from this that the movement origin
ated : a movement which has produced a group that
might be called the nosological group.

Let us now quote an example taken from French.

There is a pejorative suffix -itre, which forms words such
as maråtre, bellâtre, douceatre. The history of this suffix
is instructive, but we must take it up rather far back.
Its birthplace is to be found in the Greek language,
where there were verbs in açw with no disparaging signi
fication : Davuáów , “ I admire " ; otrovôácw, “ I apply
myself ” ; oxolácw, “ I take leisure . ” From thence arose
the substantives in aornp, such as dikantúp, “ judge ” ;
dpyaotúp, “ workman .” Among these, we already see
certain words of suspicious appearance : tatpaotúp, “ he
who plays the father" ; untpáotelpa, " she who plays the
mother ;" è alaotńp, " that which counterfeits the olive
trec ” (that is to say, the wild olive).
This class of words pleased the Romans. As a general
rule, it may be noticed that evcrything that appeals
to malignity passes casily from one people to another.
The Latin tongue had, therefore, the words patraster,
filiaster. Cicero, in his correspondence, coins the vocable
Fulviaster, “ he who imitates Fulvius, a second Fulvius.”
From Latin, the formation in aster passed to the

1 This formation in ca, however, does not in itself imply anything

of the kind : åpuovia, “ union ” ; 8.800 karla, " teaching ” ; 'ueonuBpla
"noon," etc.
 IRRADIATION 43

derivative languages, in which it had a great success .

All the Romance languages use it. French has taken
possession of it, and employs it with more licence than
ever did Greek or Latin. Roussâtre, verdâtre, saumatre,
opinidire, medicatre, arc in common usage. The pejorative
sense, which barely existed in Greek, but which begins
to show in Latin, has therefore definitely entered into
this suffix .

Modern German has a class of verbs that may be

called " depreciative," for they express the action coupled
with a suggestion of disesteem and irony. They are
terminated in eln . Thus of klug, " clever, " is made
klugeln, “ to pretend to be clever, to split straws” ; of
wita, " wit ," is made witseln, " to try to be witty, to
talk nonsense " ; of fromm , “ pious," is made frommeln ,
" to cant.” Sometimes the verb in eln is taken directly
from another verb, deuten , " to interpret " ; deuteln , “ to
gloss." The depreciative idea entered as an afterthought
into this inflection, which had originally no disparaging
signification. The formation in -eln comes from ancient
substantives in -el, as is seen by Zweifel and zweifeln ,
Sattel and satteln, Wechsel and wechseln, Handel and
handeln. But as among these substantives there were
some with a diminutive sense, such as Würfel, “ thimble ” :
Schnitzel, “ shaving, shred " ; Äugel, " ocellus,” this cir
cumstance has sufficed to imbue the verbal infection
with a special savour. To say that these are products
of Analogy is an insufficient explanation. The popular
mind has multiplied these verbs because Irradiation had
 44 SEMANTICS

given them a special signification. The diminutive

idea is itself, if I may so say, an afterthought. The
suffixes which, in Greek or Latin , were used to form
diminutives did not in the beginning possess this
mcaning. But once they have acquired it, they are
indefinitely propagated. Thc fecundity displayed by
Latin on this point is well known. Like a gardener
who applies himself to variegating a flower adopted
by fashion, the popular mind, once imbued with the
taste, produces diminutives of all shapes. We then
see the diminutive suffix attach itself even to pronouns ;
ullus ( for unulus), singuli, ningulus are examples of
this. Everybody knows the richness of Italian in this
respect. Something similar may be observed in certain
dialects of modern German .

Irradiation may become a source of crror to the

philologist, if he persists in trying to find in the word
the textual statement of what it conveys to the mind. I
know scarcely any significative suffix, of which some
scholar has not attempted an explanation, by means of
a substantive or verb. Quite recently the verb memini
was detected in monumentum , argumentum . Pott, again,
1 Similar observations might be made on French words ending
in iller, such as sautiller, in eté, such as tacheté, etc.
* To give examples : Animula, apicula, avunculus, angellus,
corolla, bacillum , etc. A diminutive is the basis of somnolentus,
fraudulentus, violare . . .
• See Grimm , Grammaire allemande, iii. 688.
* It is well known that the suffix mentum is the development of
men : augmen , augmentum ; segmen , segmentum .
 IRRADIATION 45

tried to discover in patronymics such as' Ατοείδης, Πηλείδης,

the substantive židos, “appearance," although names like
Πριαμίδης, Τελαμωνιάδης, in which the same suffix appears
in a different form , should have suggested some doubt
to him . In the same way, Corssen thought that he
discerned a verb kar, “ to make,” in such words as volucer
or ambulacrum , a root bhar, "to bcar," in celeber, cribrum .
It is true that the mistake made by scholars is made
also by the people. But it must be confessed that the 1

popular mind deceives itself with a finer ingenuity. The

English sweet- heart, written as if it meant, " my sweet
heart," is formed with the same suffix as niggard,
sluggard, coward. It should therefore be written
sweetard, "sweetish." 1 But sweet - heart has certainly
more colour.
Similarly, in German, adjectives like triibselig, armselig,
give now the impression of being derived from Seele,
“ soul,” whereas they are the development of an abstract
suffix - sal, which is preserved in triibsal, mühsal. The
impression is so general that adjectives such as arbeit
selig, vertrauenselig, seem to be formed quite regularly,
and that seelenarm has been constructed in imitation of
armselig

There exists in Latin a form of participle intended , if

we are to believe the grammars, to express the idea of
obligation. It is found sometimes in the active : Nunc
est bibendum . - Denegandum est exceptionem . - Dandum
est operam ; sometimes in the passive : Asperum et vix
1
Sayce, Introduction to the Science of Language, ii. 346.
46 SEMANTICS

ferendum.- Urbem dux militibus diripiendam dedit.

Danda opera est. But whatever the construction, the
grammars assert - and the impression that we have of
the Latin would confirm their assertion — that a con
ception of obligation is contained in the participle.
This conception of obligation has nevertheless made
its entry after the fact. In fact, the participles in dus,
da, dum , as well as the corresponding gerundives,
expressed originally nothing more than the idea of
action , whether passive or active. This is clearly shown
66
by the ancient official formulas. " Were present at the
drawing up of the act ” was expressed in Latin by :
scribendo adfuerunt. "Presided at the execution of the
work ” was expressed by : præfuit operi faciundo. Latin
writers have left us, morcover, a fair number of examples
of this purely active or passive sense. Livy relates how
the Gauls were hewn in pieces while receiving the gold
for the ransom of Rome : inter accipiendum aurum cæsi
sunt. Cicero, in his De Officiis, speaks successively of
injustice committed or suffered. He ends the first part
with these words : de inferenda injuria satis dictum est ;
“ Enough has been said of injustices committed by
oncself."
I have purposely multiplied the examples on account
of the false ideas which still prevail on this point.?
Necessity is but a subsidiary shade of meaning which
1 Or even by operis faciundo ( Orelli, 5757), making
faciundum a neuter substantive, similar in meaning to the French
confection.
3 The true solution has been given by M. L. Havet. The
examples have been collected by our lamented pupil, S. Dosson.
 IRRADIATION 47

has penetrated by supererogation into forms of this kind.

To explain how it has so penetrated, certain formulas
such as : Decemviri creati sunt legibus scribundis.
Quattuor viri viarum curandarum , must be considered.
If in these formulas a substantive be put in the place
of the verb, the sense remains the same. Yet the
substantive contains in itself nothing that indicates the
idea of obligation.

Everybody knows the distinction made by Philology

between the “ material ” and thc “ formal ” element in
words. In every age it has been asked whether these
two elements had the same origin, or whether there is
between them some difference of nature. I need not at
present discuss this subject. I only wish to show that
we may happen to consider as belonging to the “ formal
element ” letters or syllables taken from the " material
element." This is a phenomenon of Irradiation.
We are furnished with an example by the Greek
perfects in κα such as λέλυκα, πεφίληκα. Georges Curtius,
with rare perspicacity, has shown that this k is in no way
different from the c of facio, jacio, and that it is still
enclosed in the “ material ” part of the word in certain
verbs such as ήκω, ερύκω, δλέκω.! The fact of its being
neighbour to the inflection sufficed to transform it into

De Participii Gerundivi Significatione, Hachette, 1887. See also

what I said in the Mémoires de la Société de Linguistique, viii. 307.
· See his Grundzüge, 5th edition, p. 61. M. Ascoli had already
conjectured something similar. It is the same c that we find in
Latin in fecundus,jucundus.
 TICS
48 SEMAN

an inflection itself. To call such a phenomenon " attrac

tion ," or "adherence,” is to name without explaining it.
The need of a clear and; convenient exponent has
brought about this metamorphosis : it has caused the
incorporation in inflection of what did not previously
belong to it, and has enriched the formal at the expense
of the material clcment. It was in certain perfccts like
déowka, Cornka, that the process began. But when once
the « had become a significative clement, it entered into
all verbs.
Here are two cxamples taken from the other end of
the history of Indo-European languages. Mr. Whceler
tells us that the Americans find mcans to provide a
singular for words taken rightly or wrongly for plurals,
such as Chinese, Portuguese. Corresponding to Chinese
they have made a singular Chince ; corresponding to
Portuguese, thcy have made a singular Portuguee. In
this way the inflection se passes to thc condition of a
“ formal ” clement."
hear German spoken, it might be thought that
there exists a second person of the verb ending in e :
Da biste ? - Lebste auch noch 2 - Was meinste 2 - Jetst
haste's. The origin of this e is not doubtful: it is plainly
a remainder of the pronoun of the sccond person du , of
which the consonant has become extinct, and the vowel
cmbodicd in the verb. But if these second persons
came down to us from a remote age, the vowel would be
taken for the remnant of an inflcction.

I " It would appear even that for the French word chaise, a singu
lar shay has been found. Wheeler, Analogy, p. 14.
 IRRADIATION 49

These cxamples, one of which takes us back to the

earliest periods of the Greek tongue, while the other
two are of our own time, show that borrowings take
place between the formal and the material elements,
Irradiation being the cause of this process of trans
formation ,
 CHAPTER IV

SURVIVAL OF INFLECTIONS

What this is-Examples drawn from French grammar - Archaism .

WHEN an inflection disappears, whether through the
action of phonetic laws, or for any other reason , it does
not follow that it will cease to exist for the mind.
There it may still live a long time, thanks partly to
tradition, partly to the position occupied by the word in
the phrasc, partly also to the comparisons which our
memory instinctively makes with analogous construc
tions. This Survival of the inflcction is not an un
important matter, nor is it without influence upon the
syntax.
This will become clearer from a few examples.
There exists in French grammar a rule which may
at first sight appear arbitrary, but which none the
less rests on a sound instinct of the language. А
word may not be used as complement to two verbs
which require different cases. Even if the word in
question is outwardly identical for both cases, the pro
hibition holds good. It is not permissible to say, for
50
 SURVIVAL OF INFLECTIONS S1

instance : " Vous saves que je vous ai toujours respecté et

porté une vive affection ."
Whence comes this prohibition ? It comes from the
Survival, in the background of our minds, of a declension
that has been practically abolished. The idea of the
dative, which continues to exist among us, admits no
mixing with the accusative, although, in the present
example, dative and accusative are identical in form .
The rule, I rcpcat, is not artificial: we are all conscious 7
of it, in reading the faulty phrasc. The fact is that there
is a reminiscence which serves us as guide. Were the
phrase transferred to the third person , it would be found
necessary to say : " Vous savez que je le respecte et lui
porte une vive affection ." The recollection , partially
present to the memory, of le and lui prevents the
confusion of the two vous.
for the same reason, it is necessary to repeat the
pronoun, although the pronoun does not change in the
phrasc : "Je te remercie et te serre la main,” 1 Thus an
inflection that has been destroyed still continues to
influence the mind, owing to association with a similar
form .
It may be said that by means of a few precious
fragments of this kind the declension of pronouns {
survives almost complete in the French language. The
I In his Remarques sur la Langue Française Vaugelas mentions
this rule : “ This rule," he says, “ is a very good one, and most
conducive to the purity and clearness of the language." Guillaume a
de Humboldt again expresses the same thing in these terms : " Es
sinken die Formen, nicht aber die Form, die vielmehr ihren alten
Geist über die neuen Umgestaltungen ausgoss."
52 SEMANTICS

dative continues to make itself felt in such phrases

as : “ accorde-moi la protection, donne - toi du repos, ne
nous faisons pas d'illusions, n'allez pas vous chercher des
regrets."
The accusative exists in a like manner. It would be
somewhat wounding to our internal syntax, were we to
say, in one single phrase : “ Oi se sont cachés, qui a
disperse nos amis 2 "

Another Latin forin that still survives, although

scemingly cxtinct, is thc ncuter. Perhaps, cvcn, the
French use it more than the Latins. They say : " Le
beau, le vrai, le bien, l'honnête, l'utile, l'agréable, l'infini,
l'intelligible, le contingent, le nécessaire, l'absolu, le divin ."
Their philosophical language is full of it. Their literary
criticism also, “le fin, le délicat, le romanesque, l'atroce.”
" Xavier de Maistre," says Sainte -Beuve, " a trouvé sa
place par le naïf, le sensible et le charmant." La Bruyère
speaking of Rabclais : “ Où il est mauvais, il passe bien
au delà du pire. ... Oi il est bon, il va jusqu'à l'exquis
et à l'excellent."
This faculty of employing adjectives in a gender that
sccms to have dropped out of the language, is connected
with thc presence of a certain number of ncutcr pro
nouns that . have been saved from the wreck, to wit le
("je ne le soufrirai pas, me le pardonncres vous ? " ), ce
(" ce fut la cause de ses malheurs, ce n'est pas qu'il soit
méchant, c'est à vous de commencer ." ), que ( “ que
.

ferons-nous, que vous en semble ? " ), quoi (“ quoi de plus

insensé, un je ne sais quoi ..."). These words, and a
 SURVIVAL OF INFLECTIONS 53

few other similar ones, have been sufficient to maintain

the neuter gender in the mind and in the language, and
to allow it an extension that is by no means coming to
an end. We even see that feminine substantives, such
as quelque chose, rien, have lost their gender and turned
into neuters.

There is also an example of Survival apart from pro

nouns. Thc French language has lost its declension,
and yet continucs to usc absoluto ablatives. “ Lui mort,
toutes nos espérances sont andantics.” “ La nouvelle s'ctant
répandue, des attroupements se formerent." What have
we here other than absolute propositions after the
manner of the Latins ? Before a construction of this
kind, our logical analysis is at a loss. This is one of the
examples which show how difficult it is to disconnect a
language from its origins, and how great would become !
the obscurity of the French language were it no longer
illuminated by the light of Latin.
Another example is the genitive, which, as is well
known, persisted for a long time in certain expressions :
l'Hôtel-Dieu, le parvis Notre- Dame, les quatrefils Aymon.
But, as will be seen in a moment, this construction,
having become obscure, was transformed by the popular
intelligence.
Thesc survivals are instructive, bccausc thcy allow us
to inſer that it was in no way different with ancient
languages, and that where we find some unexplained
interdiction or toleration, we are perhaps in presence of
the protracted action of a former state of things. It is
 54 . SEMANTICS

thus no doubt that must be interpreted the rule known

by the formula of τα ζώα τρέχει..

The law of Survival, like the law of Differentiation,

has its limits. When an inflection is no longer repre
scntcd save by a small number of specimens, and when
these specimens have themselves become unrecognis
able, human intelligence, deprived of guidance, no longer
knows whither to turn. An instinctive prudence, the
product of many unsuccessful attempts, causcs the
abandonment of constructions that have become too
difficult to understand. It rarcly happens that the
pcoplc fails in this precaution. What it does not
understand, it abandons or transforms.
It has, for instance, transformed the genitive con
struction of which mention has just been made. In
cxpressions such as : Regent Street, Queen Anne's Gate,
ctc., it is no longer a genitive that we pcrccivc: we
imaginc ourselves to be pronouncing the very name of
these public thoroughfarcs. Thus a construction has
been formed and most widely developed to which we
owe for the most part the names of our streets, of our
squarcs, and of our parks, not to mention a thousand
industrial inventions.
It may happcn that thcsc survivals arc prcscrved in the
language of literaturc, after they have disappeared from
thc language of the people. It is in this way that poetry
} has kept thc habit of inversions, which are nothing but a
licence of ancient days. So long as they do not detract
i See above , p. 37.
 SURVIVAL OF INFLECTIONS 55

from lucidity, these relics of a former age are precious :

they endow the language with dignity, grace and power.
But the deviation must not be too great. If the licence
of syntax supposes the existence of inflections long since
abolished or forgotten, a certain obscurity cannot fail to
result. The most subtle form of archaism is to appeal to
grammatical methods that no longer exist in the popular
consciousness. If it be comparatively easy to restore
ancicnt words to circulation, it is much more difficult to
revive and render intelligible the ancient turns of phrasc.
Survival is therefore a law of Language, whose true
limits every one must measure, in due accordance with
the idiom and the occasion ."
1 See what I said on the subject of German in my book, De
Penseignement des langues vivantes, p. 65.
 CHAPTER V

FALSE PERCEPTIONS

False inflections of the plural - False inflections of cases

Apophonia.
We are thus led on to spcak of a phenomenon nearly
related to the preceding one : " False Perception ."
We often imagine that we perceive an inflection where
nonc exists. Thus an Englishman, pronouncing the
plural oxen , thinks that in the syllable en he recogniscs
the distinguishing mark of number ; yet this is merely
the Anglo-Saxon stcm oxen ; Sanscrit, ukšan. The rcal
mark of the plural has been dropped.
It is casy to scc the source of this illusion. It is that
the singular, having lost half its stem, is reduced to the
syllable ox. Thenceforward there is a difference between
thic singular and the plural which serves to denote
number. The pcople, has : focling for utility, but docs
üt trouble at all about history. It uses whatever it
possesses : if it suffers losses, it utilises what is left. It
introduces a meaning into syllables that were meaning
less. Its Perception is therefore false from the point of
56
 FALSE PERCEPTIONS 57

view of history, but from the point of view of history

alone.
The same example may serve for German It has
even come to pass that the German language has been
so convinced of the existence of an inflection, that it has
mobilised and made liberal use of it. Not only does it
decline : der Ochs, die Ochsen , but it makes also : der
Mensch, die Menschen, and even, declining words of
foreign origin, der Soldat, die Soldaten ,
But German possesses another syllable, the history of
which is yet more instructive. When we are told that
in the plural Kind makes Kinder, we are given to
understand that er is the inflection of the plural; yet
er is nothing more than the suffix es or er that we
have in the Latin gener-is, in the Greek yéve(o )-os.
Which fact however has not prevented a whole cate
gory of words from imitating this model: die Weiber,
die Lämmer, die Dächer, die Biicher, die Götter. It
may thcrcforc bc said that the instinct which at the
prosent day brings about the recognition of a plural
termination in Kind -er, Weib-er, Häuser, is from the
point of view of history a False Perception ; this how
ever has not prevented its becoming a regular inflection
of the language.
Facts of this kind are easier to observe in modern than
in ancient languages. The reason is casily guessed, for
it consists in nothing but the lack of anterior documents.
The English word child, which formerly made in the plural
cildru , cildre, has added on to this the syllable en : children. On the
original identity of Kind and child, see the Mémoires de la Société
de Linguistique, vol. vii. p. 445.
 S
NTIC
58 SEMA

Still, in Latin the e of dulce, nobile, appears to be the sign

of the neuter, although the neuter is simply recognisable
by the absence of inflection . We have only to compare
the Greek ίδρις, neuter ίδρι, or εύχαρις, neuter εύχαρι, to
see that the e of dulce occupies the place of an ancient
termination i. Were it possible to question a con
temporary of Augustus on his impression of words
like onus, scelus, he would no doubt say that the syllable
us is there to mark the inflection. The Greek thought
that in the imperfect έλυε, in the aorist έλυσε, he
discerned the third person, albeit the mark of this
third person ( a t) had disappeared.
Another kind of False Perception is a belief in the
presence of grammatical forms which have never existed.
In Latin, the declension is shorter by one case in the
plural than in the singular : to wit, the dative and ab
lative possess, and probably have always possessed, a
single plural infection. Yet this deficiency is not felt.
So little is it felt that philologists are not yet agreed as
to which of the two cascs is missing.
We have just scen that the loss of an inflcction may
add to the significative value of that which survives.
The well-known phenomena of Umlaut and Ablaut draw
from that fact the greater part of their importance.
It is recognised that the difference of vowel between
man and men , between Vater and Väter, is by no means
primitive, but that the "softening" of a into e or ä is
duc to the influence of a final syllable originally present
· but later on abolished by the wear and tear of ages.
This difference of vowel is sufficient to distinguish the
 FALSE PERCEPTIONS 59

plural from the singular. It possesses indeed all the

greater value in that it is alone at the present day in
marking an important grammatical relation. Could it
but have been introduced universally, this manner of
denoting the plural would have had the merit of elegance
and brevity.
It is impossible to consider the difference between man
and men without thinking at once of the difference that
exists in conjugation between the diverse tenses of certain
verbs : sing, sang, sung. Here also the present instinct
of Language is not in agreement with history. It would
seem as if this variety of vowels had been invented ex
pressly to mark the variety of tenses. Yet that is not
the case : by going back a few centuries, it may be
proved that it is only the accompaniment of other cx..
ponents, which are the significative and veritable ex
ponents. The diversity of vowels is produced by second
ary causes, causes of accentuation or of contraction .
But the idea suggested by modern language, is that the
change of i into a is intended to indicate the preterite,
and that the change of i into u is made to mark the
participle. Though not significative in its origin, this
change of vowel has ended by becoming significative.
Perhaps even there may be a more intimate connection
between this advent of meaning and the downfall of the
flectional apparatus, for it may be suspected that the
people does not abandon what is useful to it, until
conscious that it already possesses a substitute.
 CHAPTER VI
ANALOGY

False idea of Analogy - Cases in which Language allows itself to

be guided by Analogy- (a ) To avoid some difficulty- ( ) To obtain
greater clearncss_ (c) To emphasise either an antithesis or a
similarity- (d ) To conform to some ancient or recent rule
Conclusions on Analogy.
In the philological treatises of the last fifteen or twenty
years Analogy occupies a considerable space, and that
not without reason , since man is by nature imitative : if
he has to invent an expression, he does it more quickly
by modelling it on some existing type, than by limiting
himself to original creation. But it is a mistake to
represent Analogy as a cause. Analogy is nothing more
than a means. We shall try to demonstrate the true
causes.

Languages have recourse to Analogy :

(a) To avoid some difficulty of expression.-- A more
convenient formation having been found, the ancient
" I suppose that is unnecessary to repeat what I said at the
beginning about the will, semi-conscious and but feeling its way,
that presides over the evolution of Language.
60
 ANALOGY 61

formation is, in a manner, arrested in its power of ex

tension, reduced to that which it actually possesses, and
deprived of all opportunity of further enriching itself.
But from the moment that it ceascs to be enriched it
becomes impoverished . Habit effectsthe abandonment,
now on one point, now on another, of the ancient forma
tion. It ends by keeping only a small remnant of faith
ful specimens, specimens themselves increasingly incom
plete and uncertain .
A striking example is furnished for us by Greek, with
its two conjugations in ui and in w, which we find since
the earliest ages in concurrence, the one (in ui) con
stantly retiring, the other (in w) as constantly pro
gressing
The conjugation in Mi is without doubt the more
ancient," as it is the more complicated and the more
difficult of the two. It is therefore also a closed forma
tion , reduced to some hundred verbs (of great import
ance, indeed), the number of which cannot increase. As
early as in the Homeric age, the conjugation in pi is not
only blockaded , but also attacked in its own quarters.
By the side of δείκνυμι, a verb δεικνύω is seen to arise.
The verb elul, “ to be," makes in the participle öv, after
the model of aúwy. The verb elu , " to go," makes in the
optative locus, after the model of dúolu. Reduplicated
verbs, like πίπτω, μίμνω, γίγνομαι, which were of the
same kind as τίθημι, δίδωμι, κίχρημι, have decisively
Some philologists in these last few years, have maintained that
the conjugation in the was the more modern. We can see in this
thesis but an ingenious paradox, the birth of which a mere glance
at Latin should have obviated .
62 SEMANTICS

abandoned the conjugation in me, in order to pass over to

verbs in w.
The conjugation in de presents therefore the spectacle
of a formation that has been stormed and sacked. Each
of the losses that it has undergone has been a gain for
the conjugation in w. The memory does not willingly
burden itself with two mechanisms working concurrently
towards one and the same end : let it but hesitate, low
cver slightly, and the forms most commonly in use will
be the first to present themselves.
The conjugation in w offered the advantage of a more
uniform accentuation, of a lesser varicty of vowels, of a
more visible symmetry ; the o or e that insinuates
itself between the root and the inflection (Aú-o-jev,
dú-e-te) is as a buffer that prevents conflicts. This
grcater facility was bound to ensure victory to the con
jugation in w.
In Latin, matters are yet further advanced. The
struggle is already at an end. Who would imagine,
without the light shed by cognate languages, that sistere,
bibere, gignere, serere, were once reduplicated verbs,
similar to τίθημι , δίδωμι ? The survivors of the ancient
conjugation, esse, ferre, velle, and a few others, are classed
among the irregular verbs. Yet they are only irregular
in some of their forms. The work of adjustment being
carried on by the people, velle has become in low Latin
volêre, whence the French vouloir ; posse has become
potére, whence the French pouvoir. The last remnants
have therefore becn little by little absorbed.
Nevertheless , such is the slowness of these evolutions
 ANALOGY 63

that even at the present day there remains in every

Romance language a witness, unique indeed, to the con
jugation in Mt. This is the verb étre, " to be," which by
its anomalies betrays its earlier origin. It has moreover
been vigorously cut about. In Spanish they have somos,
sois, son, as if the Latin were sumus, sutis, sunt. Italian
derives a gerundive essendo from an already modernised
infinitive essere.
What has happened to the verbs has taken place also
with the substantives. A declension that is easier and
clearer gains ground from the other declensions. Already
in theinscriptions of Delphi we find τεθνακότοις , αγώνοις ,
εν άνδρoις τρίοις , εν τοις οκτώ ετέοις, etc. This is a begin
ning that foreshadows what will happen in course of
time to this third declension, which is too delicately
adjusted. A nominative ăywvov presently arises in
imitation of the dative àyóvous. It is in this way that
modern forms like άρχοντοι, γέροντοι , are fashioned. At
quite an early stage, we find the nominatives púrakos,
μάρτυρος, διάκτορος, side by side with φύλαξ , μάρτυς,
διάκτωρ.1
Something of the same kind happened to the feminine
gender. Nouns of the third declension have been
changed into nouns of the first : instead of 16 ,
modern Greek says η φλόγα, instead of την ελπίδα, την
ελπίδαν ,,
It is evident that the dative plural was the stone of
stumbling : it is at this point that declensions invariably
1 The facts are the same in India. See Otto Franke, Die Sucht
nach a - Stämmen im Pali (Annals of Bezzenberger, xxii. p. 202).
64 SEMANTICS

begin to run off the rails. The present participle åkoúw

should have produced the scarcely convenient formation
ακούουσι . But as early as in the Homeric language we
find åkovóVTECO1.1 These forms in cool, which had their
birth among themes such as teixos, become very frequent
on inscriptions, where we find, for example, åpxóvredo ,
εόντεσσι , ελθόντεσσι , αγώνεσσι, πάντεσσι , ευεργετησάντεσσι.
In comparing αγώνεσσι and αγώνoις, we are convinced
that on both sides the aim is the same : to avoid dy@ol.
In Latin, we find again the same facts, and in a still
more visible manner. The declension of consonants is
already half remodelled. It is to the type of the
dcclcnsion in i (avis, collis) that the diſferent inflections
have been brought back. This may be easily verified
by comparing, for example, thứ Greek pepóvr -wv, with
the Latin fercnt- ium , the Greek pépovt -a with the Latin
ferent-in, the Greek pépovt-es with the Latin ferent- ês.
( for ferenteis ). It must be remembered that the Latin
pronunciation compresses words, abridges or extinguishes
final syllables : all of which are causes that rendered
the declension indistinct. The remodelling spread by
degrees as far as certain nominatives : thus juven ,
" youth " (Sanscrit juvan ), whence juven -tus, has become
1
Odyssey, I., 352.
? There are still a few rare traces of the anterior state of things.
Aulus Gellius (xix. 7) quotes from Lævius the expression silenta loca .
Silentu is a plural neuter after the manner of piloûvtoa. But Latin
has lost the habit of thcsc ncuters : it says veloc-ill, locupletiu ,
simpliciu . In the genitive plural there is still purentum , animan.
tum : but thc ordinary form is -ium (alulescentium , infantium ,
discordium ). ?
 ANALOGY 65

juvenis ; aus, " car ," whencc au (s)dire, auscultare, “ to

listen ,” has become ausis, auris.
( 6) To secure greater clearness .- So far as it can be
avoided , grammatical forms must not give rise to
ambiguity. If they are too short, too blunt, they
threaten to become unintelligible. That is what has
happened for instance to the genitive plurals of the
second declension. The ancient genitivc in um (Greck
wv), of which examples still exist in stereotyped expres
sions,' gives way to a genitive in örum borrowed from
the pronouns, and having in addition the advantage of
being symmetrical with the forms in irum of the first
declension .
The superlative originally ended in tos. Of this very
simple formation there have remaincd τρίτος , τέταρτος,
dékatos. We know indeed that the ordinal numbers are
formed by the help of the same suffixes which scrve to
mark the degrees of comparison. But this exponent
Tos, too simple and too short, might have given rise to
misunderstandings. By detaching the a from déka,
Greek obtains a more complete suffix, atos ; whence
such superlatives as ύπατος, έσχατος, πύματος. For the
sake of clearness, the language added the r of the com
parative repos to the suffix, atos ; thenceforward there
was the suffix tatos, which permitted the opposition of
φίλτατος tο φίλτερος.2?
The desire for explicit forms, cxplains how , in French,
Pricfectus fabrum ,duo milia sestertium , templa deunt, ctc.
* We owe this model of historical study to M. Ascoli in the
Studien of Curtius, ix. 342.
F

1
66 SEMANTICS

troisième, quatrième have been substituted for the

.

ancient ordinal numbers tiers, quart, quint (le tiers parti,

un quart voleur survient ...). Of the ancient Latin
ordinals, only the first two are left : but already deuxième,
instead of second, is familiar to French ears.
In the conjugation of verbs, certain past participles
threaten to become estranged from the verb from which
they are derived. Who realises now - a -days the relation
ship of the French poids,which should be written pois,with
pendre, of toise with tendre, of route with rompre ? It was
useful to have a form which emphasised the affinities to
a greater degrec. This explains the favour with which
thc participle in ntus was reccived ; pendu, tendu, rompu .?
The movement started with a few rarc forc - runners
which we find in low Latin : pendutus, decernutum ,
incendutum . They are themselves a result of imitation
(Latin solutus, statutus). Thanks to this final syllable,
French has re -established the disordered lines of its
conjugation.
Instead of nous prenmes, nous faismes, which should
have been the outcome of the Latin prendimus, facimus,
nous pren -ons, nous fais-ons, were formed : instead of
1 As late as in the sixteenth century, fractions, in mathematics,
were called nombres roupts. Route designates a way made by
breaking through (rompre) forest and field .
* Children when they say j'aiprendu, are conforming themselves
to the models furnished by language. They have long been
recognised as active auxiliaries of grammatical regularity. In.
stead of I came, I caught, tlicy are often hcard to say, I comcil, I
Cutiheit.
3 Latin vcrbs having their perfect in ui, likc hubui, tonui, were
the first to take a participle in utus.
 ANALOGY 67

vous prents which should have resulted from the Latin

prenditis, we find vous pren -es. Whence come these
fuller inflections ? The second person of the plural
sufficiently indicates this. They have been borrowed
from the first conjugation ."
Let us give another example drawn from the Greek
conjugation .
In the third person of the plural, the second aorists
of verbs like τίθημι have a very short inflection : έθεν ,
έβαν, έσταν, έφαν, έφυν, etc. Homeric language abounds
in forms of this kind. But its drawbacks are obvious :
these third persons of the plural too closely resembled
the first persons of the singular. The means employed
were very simple : thanks to an addition borrowed from
the first aorist, there were formed Broav, Cotacav,
έφασαν, έφυσαν, ανέθεσαν.2
It is a fact, surprising at first sight, but attested by
numerous proofs, that the suffixes most frequently in
use in our modern languages are borrowed. Thus
Greek has helped us to form words in -ism , such as
optimism , socialism ; in -ist, such as artist, florist; in -ise,
such as authorise, fertilise. German has furnished the
suffix -ard, as in the French vantard, bavard, the English
dastard, coward, bastard. Italian the suffix -esque, as in
gigantesque, romanesque, picturesque. Strictly speaking .
even French words in -al, such as national, provincial, in
" (iteur, like ordonnateur, provocateur, are formed by the
1 The only survivors which have not been remodelled aro : von
dites (clicitis), vousfaites ( facitis).
* Curtius, Das Verbum , i. 74.
68 SEMANTIC
S

help of Latin suffixes, since these same suffixes, when they

have made their entry into French by way of popular
usage, have worn a different aspect. It is the want of
explicit forms, which shall stand out clearly before the
eyes, that has procured this special favour for foreign
inflections : the native inflcctions, having been subjected
to the wcar and tear of ages, and having mingled with
the anterior part of the word, are not so obvious.
The same practice is observable cverywhere. It is
well known what a success has been obtained in German
by the French inflection -ie, which has given the sub
stantives in -ei, like Bäckerei, Zauberei. English again
has borrowed from the sccond conjugation of French
that syllable -ish that is often found not only in finish,
nourish, where the model is furnished by French, but
also in publish, distinguish, where the suffix is trans
ferred by imitation.
In every age and in every nation there have been
purists who protested against these borrowings. But
those who really form the language, anxious above all
to be understood, and that at lcast cost, trouble thein
selves but little about the origin of the materials with
which they work.
( c) To emphasise either an antithesis or a similitude.
Language here reveals to us a fact of psychology ; the
mind, which naturally associates ideas in couples, likes
to solder together contraries by giving them the same
exterior. While this helps the mcinory, it gives greater
relief to speech. " Nothing is more natural,” says
Bain , “ when we consider a quality, than the disposi
 ANALOGY

tion to return to the other quality, which forms its

contrast. "
We will begin by the simplest examples.
Day and night form an antithesis as old as the hills :
on the model of diu, Latin, diverting the ablative nocte
from its declension, has made noctu . On the model of
diurnus it has made nocturnus,
Another antithesis no less ancient is that of life and
death. On the model of vivus, Latin has made mortuus.
According to the rules of the Latin language, morior
should make mortus, as orior, experior, make ortus,
expertus. But as occasion for antithesis arises per
petually , the final syllable of the one has communicated
itself to the other.
The expressions " before " and " after" are likewise of v
a nature to influence one another. By the side of the
adverb antid, that became later on ante, Latin formed
an adverb postid, that became later on poste or post.
Postid was preserved in postid -ea, which is modelled on
antid -ea. At its basc is the syllablc pos, " aftcr.”
Wc scc from this that, in order to dctcrminc a crcation
by Analogy, there is no necessity for a language to
furnish a large number of models. In the cases that we
" It has lately been maintained that it is noctu that has influenced
diu : but to establish the true filiation, it is sufficient to recall the
Sanscrit divas or djus, " day ":( pūrvē-djus, " yesterday").
: The form morius is, in fact, that to which the verb has reverted
in the Romance languages.
8 Mortuum aut vivum . - Vivo et mortuo. C. 1. L., vi. 6467 ; ix.
4816, etc.
4
Sanscrit pas, “ after," in paçocat.
70 SEMANTICS

have just mentioned , a single word has sufficed ; but

that is because the two terms were in direct opposition.
It might be said that Analogy makes its power felt in a
direct ratio to the situation. Thus in French we find
the adjective méridional, for whose suffix, unknown else
where, it would be impossible to account, were it not for
the antithesis septentrional.
Such a form is inexplicable, until compared with its
opposite. So also čutów (in speaking of a difficulty,
an obstacle) is only explained by ékmódwv, “out of the
way of one's feet.” 1
The Greeks, who already knew Analogy by an
tithesis, called it by a pretty name : ovverðpoun kar'
évavtiórnta. The image is taken from a beast that
breaks loose from its own herd and follows another.
We will now give some examples of Analogy serving
to emphasise a resemblance.
As names or relationship like πατήρ, μήτηρ, θυγάτηρ,
had their dative plural in -aoi, the Greek viós, " son," for
no other reason, made viáo.. M. J. Wackernagel points
out a similar case in Sanscrit . The word pati, which
mcans at the same timc “ master ” and husband ," has
two genitives, the onc ( regular) patis, when it significs
master," thc other ( irregular) patjus, when it significs
“ husband .” This patjus comes from genitives like
pitus, " of the father" ; matus, " of the mother."

· Analogy by opposition appears equally in the antithesis fieis

and upeis, warpós and winpos. See also (Mem . Soc. Ling., ix.) what I
said of the adverb σιωπή..
? Journal de Kuhn, xxv. 289.
 ANALOGY 71

Greek possessed a substantive ουθαρ (genitive oύθατος ),

“ breast," the antiquity of which is attested by the Latin
uber, and the German Euter, as well as by the Sanscrit
idhar. These nouns in -ap, -atos have been multiplied
to mark some portion of the body. We find yóvate,
" the two knees" ; varc , " the two cars " ; " porúnate, “ thc
two eyes,” and even a 'pnap, thc hcad ."
Briefly, we may find in cvery language some words
which from being similar in mcaning have approx
imated in form . Greek, for example, has Adpuys, and
φάρυγξ, σύριγξ and σάλπιγξ. Sanscrit has anguftha,
“ the thumb " ; öštha, “ the lip ” ; kòštha, “ the belly ” ;
upastha, “ the lap." Celtic languages have their words
in arn and in orn ; vague relics of classification , more
than half effaced, comparable only to those alignments,
which , on the sites of vanished cities, still bear witness
that in former days men tried to build their dwellings
with some show of order.1
It is in Syntax especially that this kind of symmetry
is observed. Many constructions rcpugnant to pure
logic find herein their explanation. If verbs signifying
" to tako, to deprive, to snatch away,” arc in Latin
constructed with thc dativc, it is bccausc “ to give, to
assign, to offer," have that construction. If diffidere
alicui was said , that was because credere alicui was
said. If obliviscitur nostri with the genitive was said,
that was because meminit nostri was said, also with the
· See Bloomfield, On Adaptation of Suffixes in Congeneric Classes
of Substantives, Baltimore, 1891.-Zimmer, American Journal o )
Philology, 1895, p. 419.
72 SEMANTICS

genitive. Finally, if in urbe was said with the ablative,

which would seem to imply a contradiction , since the
ablative denotes an idea of removal, that was because ex
urbe, ab urbe, were said. Thus also in German in dem
Haus, zu dem Haus, in which in and zu are used with
the dative, have led to the cmployment of the dative in
such expressions as aus dem Haus, von dem Haus. So in
English because we say agree with some one, we say too
differ with some one.
We have only to listen to the conversation of people
who arc but imperfectly acquainted with a language,
and to take note of the mistakes which they make, to
scc that it is by associations of this kind that they are
usually inſlucnccd.
(n ) Analogy in order to conform to some ancient or
recent rule.-- These words nccd an explanation . We
arc now spcaking of a rulc not yet formulated ; a rule
at which mankind strives to guess, and which we see
children trying to discover. By pre- supposing its cx
istence, thc pcople actually creates it. The idea that
Language is obcdicnt to fixed laws is profoundly im
pressed on the popular mind ; indccd nothing can be
more reasonable, since, without laws, Language, would
ccasc to bc intclligible, and would fail in its primary
and only objcct. Wc scc that with the ordinary man
an infringement of what he supposes to be the rule
provokcs cither laughter or contempt.
i Obliviscor means literally " to yellow, to be effaced .” The
mctaphor is drawn from a writing that fades. Varro ( De L., L.
v. 10 ) calls words that have dropped out of usage : oblivia verba.
 ANALOGY 73

Forms which bewilder by an unusual aspect, are

thcrcfore regarded as faulty, and brought back to the
80 -called regular type. It is in this way that exceptions
become less and less numerous and finally disappear.
Philologists, conservative by profession, are usually
unfavourable to this kind of adjustment. Here, how
ever, Analogy fulfils a necessary function , without which
there would soon bc nothing but obscurity and disorder.
But the people must not be set problems too difficult
for solution : if there are pitfalls in the path it will fall
into them. Isidore de Séville records a verb of the first
conjugation, in use in his day, prostrare, “ to throw to
the ground” ; it is prostravi that has produced this verb,
the road that lcd to prosterno having becomc too difficult
to find. Even in classical Latin we find delere, “ to
cffacc, to dclctc," drawn from the perfect delovi, which
is a compound of linere. There was a verb prestare,
compound of stare, which formed in thc perfect præstiti,
“ I have excelled ” ; another verb præstare, derived from
prestus (pre -situs), " prepared, ready," has therefore also
formcd præstiti, “ I have prepared, I have furnished.”
The memory of the people is short. We find a plural
like omnes ( for homines) cnriching itself with a ncuter
omnia, and a singular omnis : wc scc a feminine felix
.

( from fela, “ brcast " ) producing a masculine and a

neuter.
It is interesting to note how scrupulously the rule,
once admitted, is obeyed and applied. The philologist
who watches this process and who, knowing the elements
i Felicia arma. Felix omen .
74 SEMANTICS

that have been set in motion, secs the most incongruous

materials going through the mill, cannot help admiring
the performance. This has been erroncously called a
compulsion ( Systemswang ). There is no compulsion :
there is only voluntary obedience to the rule.
Here are a few specimens.
We are accustomed to find Greek verbs taking the
syllabic or temporal augment in the imperfect and in
the aorist. But wc arc not prcparcd to find the augment
modifying an adverb or a pronoun. This, however, is
what happens when, in Xenophon, compound words like
ÓTLOOopúlaç, "rcar-guard,” avrópodos, " deserter," give
birth to imperfects like úhlo obvláket and to aorists like
avtomódnoe. No one is surprised excepting the philo
logist, who sccs in it an example of popular logic. In
modern Greck, where the augment survives, it is placed
unhesitatingly before prepositions; it is usual to say ,
for example, επροτίμων, “ I liked better ” ; ήνόχλησα, “ I
have disturbed . " Ancient Greek had already made a
beginning by saying έκάθευδε..
There is nothing very surprising in the fact that
Latin took a passive or middle participle like amamini,
laudamini, and made out of it a second person of the
conjugation, with estis implicd : it is as if we found in
Greek φιλούμενοι έστε, τιμώμενοι εστε. But the point at
which Analogy begins its work is when we find amabi
mini, amemini, amaremini, forms which are heteroclite,
but perfectly intelligible.
Analogy is specially interesting to observe when it
grapples with an unforeseen difficulty ,
 ANALOGY 75

The reduplication of the initial syllable of verbs,

obligatory in the perfect, became almost impossible in
the case of the groups on , or, ok, or of the letters $, E.
We know how Greck gets round the difficulty. In such
a case , instead of reduplication, it is satisfied with the
augment. We seem to witness some such compromise
as appears in the history of institutions and of laws.
Or, if this comparison gives too much prominence to
self-conscious reason , let us say that we are watching
the labour of some ingenious animal, which builds itself
a house with materials unequally fitted for the purpose.
It is especially important to notice the aim which is
thus darkly pursued. It is impossible for any one who
studies the Greek verb not to recognise an intention of
completing thc design : by thc side of thc aorist indica
tive έλυσα there are found an aorist imperative λυσάτω,,
an aorist optative λύσαιμι, an aorist participle λύσας.
Thc a that rcappears in thesc various forms is, as it
were, its signature. The intelligence of the masses is
here seen on one of its most interesting sides : by the
simplest means, it wins through the difficulties which, in
every profession and every art, the material opposes to
the workman .

The foregoing observations will show in what light

Analogy should be regarded. If we considered the use
made of it in certain books of recent date, we should
take it for some great sponge, passing vaguely over
grammar, mixing and confusing the common forms, for
the aimless cffacement of legitimate and uscful dis
76 SEMANTICS

tinctions. But such is not its character : it is on the

contrary in the service of reason , a reason indeed some
what short of memory, but none the less a true and
necessary motive power of Language.
The question has been often discussed whether “ in
the childhood of our languages ” Analogy had as much
power as it has to -day. “ Can we,” says Curtius,
admit analogical formations at such a remote epoch ?
Analogical formations secm to me probable in recent
periods only. . . It was certainly not by chance that
attention was first drawn to thesc facts with reference
to modern languages, more especially the Romance
languages.”
We cannot on this point agree with the Icarned
Hellcnist. If attention was first drawn in this direction
with reference to the Romance languages, the reason is
that the Romance languages Icave their origins cx.
poscd to vicw, an advantage which is wanting to the
ancient cpochs. But the causes which bring about
the changes being causes inherent in the mind and
imposed by the conditions of all Language, there is
no reason to believe that they acted less powerfully in
the past.
Is it true, as has also been said, that Analogy is a
blind force, persisting to the very end, allowing nothing
to arrest its progress ?
It is difficult to believe that this is the case when,
abandoning theory, we face the facts. Expericnce
proves on the contrary that Analogy has its limits, which
indeed form a study at least as interesting as the actual
 ANALOGY 77

phenomenon. Motives of clearness or of harmony

suffice to hold it in check.
One last question must be put : does Analogy deserve
the disfavour to which certain philologists have con
demned it ?
If unduly pressed, Analogy would make languages
tou uniform and, in consequence, monotonous and poor.
The philologist, the writer, will always, by taste as by
profession , be on the side of the vanquished, that is to
say of the forms which Analogy threatens to absorb .
But it is thanks to Analogy that the child, without
learning one after the other all the words of a language,
without having to test them onc by one, yet attains to
mastery over them in a comparatively short time. It is
thanks to Analogy that we are sure of being heard, sure
of being understood, even if we chance to crcate a new
word. Analogy must thcrcforc be regarded as a prim
ordial condition of all Language. Whether it has been
a source of fecundity and of clcarness, or whether it has
been the cause of sterile uniformity, this the individual
history of each language alone can teach us.
 CHAPTER VII

NEW ACQUISITIONS

Need of indicating acquisitions together with losses — The infinitive

– The passive - Adverbial suffixes - Historical conclusions.

It is easier to distinguish the gaps which occur in a

society than to notice the new forces which are making
themselves manifest. In thc same way we more often
hear of the losses undergone by Language than of the
reinforcements which come to its aid. Grammatical
evolution is brought about so slowly, and by so obscure
a process, that, for the most part, it escapes the notice
of the observer. Yet it is hardly credible that during a
period of four thousand years, the Indo-Europcan lan
guages should have suffered a constant waste, without
compensation of any kind. The history of the losses
has been often written : that of the acquisitions still
remains to write. We propose, by way of indication, to
enumerate a few examples.
There can be no question, be it well understood , of
creations ex nihilo. The form which progress assumes
78
 NEW ACQUISITIONS 79

is the appropriation to new usages of material trans

mitted by the past.
Let us first take the infinitive.
This exceedingly valuable form , the first to be learnt
by children, the first to pass between two peoples when
they come into contact and try to understand each other,
has nevertheless not always existed. It is, on the con
trary, the product of a slow selection : it is the fruit of a
tardily accomplished union between the substantive and
the verb. The relatively recent date of the infinitive
can at once be ascertained, by noting how Latin and
Greek, in agreement concerning the rest of the con
jugation, here part company with each other. There is
no resemblance between the inflection of Aéyel and that
of legere, between cival and esse. And even, without
going outside the Greek language, by coinparing dia
lectal formations like fupev, eivai, fuerat, it can be ascer
tained without a doubt that until a fairly recent cpoch,
Greek had not made its definite choice. Latin seems at
first sight more decided ; but a little attention shows
plainly that it is even farther from realising the unity of
the infinitive, since it divides the function between three
forms: the infinitive properly so-called, the supine, and
the gerundive. It is only in modern languages that this
unity becomes an accomplished fact.
The infinitive represents the verbal idea, disburdened
r
of all accessory and adventitious elements. It knows
neither person nor number. The idea of the voice
(active, middle, passive) is in the main foreign to it.'
' A wine pleasant to drink, a counsel difficult to follow , an insult
80 SEMANTICS

Even the idea of tense only entered in by a sort of

redundancy and thanks to bclated after- touches. Certain
grammarians would turn the infinitive into a mood of the
verb ; but it is not a mood. It is, as the ancients justly
said, the most general form of the verb (rò yevik“ TATOV
ρημα), the name of the action (όνομα πράγματος).
To realise the importance of this form , it is sufficient
to read a few lines of some modern language. Halfverb
and half substantive, the infinitive, though it does not
carry the cumbersome baggage with which both these
species of words are laden , yet renders identical scrvices.
Like the verb, it has a transitive power ; it is able, like
the verb, to take to itself a subject ; like the verb, it can
be accompanied by an adverb or a negation. But, on
the other hand, cmployed as a substantive, it can be
either subject or object ; it is placed after prepositions
like to, of, for, without, and always without the trouble
of inflections. It is able to express an exclamation,
a desire, or an order. Finally, it is less exposed to
that Concretion of meaning, to that crystallisation, that
coagulation, of which we shall have to speak later on,
and with which all substantives, even abstract sub
stantives, are threatened . ?

impossible to forgive. In Greek, καλός οράν, άξιος θαυμάσαι, ράδιον

mabeir. In Latin, Mirabile visu, difficile dictu, etc. Cicero (Ad.
Fam ., ix. 25) gives us in passing this example of a change that has
come over the meaning : Nunc adles ad imperandum , vel ad paren .
dum potius : sic enim antiqui loquebantur.
1 Infinitorum vis in nomen rei resolvitur. (Priscian.)
• Compare, for example, frui and fructus, regere and regio, etc.
Sce farther on, the chapter on abstract words.
 NEW ACQUISITIONS 81

In face of such advantages, one wonders what could

have retarded to such an extent the creation of the
infinitive. To answer this question, we must for a
moment glance backwards and consider the general plan
of our languages.
Every time that there is question of classifying lan
guages according to their greater or less perfection , we
are accustomed to speak of the Indo -European family as
situated on the topmost rung of the ladder. Yet no
long scarch is required to find in this family what we
regard as a characteristic of but slightly advanced
idioms. )
head, your head, his head," but not " head " in general.
This indeed is barbarous. But it was in no way dif
ferent with the Indo-European verb, which could say
φέρω, φέρεις, φέρει , but not φέρειν. In the primitive plan,
action had always reference to a person. A form like
dlowji, oldot, represents in itself alone a whole pro
position ; it contains at the same time both the verb and
its subject. Our languages are not therefore so far
removed from the state known as holophrastic, in which
the word in itself was a phrase.
The infinitive is a conquest of abstraction. It had to
be sought for outside the verb, among the substantives.
The elaboration of the infinitive had alrcady been begun,
but not completed, at the pro -ethnic epoch. Centuries
were needed for each idiom to fix its choice on a certain
form of substantive, and to be put in possession, to the
exclusion of others, of some of the essential properties
of the verb .
G
82 SEMANTICS

It is here that you appreciate the advantages of what

is called the lack of transpicuity or phonetic corruption .
This so -called decadence has contributed not a little
towards developing the full utility of the infinitive. It
is difficult to know with certainty to which case of the
declension such Greek forms as ζευγνύμεναι, ιδείν, φέρεσθαι
belonged. But this indecision only rendered them easier
to handle. It is the same with the Latin infinitive. If
the forms modelled on videre, audire ended by ousting
the forms modelled on visum , audituin, it was perhaps
because in them the mark of the declension is to a
greater degree obliterated.
This reminds me of a fact that well demonstrates the
importance which the infinitive has assumed in our
languages. When, in the thirteenth and fourteenth
centuries, German enriched itself with a quantity of
French verbs, it adopted them under the guise of the
infinitive, superadding, oddly enough, German inflections.
It is owing to this that we find in Wolfram von Eschen
bach fischieren , "to fasten" ; leischieren, "to leave" ;
loschieren , " to lodge ” ; parlieren, " to speak ” ; and many
others. The result is that at the present day when the
German says ich spasiere, he adds to the infinitive
espacier the inflection of the first person . Nothing more
clearly proves how, in our modern languages, the idea
of the verb has incarnated itself in the infinitive.1
1 This explanation of the German verbs in -ieren has been dis
puted by Mr. Leo Wiener (American Journal of Philology, 1895,
p. 330 ). This man of science thinks that the origin must be sought for
in nouns in -ier, -ierre, like floitierre,“Autist," whence floitieren ,“ to
play the flute.” But the facts seem scarcely in agreement with
 NEW ACQUISITIONS 83

It will be asked how Greek, having once possessed the

infinitive, could have let it fall into disuse during the
Middle Ages. This loss is indeed one of the most sur
prising facts of Indo-European philology ; for to say, as
has been said recently, that the Greek infinitive was lost
through being too freely used, is to give an explanation
beyond the grasp of ordinary intelligence. But it must
be remarked that the blank caused by the absence of
the infinitive became especially painful when modern
Greek, finding itself in the company of the other languages
of modern Europe, felt the need of equalising the re
sources of its syntax. It may be supposed that neither
the liturgics of the Church, nor the folk - songs, with their
brief and simple language, had been conscious of any
The idiom θα (θέλει ένα) with the subjunctive
was used instead. An intellectual tool is lost by non
usage : a form too rarely employed is effaced from the
memory .
By a strange reversal of things, it was formerly be
lieved that verbs started with the infinitive. “ Men ,”
says a writer of the beginning of this century , " ex
press themselves at first only in a general manner ;
and it is but later that they attain to analysing, to
particularising each idea. In proportion as languages
this explanation. The substantives which have to be presupposed
are usually wanting. Moreover, we clearly see two inflections
overlaid in verbs like condewieren, in French conduire, “ to con
duct ” ; we are therefore justified in admitting an analogous super
position for the others.
1 We find as early as in the apocryphal Gospels : oéia Iva
επιβουλεύσωμεν. - 11ρέπει να αποστείλωμεν.
 84 SEMANTICS

arrive at their maturity, the infinitive forms disappear,

but in a just proportion : they are still of use in giving
variety to the style, although it may be already percep
tible that they are becoming less frequent.” It would
be impossible to shut the cyes more resolutely to the
+ truth. The infinitive sums up centuries of effort : it is
the most recent of verbal forms.
Like the infinitive, the passive is of the number of
those methods of expression which one is tempted to
imaginc much more ancient than they are.
Sylvestre de Sacy, who wrote for the use of his
children the Principes de Grammaire Générale, represents
the passive as one of the two necessary forms of the
verb. He gives for this three reasons. The passive is
necessary : first, when you wish to express an action
without naming the acting subject : " I am afflicted ” ;
secondly, when you wish to bring out the object which
suffers the action rather than the subject which performs
it : “ The Roman empire was founded by Augustus ” ;
thirdly, to vary the discourse and prevent monotony.
A philologist of a different school, but too prone to
thcorics, Hartung , explains the active and passive by
reducing them to directions in space. The active
answers to the question quo (whence the accusative); the
passive responds to the question unde (whence the
ablative or genitive ).
It is unnecessary to point out the artificiality of these
cxplanations. The passive is not an ancient form : this
may be guessed by merely noting how widely pépopal
· Encyclopedia of Ersch and Gruber, iii. vol. xiii. p. 172.
 NEW ACQUISITIONS 85

and feror differ as to inflections. The passive is a form

acquired gradually by the diverse Indo-European
languages long after the system of their conjugation was
completed in its principal lines. It was by taking
possession of the reflexive form that the greater part of
them, and especially Latin and Greek, contrived to
create a passive voice.
To understand how the reflexive form can take the
place of the passive, it will be enough for me to quote
a few French sentences, in which, even at the present
day, the same turn of phrase is used :
“ Les grands poids se transportent mieux par la voie
maritime. "
" Cette forme de vêtement ne se porte plus."
“ Ces événements se sont vite oubliés.”
" Le monde de la nature se divise en trois règnes.”
And in Italian : dicesi, temesi. And even : avvenimenti
compiutisi.
Not that the idea of the passive was difficult to con
ceive : “ I am stricken is no less easy to understand
than “ I strike. " Thc difficulty arose elsewhere : it
arose in the general plan of our languages, which is
opposed to the passive idea ; for the Indo -European
languages have always presented the phrase in the form
of a little drama wherein thc subject is always active.
Faithful to this plan, they still say : " The wind shakes
the trees. ” “ The smoke rises into the sky.” “ A
polished surface reflects the light." " Anger blinds the
mind ." "Time passes quickly .". " Two and two make
four. " Each of these propositions contains the enuncia
86 SEMANTICS

tion of an act attributed to the subject of the phrase.

Even the passive itself had to be imagined under the
form of an act.
It is this indeed that our languages have realised.
They crcated the passive more or less tardily by present
ing it in the form of an act returning upon the subject.
Pascitur mcant “ hc nourishes himself," before it meant
" he is nourished." Aldokopal mcant " I tcach.myself,"
before it meant “ I am taught.” On this point the
Germanic and Slav languages are particularly instruc
tive. We find in them the successive stages of the
metamorphosis. In old Norsk their finna sik means :
they find each other. " Thence arose a form their
finnask, meaning " they are," " thcy exist,” and finally,
" they are found ” (that is to say, inveniuntur ). The
same thing appears in Lithuanian and in Slay. It was
indccd the Letto-Slav family which, by the transpicuity
of its forms, first pointed the way to the origin of the
passive.
We have herc a fresh example of the intention that
presides over the evolutions of Language, siinultancously
with an example of the almost childish simplicity by
which that intention attains its ends. The passive
seemed directly opposed to the idea expressed by our
verbs : nevertheless, the passive has, by an identical
method, found its expression in idioms far removed
from cach other.
I will give yet another example of this intelligence,
hidden and yet so alert, which takes advantage of the
smallest accidents to furnish thought with new resources.
 NEW ACQUISITIONS 87

Everybody knows that the adverb is an ancient adjec

tive or substantive which has abandoned the regular
plan of declension. It is in this way that primum ,
ceterum , potius are ancient accusatives, that crebro, subito,
vulgo arc ancient ablatives. But whcncc comc the
adverbs in -e, like pulchre, recte ? This has not up till
now been sufficiently investigated.
Latin was wont to change its substantives or adjec
tives from one declension to another, when they were
lengthened by a prefix , or entered into a compound.
Animus made exanimis, fama made infamis, clivus made
proclivis, poena made impunis, and so on. The ablative
of these words in -is was eid or e . At an epoch when
the Latin language was not yet fixed there was therefore
a choice between infirmus or infirmis, præclarus or
præclaris, the ablatives of which were infirmo or infirme,
præclaro or præclare. Usage has not failed to profit
by this double formation : it has given the preference
to the form in e, which stood out better from the ordinary
declension. Not only was this form prcfcrred for the
adverb, but it was also gencraliscd, so that it produced
words like firme, clare. The Oscan amprufid, which corre
sponds to the Latin improbe, is a witness which permits
of no doubts as to this origin. The Latin language thus
entered into possession of an adverbial inflcction, of
which it has, as is well known, made the widest use."

An observation of a rather different nature here

presents itself. We have just quoted two or three
· See Mem . Soc. Ling., vii. 188.
88 SEMANTICS

examples of acquisitions made by our languages. They

are certainly both valuable and important. Yet how
ever useful they may be, they do not either in value or in
number, approach the acquisitions previously capitaliscd.
By thcsc I mean that grammatical machinery which
constitutes the common stock of the Indo -European
languages, and which was alrcady of ancient date and
completely fixed when Sanscrit, Greck, Latin, Germanic,
Slav, Celtic appeared for the first time. Therein , if I
mistake not, lies a means of weighing the antiquity of
the Indo -European languages.
By the antiquity of the Indo- European languages, I
do not mcan thc antiquity of a race, a thing difficult
both to conceive and to understand, but the antiquity.of a
civilisation. For a grammar and a morphological system
to attain to the degree of unity and of fixity that we
find at the base of Aryan languages, a certain perpetuity
in tradition is ncccssary. This perpetuity presupposes,
if not a literature, at Icast forinulas, folk-songs, sacrcd
texts transınitted froin age to age.
As there is no rcason to suppose that in thosc carly
ages history pursued a swiftcr course, we are thus
cnabled to cstimate the extent of the past. We have
just scen how long a time was needed for each of our
languages to acquire an infinitive, a passive, adverbial
1 We might also cite, in the Slav languages, the creation of the
" animated gender," which rests on the morphological distinction
between substantives designating beings endowed with life and
those which are without it. This distinction arose subsequently
and owing to a mere accident of language. See the work of A.
Meillet, in the Bibliothèque de l'École des hautes études.
 NEW ACQUISITIONS 89

inflections. And further, a choice is not definitely

determined till after long centuries, Moreover, the
acquisition of new instruments, such as the article, the
auxiliary verbs, has required no shorter timc. Wc
should thereforc conccdc to the period that went beforc,
in itself of far grcater importance, at Icast an cquivalent
number of conturics. As the historical period that we
are able to survey, from the first Vedic hymns until our
own time, amounts to about three thousand years, it is
not too much to ask another three thousand years for
the period which went before it. No less was necessary
to separate noun and verb , to establish the conjugation
and the declension, to eliminate the useless parts, to
create the mechanism of noun -formation, to devise a
pronominal declension corresponding to the substan
tival, to allow Analogy to establish the beginning of its
cmpirc, finally to lay thc foundations of Syntax.
If wc admit for the past the samc mcasure of time as
is furnislied by obscrvation of modern epochs, six thou.
sand years are the ininimuin which can be estimated as
the period of civilisation represented by our family of
languages .
 CHAPTER VIII

EXTINCTION OF USELESS FORMS

Difficulty of this study - Superabundant forms produced by

grammatical mechanism-Advantages of Extinction -Are there
any forms irrevocably doomed to disappear ?
THE Extinction of useless forms is not to be understood
only of those which, after having existed for a longer or
shorter period, have dropped out of usage ; but also of
those forms which, virtually possessing the right to live,
have yet never attained rcalisation. It will be under
stood that we are now entering the domain of hypothesis.
Nevertheless this kind of verbal infanticide has its place
in the history of Language.
If you consider the matter from the standpoint of a
mere statistician, over-production will seem inevitable.
Suppose the Greek language carried out, through all
tenses and all moods, the three verbs, leitw , lítw, and
Aluntávw , which all three signify “ to leave," or the thrcc
verbs βίβημι, βαίνω, and βάσκω, which all three signify
“ to walk ," there would be such an abundance of forms
that the mind would be overwhelmed." But everybody
I See above, p. 37.
90
 EXTINCTION OF USELESS FORMS 91

knows that it is not so ; the semi- conscious wisdom

which presides over the elaboration of Language brings
about the elimination of uscless forms. What is of no
use is suppressed. Hence the composite conjugations.
Hence paradigms such as : λείπω, έλιπον ; βαίνω, έβην ;
λανθαίνω, έλαθον.
Although composite, these conjugations do not cease
to be regular. As it is in the nature of the popular mind
to proceed with regularity, it imports order also into its
radiations. The second aorist has cverywhere inherited
the shortest forms, while the present has usually kept
what remains of the forms which are most developed.
The play of the Greck conjugation is therefore due to
a succession of affluence and poverty. Not that there
remain no useless riches. Sanscrit has as many as seven
different formations of the preterite. Certain Greek
verbs have two aorists, two futures, two perfects. But in
proportion as languages grow old , they rid themselves of
thcir superfluity. This fluctuation, which gives to thc
Homeric language a choice between three or four forms,
no longer exists in thc Grock of Lucian.'
The Extinction of useless forms goes the length of
uniting different verbs in one and the same conjugation :
fero, tuli ; opów, cidov ; déyw , citov, cipnka ; je vais, j'irai,
je suis allé ; I go, I went ; I am, I was, I have
been, etc., etc. Our grammars represent these as de
fective verbs which have reciprocally completed each
· It has moreover been imagined, not without some semblance
of truth , that the fluctuation of the Homeric tongue would have
been less great but for the adınixture of several different recensions.
1

20
S
 EXTINCTION OF USELESS FORMS 93

is what occurred in Latin with tõtus, which presupposed

1
a correlative quötus. The Latins must first have said
tota terra , quota est. We see how, by mcans of Suppres
sion, the Latin language acquired a word signifying
" all ." The same thing occurred in Grcek . At first
a pronoun râs inust have answered to tâs. Suppres
sions of these various kinds are not losses : on the
contrary , Language gains thereby in rapidity and in
force.

Languages may be judged by what thcy pass over in

silence as well as by what they cxpress. When we
obscrvc other familics, we scc that those who laid the
foundations of the Indo-Europcan grammar werc rcla
tively modcratc. Thc declension sccms never to have
had more than a limited number of cascs. The conju
gation, though more exuberant, never attained to the
developments which we find elsewhere. It does not
mark the gender ; it docs not make the distinction
between momentary and prolonged action ; it has
avoided idle honorary distinctions ; it has not tried to
include too many things in one single word .2
Our languages have, in general, abstained from mark
ing many idle distinctions, which, being merely super
ficial, are but a frivolous expenditure of intellect. In
· Not to be confounded with quotus, which is a derivative of the
noun of number, quòt.
? It says, for example, in one single word : loraya.— “ I place
myself” ; Totadai, “ thou placest thyself ”; Totatai, " he places him
self.” But it has never tried to say in one word : " I place you," or
" he places me."
94 SEMANTICS

Japanese, for example, words change according as the

speaker takes count of quadrupeds or fish, days or
dimensions. In Basque there is a ceremonial conjuga
tion . As there are profound differences in the art of
different peoples, this delighting in details, that grasping
at nature in her grander lincs, so in Language also there
may be obstruction and superfluity . The Extinction of
useless forms, whether they perish because abandoned
by a more mature reason, or whether they are arrested
by the mind before their conception, has therefore its
nccessary part to play.

It is interesting when the same idea is represented by

two synonymous terms, to see how Language gets rid of
one of the two, though not so completely as to leave no
traces. The name for an old man is yépwv in Greek,
senex in Latin : the two terms co -existed alongside of
one another in an anterior period, and we find in Sans
crit, by the side of garan, which corresponds exactly to
yépwy, the word sanas, " old,” which is of the family of
senex . Greek made its choice, Latin did the same : but
they chose differently. Greek, however, still says žval
åpxai in opposition to véai) to designate magistrates
giving up office ; it also says & vot kaptol to designate the
fruits of the past year. The language of politics and
agriculture therefore retained in an exceptional manner
the synonym that had otherwise dropped out of use.
Latin, on thc other hand, designated a man worn with
" Sayce, Introduction to the Science of Language, i. 205, third
edition.
 EXTINCTION OF USELESS FORMS 95

years, by a - ger ( for ævi-ger), a compound the second

part of which is the root of yépwvel Composition has
here saved the synonym which everywhere else has been
sacrificed. This demonstrates all the more clearly the
adjustment which took place in both languages.
Latin having expressed the idea of hcaring by the
periphrastic expression audire, which means strictly “ to
receive in one's car,” ? the ancient verb cluo became
thenceforward useless, and was bound to disappear. But
it is the substantive cliens ( cf. the German der Gehörige )
which proves that it did, in a more distant age, exist in
Latin.

Are there any Extinctions of words or of forms which

are imposed by Phonetics ? This has been frequently
maintained . Yet, when we see with how little difficulty
the popular instinct saves what it does not wish to lose,
we begin to doubt this so-called necessity. If there was
one word more than another threatened with Extinction
during the transition from Latin into French it was the
word avis (oiseau ). Yet we see how easily it was
preserved and multiplied under the forms oiseau (avicel
lus), oie (avica, anca ), oison (aucio). If a verb is in
question, the frequentative comes to take the place of
the simple form : premere, pellere would have had trouble
1 In Sanscrit, gar, “ to be worn out, to grow old. ” The participle
girna, for example, applies to worn-out clothes. The contraction
of the first member is the same as in @ -tas ( for ævi-tas), a - ternus
( for avi- ternus).
? From aus (Greek oớc), “ ear," and dio (cf. con -dio), “ to place."
Compare the synonym aus-cultare.
 S
NTIC
96 SEMA

to gain admittance in French : but presser, pousser are

used . The verb flare yielded but little ; but compounds
like suflare ( souffler), conflare (gonfler), were adopted.
Latin , it sccms, found a difficulty in distinguishing
between certain synonyms. There were two verbs,
luere, one signifying " to wash ," and the other of a
precisely opposite sense, since it meant " to sully, to
soil " ( cf. lues, a stain). But the language avoided the
confusion without difficulty, by means of the compound
polluere, which assumed the signification of the simple
verb .
Here again, as in all the laws which we have studied
in this first part, we find an intelligent thought, not a
blind necessity, at work.
Wherever we look attentively, we see vanishing that
so-called fatality which, we are told, should be the law
of Language. The phonetic laws do not reign uncon
trolled ; they are no more capable of destroying an
indispensable or even merely useful word , than thcy are
able to prolong the existence of a superfluous form ,
 PART II

HOW THE MEANING OF WORDS IS DETERMINED

H
 CHAPTER IX

THE SO -CALLED TENDENCIES OF WORDS

Whence comes the " pejorative tendency " ? — The " tendency to
deteriorate "-Other tendencies no less imaginary.

In this second part we propose to examine by what

causes words, once created and endowed with a certain
meaning, are induced to restrict, to extend, to transfer X
this meaning from one order of ideas to another, to raise
or to lower its dignity, in short to change it. It is this
second part which, properly speaking, constitutes
Semantics or the Science of Significations.
An illusion against which warning would seem to be
superfluous, which is none the less of frequent occurrence,
and which at times arrays itself in the garb of science,
is the error summed up by the name “ tendencies of
words.” Nothing could be more chimerical. How
should words have tendencies ? Nevertheless, we hear of
a pejorative tendency, of a tendency to deteriorate, etc.
One eminent philologist has published a very instructive
work, entitled : Ein pessimistischer Zug in der Entwick
lung der Wortbedeutungen . Another writer, Mr. Abel,
i Reinhold Bechstein in the Germania of Pfeiffer, vol. viji.
99
 /

100 SEMANTICS

in a memorandum on English verbs expressing an idea

of command , says that to command has a tendency to
descend, but that it inclines always in a good sense .
These tendencies must take their place among those
" forces " with which the science of the Middle Ages
peopled nature. We might as well interpret our econo
mists literally, when they say that the metal silver has a
constant tendency to fall in valuc.
The so -called pejorative tendency is the result of a
very human disposition which prompts us to veil , to
attenuate, to disguise ideas which are disagreeable,
wounding, or repulsive. Aulus Gellius draws attention
to the fact that the word periculum could formerly be
taken in a good sense : and in fact its literal meaning is
"experience." ! That it has come to possess a disagree
able sense is merely the result of euphemism ; in the
same way in French an army put to the rout is said
to have been " éprouvée." Valetudo signifies “ health " ;
but it has come to mean the contrary , as when we say
that So-and-So has “resigned on account ofhealth .” To
tell a man that he is lying is a serious matter ; we prefer
to speak of his imagination. And this was first
expressed by the verb mentiri, which is formed from
x mens as partiri from pars, or sortiri from sors. The
German List, “ cunning," began by being a synonym
of Kunst, " knowledge, cleverness.” 2 Gottes List, “ the
wisdom of God ," is an ancient expression. The English
silly answers to the Anglo-Saxon saelig, to the German
1 of the same family of words which has given experiri, peritus.
From the Gothic leisan, “ to know ."
 THE SO -CALLED TENDENCIES OF WORDS IOI

selig, and meant originally " happy, tranquil, inoffensive." 1

Examples might be indefinitely multiplied. There is
nothing in it all save a feeling of consideration , a pre
caution against unnecessary shocks,-a precaution which
whether sincere or feigncd is not long efficient, since the
hcarer seeks out the thing behind the word , and at once
identifics thcm.
The so -called pejorative tendency has yet another
causc . It is in the nature of human malice to take
pleasure in looking for a vice or a fault behind a
quality. The French have the adjective prude, which
had formerly a good and noble acceptation, since it is
the feminine of preux. But the spirit of the narrators
( perhaps also some feeling of rancour against the loftier
virtues) turned this adjective aside towards the equivocal
sense that it now bears. Words which refer to the rela
tions of the sexes are especially exposed to changes of
this kind. We remember what a noble signification
amant and maitresse still possessed in Corneille. But
they are dethroned, as was Buhle in German. Here we
sec the inevitable results of a false delicacy ; honourable
names are dishonoured by being given to things which
are dishonourable.
In Middle-High -German , Minne expressed in a general
way the affections of the soul : remembrance, friendship,
love, and even the love of God . But towards the end
of the fifteenth century, the word had to be banished
1
Cf. the German albern, “ silly," which corresponds to the Old
High -German alawar, " good, friendly. " Likewise simple in
French, einfältig in German.
102 SEMANTICS

from the language as opposed to decency. It is only in

our own days and thanks to studies on the Middle Ages
that, after a long period of exile, it has been brought
back to honour.

With regard to this so-called pejorative tendency, we

should, to be just, also postulate a mcliorative tendency.
Politeness has singular refincinents, and affection curious
windings which cause certain terms with an unfavour
ablc mcaning to lose their disagrccablc element. Friend
ship, as though in want of appropriate adjectives,
changes blame into praise, and turns reproach to more
highly - flavoured eulogy. The Italian vezzoso (vicious) is
defined as “che ha in se una certa grasia e piacevolezza."
The English smart (the same that produced Schmerz in
German) has become synonymous with “ sprightly,
lively, pretty ."

The Deterioration of words arises from another fact

which is no less common, to wit exaggeration . Aflicted
originally meant “crushed, broken by grief ” ; it has
lost much from having been used out of season. Abimer
had in French the same fate as fatigo in Latin , which
originally possessed a very strong and noble meaning
Gåter, meurtrir, gêner, tourmenter, are examples of the
1 Virgil uses it where speaking of the persecutions of the gods :
Aspera Juno
Quæ mare nunc terrasque metu cælumque fatigat.
It is related to fatisco. Fessus, which is of the same family, has
itself lost much of its force.
* Already in Latin : Ne torseris te ( Pliny the Younger, ix. 21 ).
 THE SO -CALLED TENDENCIES OF WORDS 103

same kind. In English , to be anxious to see you merely

means to wish to see you. In modern Greek aduvw , “ to
labour at," has become the ordinary term for " to do " ;
κάμνετε μoί την χάριν, “ do me the favour.”
As is shown by the last example, Deterioration is
often accompanied by a kind of discoloration, which
ariscs from the fact that the word is used in all kinds
of associations. The German adverb sehr (which should
be written sir) signifies " cruelly." 1 The Germans used
to say : er ist sehr leidend, sehr betriibt. But the dis
coloration has been so great that they have ended by
saying : er ist sehr brav, sehr froh.

Any one who takes his stand on etymology without

paying attention to the Deterioration of meaning may
be led into strange errors. How much has been written
on the compelle intrare of the Gospels ! These words are
the translation of the Greek αν 'γκασον εισελθείν, which
signifies “invite them to come in . " ? There is no question
of compulsion.
The Latin invitare, which expresses the same idea, is
a derivative of invitus. It began by signifying " to do
violence .” But an excess of politeness caused it to be
employed on occasions which, from the time of Cicero,
gave it the meaning of “ to invite. "
The German verb nöthen or nöthigen is an example of
the same fact.
1 Versehren , “ to ravage," unversehrt, unwounded ,” are of the
same family. The head of the family is the Old -High -German
sêr, “ pain ."
· St. Luke xiv. 23.
 104 SEMANTICS

Another tendency which it is no less chimerical to

attribute to Language, instead of seeking the cause
among the facts of history, is the tendency to level.
Herr, in German, was a title reserved to gentlemen ; it
is the comparative of an ancient adjective signifying
raised . " The House of Lords at Berlin is still called
das Herren Haus. But at the present day the title is no
finer than Monsieur in French.
There are decadences that extend even to the pro
nouns. Er and sie from having been formulas of polite
- ness, like ella in Italian, have descended from their high
estate, because a refinement of obsequiousness, in order
to go one step higher, substituted the plural pronoun for
them.2
The propensity to generalise what at first was made use
of by the minority alone, accounts for some facts that are
disconcerting at first sight. Client, in Latin, meant " he
who obeys, the servant.” : A patrician at Rome had
clients. Later on, the word designated one who when
summoned before the tribunal invoked the protection of
a patron in his defence. But in our own days this
expression, having passed through the hands first of the
doctor, then of the merchant, has ended by assuming a
false meaning, for it is contrary to etymology to bestow
the name of " one who obeys” on him who gives the
orders.

1 For the lower classes the word Meister was used . For exam
ple, Herr Hartmann von Aue, Meister Gottfried von Strassburg.
* See Grimm's Dictionary, at the word er .
3 See above, p. 95.
 THE SO-CALLED TENDENCIES OF WORDS 105

In our modern societies, the meaning of words is more

quickly modified than was usual in antiquity and even
in the generations which immediately preceded us.
Herein we see the effect of party warfare, of the
mingling of classes, of the strife of interests and of
opinions, of the diversity of aspirations and of tastes.
We need only remember to what a pitch of contempt
the formerly respected term bourgeois has arrived in
France : to such a point indeed that German literature,
in order to sound the same note of depreciation,
borrows the French word , reserving for Bürger its
primitive value.
Another cause of acceleration lies in industrial pro
duction : thinkers and philosophers have the privilege of
creating new words which arrest attention by their
amplitude, and by the learned aspect of their structure.
These words pass into the vocabulary of criticism and
so gain currency among artists ; but once admitted into
the studio of the painter or sculptor, they speedily come
forth in order to spread through the world of industry
and commerce, which makes use of them without
measure or scruple. So that in a comparatively short
time the vocabulary of metaphysics is helping to nourish
the language of advertisement.
Language, as is seen , undergoes outward fluctuations
in various ways. But besides these changes arising from
extrinsic causes, there are also changes which are explic
able only by the very nature of Language ; these we will
try to make known .
1 [And English literature too .)
 CHAPTER X
RESTRICTION OF MEANING

Why words are necessarily disproportionate to things - How the

mind readjusts this disproportion.
One fact which dominates the whole subject is that
by a necessity, the reasons for which will appear, our
languages are condemned to a perpetual lack of propor
tion between the word and the thing. Expression is
sometimes too wide, sometimes too narrow. We do not
notice this want of accuracy because, for the speaker,
expression adapts itself to the thing through the
circumstances, the place, the moment, and the obvious
intention of the discourse. At the same time the atten
tion of the hcarer, who counts for half in all Language,
goes straight to the thought behind the word, without
dwelling on its literal bearing, and so restricts or extends
it according to the intention of the speaker.
As the facts of Restriction are of the most frequent
occurrence we will examine them first.
To designate the roof of a house the Latins used the
word teg -men , formed from a verb, tegere, " to cover," and
a suffix -inen , which serves to mark the instrument. But
106
 RESTRICTION OF MEANING 107

tegmen was just as suitable for and also used to mark

the shelter furnished by a troc, a cuirass, or any kind of
cover or wrapping. If in the place of tegmen I have
recourse to tectum , I find a word which, though already
more restricted by usage, offers almost the same combin
ation of verb and suffix . Tec - tum designates everything
that is covered , consequently the ceiling of a room , the
vault of a cavern, the tester of a bed, as well as the
roof of a house. We must come down to the French toit
to find the word at last sufficiently contracted by usage
and (it must be added) sufficiently obscured in form to
be uniquely and especially suitable to the covering of a
house.
By this first example a glimpse is afforded us of the
real cause of the disproportion between the name and
the thing .
It comes from the fact that the verb is the essential
and capital part of our languages, which serves to form
substantives and adjectives. Now by nature the verb
has a general signification, since it marks an action taken
in itself, with no other determination of any kind.
When this verb is combined with a suffix the verbal idea
may be attached to an active being, to an object which
undergoes the action, or to an object which is the
product or instrument of the action. But as this action
keeps its general signification, the substantive or adjective
thus formed will itself possess a general meaning. It
must be limited by usage.
1 It would be more accurate in the case of the most ancient
words, to say verbal root instead of verb .
108 SEMANTICS

From this fundamental condition of our languages

arises the enormous quantity of words with a general
signification which have in course of time assumed a
special meaning. In proportion as a word becomes
restricted, Language is compelled to have recourse a
second, a third, a fourth time to the same verb. It is in
this way that by the side of tegmen we find tegmentum ,
tectura, tegumentum , tectorium , teges, toga, all words which
began with a general meaning, and were eventually
reduced to a certain category of objects.

There existed in Latin a substantive felis or feles,

which signified " the female.” This name was suitable
to the female of all animals, at least of all mammiferous
animals." But little by little it came to mean merely
the female of cat, and it is with this signification that
it has descended to us. How are we to explain this
Restriction of incaning ? The ancients, who were not
unaware of facts of this kind, saw in it the result of a
choice, a preference (kar' Eoxúv) . But in reality it is a
simpler matter. There was no choice, or at least the
choice came about of itself. When the Greeks of the
present day call a horse adoyov, it does not mean, as has
been said, that the horse is pre-eminently the animal ,
still less " that it only lacks speech," but that the rider,
speaking of his mount, was accustomed to say "the
animal."
Each profession, each state, each class of life con
· From fela, “ breast.” It is well known that the same root fe,
“ to suckle," gave filius.
 RESTRICTION OF MEANING 109

tributes to this contraction of words, which is one of the

most instructive sides of Semantics. In Rome, hay was
called by the most general term - fenum , “ produce .” For
the Greek peasant cattle were tà ktuara, “ possessions."
In Greek, an adventurer was called melparns, from the
verb teipáw , "to try, to undertake” ; but if we consult
the usage of the language, we see that there is question
of but one kind of enterprise, brigandage by sea, piracy.

The more advanced the civilisation of a nation, the

more varied are these Restrictions of meaning ; each
class of the population is tempted to employ for its own
use the general terms of the language ; it then restores
them with the impress of its ideas, of its particular
occupations. Thus the word species, which expresses
" kind " in the most general way, was used by the
druggists of the Middle Ages for the four kinds of in
gredients in which they traded ( saffron, cloves, cinnamon,
nutmeg ), so that when the word returned to the language,
it had become spices ( épices ).
It would be easy to multiply these examples. Every
one knows the divisions by means of which dictionaries
separate the different significations of a single word.
In most cases it is but a general word, the meaning of
which has been diversified by Restriction.

No one, when employing these words, thinks of the

lack of proportion. They are , for the moment, really
adequate to their object. If, for any reason, a word
grows obselete in all its acceptations save one alone, it
110 SEMANTICS

goes down to future ages, for the greater astounding of

the future etymologist, with the unique value which has
been left to it. The German word Getreide (in Middle
High -German getregede) is a derivative from the verb
tragen, “ to bear," and could originally be employed for
anything that is carried, such as clothing or baggage; it
denoted also what is borne by the earth, more especially
corn, and it is in this single acccptation that it has survived.
The more general the signification of the verb, the
better adapted is it for the various professions. Thus
facio, in thc language of the temples, mcans " to bring
an offering, to offer a victim .” Hcncc cxpressions like
facere catulo, facere ture, " to sacrifice a dog, to offer
incense." This same verb facio, in political language, is
applied to the combined action of a party in view of a
goal to be attained. There were found on the walls of
Pompeii, which, as is well known, was cngulphed in the
midst of an electoral period, a quantity of inscriptions
with this imperative : Caupones, facite . . . Pomari,facite
... Lignari, facite . . . Unguentari, facite. ... This
means : “ Agree ! unite !” Henceforward we understand
the meaning of the word factio. What characterises
the faction is the tie, the pact which binds together all
adherents.2

· Cicero wrote that all who had lost reputation were collected
round Cæsar : Omnes damnatos, omnes ignominia affectos illac
facere. ... Compare also the expression : tecum frcio (I make
common cause with you).
* Taken in this sense the contrary of facio is deficio. What a
faction or a party is least disposed to forgive, is the defection of
one of the members.
 RESTRICTION OF MEANING III

Adulterare is a compound of alterare : it had nearly

the same meaning. The Latins said adulterare colores,
" to change colours " ; adulterare nummos, “to debase
moneys " ; adulterare jus, "to twist the law ." But as
they also said adulterare matrimonium , there has arisen
a special meaning which has passed into the derivatives
adulterium and adulter.

We see how ncccssary it is that our knowledge of a

language should be supported by history. History alonc
can give to words that degree of precision which is
needed for their right understanding. Let us suppose,
for example, that for a knowledge of Roman magis
tracies we have no help other than etymology. We
should find “ those who sit together ” ( consules), “ he
who walks in front” ( prætor ), “ the man of the tribe"
(tribunus), and so on. These words shed no light, take
no definite meaning, save through the recollections
which we may possess, from having seen them in the
narratives of historians, in the discourses of orators, in
the formula of magistrates. History, while explaining
these words, introduces into them at the same time
many accessory notions which are not expressed. It
acts after the manner of a glass, which, by contracting
the objects of sight, renders them more distinct. But
there is this difference, that while the best microscope
can show us nothing in the objects save what is actually
there, we are conscious in words like tribunus, consul, of
many ideas which are not present, and which exist only
in our memory .

1
112 SEMANTICS

Restriction of Meaning has a peculiar interest when

applied to words which bear on moral life. I will give one
or two examples, borrowed from the Germanic languages.
In German , the substantive Muth is now scarcely ever
used save in the sense of “ courage" : but we have only
to note a few of its derivatives and compounds, and to
compare a few idioms, to recover the meaning of soul
and intelligence which it formerly possessed. Gross
muth, “ generosity ”; Hochmuth,“ pride" ; Unmuth,“ dis
content" ; Vebermuth, " presumption " ; anmuthen, “ to
claim " ; cinmüthig, " unanimously ” ; Gemüth, " soul ” ;
wie ist es dir zu Muthe, “ in what mood are you ? "
muthmaassen, “to conjecture." It is no doubt owing
to having figured in compounds such as Rittersmuth,
Mannesmuth, that the word has been restricted to the
sense of bravery. The general meaning has been
preserved in the English mood.1
In the same way, Witz is scarcely understood to - day
save in a very special sense. But in former days this
term had a very lofty signification : it designated know
ledge or wisdom ( from the verb wissen ). There is no
need to go very far afield to recover traces of this
ancient acceptation : it is obvious in Aberwitz, Vorwita,
Wahnwits, and in the verb witzigen, “ to make wise ."
Here also English has remained more archaic : wit,
wits ; though in English, too, the restricted meaning is
the more common.

· Note the change of gender which has befallen some of these

German compounds : die Sanftmuth, die Wehmuth .' Originally
Muth was neuter.
 RESTRICTION OF MEANING 113

The cause of these restrictions furnishes in each case

the material of an interesting research. Sometimes it
is a synonym which extends itself, and contracts by just
so much the domain of its colleague. At other times it
is an historical event which comes to modify and renew
the vocabulary. Thus the word Busse, which meant
“ reparation ” (whether literally or figuratively ), took
with Christianity, the sense of “ penitence " : once im.
pressed with the scal of religion all other usages came
to an end ."

In addition to the restrictions of meaning to which

language bears evident and lasting witness, there occur,
in the speech of all men, perpetual applications of the
same principle, which, varied according to time and
place, leave no durable traces. “ To go to town ” is a
phrase which is familiar to all country -folk, but which,
while remaining outwardly the same, must be translated
by a different name according to the region. The events
of history may chance to take one of these expressions
out of the limited circle in which it had place in order
to throw it into general circulation. Urbs was the name
of the town of Rome for the peasants of Latium and
Sabina. But the Roman legions, carrying the word
along with them , succceded in making it familiar to the
whole antique world : for the Gaul or the Spaniard, as for
the African or the Syrian , Urbs was the name designat
ing the city on the seven hills.
· Lückenbüsser, “ stop- gap,” is however still used. There exists
at Breslau an Altbisserstrasse, " street of cobblers.” Cauer, Pro
gramme du Gymnase de Hamm , 1870.
I
 114 SEMANTICS

Restriction of meaning has at all times been a cause

of astonishment to etymologists. We know the obser
vations and objections of Quintilian on the subject
of homo : "Are we to believe," he said , “ that homo
comes from humus, because man is born of the earth,
as if all animals had not the same origin ? ” ı Yet it is
most certain that homines did signify " the inhabitants
of the earth .” It was a way of opposing them to the
inhabitants of the sky, Dii or Superi.
11.6.
 CHAPTER XI
EXPANSION OF MEANING

Causes of Expansion of meaning - Facts of Expansion are so

much information gained for history — They are a consequence of
the progress of thought.
EXPANSION of meaning is the counterpart of what
we have just been observing. It may be thought sur
prising to find two movements in opposite directions
existing simultaneously. But we must note that the
cause is not the same in both cases : while Restriction
depends, as has been seen, on the fundamental con
ditions of Language, Expansion has an exterior cause :
it results from the events of history .
A few examples will make this clearer.
In Rome a landed property that had been mortgaged
was called prædium . The word is a compound of
vadium , “ pledge," 1 and of the preposition pre. But by
i Vadium is obsolete in classical Latin, where it is replaced by
vadimonium . But it has reappeared in the Latin of the Middle
Ages. The French gage comes from it. The Gothic ga -wadjan,
the Anglo-Saxon Weddian, whence the English wed, and the
German wetten, are, in my opinion, borrowed from Latin. Legal
115
116 SEMANTICS

a remarkable Expansion of meaning, all rural property

ended by being called prædium . It was probably through
the language of law that this change came about, dower
estates being called prædia dotalia.
The particular characteristic after which an object
has been named may therefore retire into the back
ground, may even be wholly forgotten. Instead of
designating one category only, the word comes to
designate the whole species.
The French substantive gain, “ gain ," bears witness to
the agricultural life of the ancient Gauls. Gagner
(gaaignier) was to feed flocks; a gagnage was a pas
turage ; the gaigneur was the husbandman ; the gain
(gain ) was the harvest. One piece of evidence which
has never varied still remains : that is the re -gain (“ after
math ” ). As life became more complicated, the simple
gain extended its meaning, so that it has designated
the produce obtained by all kinds of work , and even
that acquired without work.
To agricultural life belongs in like manner the
Latin pecunia, which first designated wealth in stock,
and which ended by designating all kinds of wealth. It
is a less known fact that the inverse change took place
in the Middle Ages among the Celts of Great Britain.
As a compromise had been established between the
ancient system of exchange in kind and the new system
terms, ab ich it was important to have a thorough unde
standing, passed from the Romans to the Barbarians. On this
family of words, see my Dictionnaire étymologique latin, at the word
vas, vadis.
 EXPANSION OF MEANING 117

of a monetary exchange, certain terms designated in

turn either a coin, or its equivalent in land or stock . In
Old Welsh scribl (Latin scrupulum ) is a coin ; among
the Welsh of the twelfth century , ysgrubl had the mean
ing of cattle, beast of burden . In Armorican Britain,
the Latin solidus became saout, which designates stock
in general. Among the Anglo-Saxons, on the contrary,
the ancient feoh,“ cattle," came to mean a sum of money .?
These facts, of which the contemporaries were unconscious,
are explained by alternations of wealth and poverty .
It is important for the historian to observe these
transformations of meaning, since they constitute for
him indications all the more sure because involuntary.
These facts must not be laid at the door of Metaphor.
Metaphor is the instantaneous perception of a resem
.
blance between two objects. Here, on the contrary , we
are dealing with a slow displacement of meaning. The
people continued, as a matter of habit, to use the word
pecunia, even when the fortune of the Roman citizen had
ceased to consist solely of flocks.
The general ideas which humanity has acquired in the
course of centuries could not have been given names
without this Expansion of meaning. How could time
and space have been designated ? Time ( temps) meant
in the beginning " temperature, heat.” The word is of
the same origin as tepor . Then, in French , weather,
' J. Loth, Revue de l'Histoire des Religions, 1896 ; article on the
Celtic by M. d'Arbois de Jubainville.
: Thence the English fee,“ recompense, salary.”
3 The neuter tapas, “ heat," exists in Sanscrit. The connection
between tempus and tepor is the same as that between decus and
118 SEMANTICS

good or bad (temps, bon ou mauvais ),was thus designated.

Finally the abstract idea of duration was reached .
Space was the course on which were run the chariot.
races ( spatium , a word borrowed from the Greek otáðlov,
Dorian otádlov). In speaking of horses which -swerved
from the course, the verb exspatiari was used. Cicero,
when he wanted to say that eloquence had gone astray,
said : Deflexit de spatio curriculoque majorum . Then the
word acquired the general meaning of extent and space.
The verb is the part of speech which presents the
most numerous examples of Expansion. When once
Language has, in one way or another, made choice of an
expression to designate an act, the circumstance — some
times indifferent or fortuitous — which caused it to be
thus named, is promptly forgotten . Who, in pro
nouncing the French verb briller, now thinks of the
precious stone, beryllus, from which it was derived ?
Those who created the verb plumbicare, which we have
turned into plunge ( " plonger " ), soon lost sight of the
plummet (“ plomb" ) which served to weight the net or
line, and applied the same expression to everything
which goes down , which plunges to the bottom of the
water. It is in the nature of the human mind to operate
in this way, for we are much more impressed by the act
itself, which is a present impression, than by the already
decor,fulgur andfulgor. Something of the idea of temperature
still remains in the verb temperare.
1 See the Mémoires de la Société de Linguistique, vi. 3. On
the subject of the substitution of t for d , cf. cotoneum = κυδώνιον,,
citrus = Képos.
 EXPANSION OF MEANING 119

far -off circumstance which caused us in the first instance

to name it.
There existed in Rome a census which recurred every
five years, and which was accompanied by a religious
ceremony, called "purification” : lustrum , lustratio. As,
on this occasion, the magistrate and the priests surveyed
the assembled ranks of the people, the verb lustrare took
the sense of " to survey, to pass in review.” Virgil,
therefore, speaking of the Ausonian Sea which must be
traversed by Eneas, could say :
Et salis Ausonii lustrandum navibus æquor.

But few people think when they say in French,

accablés d'un malheur, accablés d'une nouvelle (" over
whelmed by a misfortune, by bad news” ), that they are
generalising an expression borrowed from a siege, and
that the substantive cadabalum , which made caable,
whence accabler, is formed from the Greek karaþonń, " over
throw." Still less did the Romans, when speaking of
the splendour of the sky or of a splendid triumph, dream L
that it was to a sickly colour of the skin, to an un

wholesomeness of complexion, that the verb splendeo

owed its origin.

Expansion of meaning is especially frequent in the

case of compound words. When once two terms have
been joined together to make a whole, the mind no
1
Σπλην , " the spleen," A man with disease of the spleen was
called splenidus (cf. rabidus, from rabies). The ancients located
the seat of jaundice in this organ.
120 SEMANTICS

longer takes heed of anything but that whole. Vindemia,

for example, which contains the word vinum , is used for
other harvests beside the vintages : vindemia olearum ,
mellis, turis. Parricidium , which means the murder of a
father, expanded with the help of phonetic alteration to
the extent of marking all kinds of crimes, to such an
extent, indeed, that the Romans even were already seek
ing out remote etymologies of the word . We here touch
on what ancient rhetoric called an Abuse of Language
(catachresis). The truth is that catachresis exists only
in an early stage, and for one who clings to the
letter : for the generality of mankind, these expressions
quickly become natural and legitimate. Thus, in
Sanscrit, a stable for horses is called aqua-goshtha,
although goshtha is a compound containing the word
go, “ cow .” We find the same in Homer :
Του τρισχίλιαι ίπποι έλος κατά βουκολέοντο..
And the same Abuse of Language, under a slightly
different form , is to be found in this other line :
'Αρνών πρωτογόνων ρέξεις κλειτην εκατόμβην. 1
Just as it is right to recommend unmixed metaphors,
so also would it be puerile to hinder, by the recol
lection of their birth , the employment of words which
have been alienated through long usage from their original
signification, and in which there has never been any
question of Metaphor but only of Expansion of mean
ing. The progress of a language consists in being freed
1 The word βούτ, “ ox, ” being contained in βουκολίω and εκατόμβ
 EXPANSION OF MEANING 121

without violence from its origins. Speech would cease

if all words had to be restored to the exact mean
ing which they possessed in the beginning. Armare
naves is a common expression ; but it hides an Abuse
of Language, since armare signified " to cover one's
shoulders. " 1 We must leave to the philologist the task
of investigating these remote points of departure. The
Expansion of meaning is a normal phenomenon, which
must have a place among all peoples whose life is
intense and whose thought is active.
· Armus, “ shoulder," made armare, whence arma, which began
by designating defensive weapons, in opposition to tela, offensive
weapons. Armorum atque telorum portationes (Sallust).
 CHAPTER XII

METAPHOR

Importance of Metaphor for the formation of Language -

Popular metaphors - Divers origins of metaphorical expressions
They pass from one language to another.
UNLIKE the preceding causes, which work slowly and
imperceptibly, Metaphor changes thc mcaning of words
and creates new expressions on the spur of the moment. 1

It is born from the instantancous glimpse of a similarity

between two objects or two acts. If it be accurate or
picturesque, or even if it merely fill a gap in the
vocabulary, its adoption is assured . But the metaphor
remains such at its outset only ; soon the mind becomes
accustomed to the image ; its very success causes it to
pale ; it fades into a representation of the idea scarcely
morc coloured than the proper word .
It has been said that the metaphors of a people are
an indication of its genius. This is true in some cases ;
but it must be admitted that for the most part metaphors
According to the remark of Quintilian (viii. 6 ) it is owing to
Metaphor that everything seems to have its name in Language.
I 22
 METAPHOR 123

teach us little save what we knew already ; they demon

strate only the universal intelligence, which does not
vary much from one nation to another. We will cite
a few examples, with apologies to the reader for their
simplicity. Our business is not to claim admiration for
these images, which indeed have ceased to be images,
but to show how languages are full of them.
As we must fix some limit, we will borrow them all
from the same language - Latin. Let us see, for
instance, in what manner the Roman people namewhat
is good and what is evil,
What is good : that which goes straight and in order
( recte atque ordine), which is full and weighty (integer,
gravis). But lightness is a bad sign ( levis, vanus,
nullius momenti). What is crooked becomes the symbol
of all perversity (Aravus). Intelligence is like a point
which penetrates ( nucument), but rolly resembles a blunt :
knife (hebes), or a dish which lacks salt ( insulsus). A
simple character is compared to a garment which has
but one fold ( simplex ); motives which are wrongfully
alleged are borders to conceal the faults of the stuff
( prætextum ). The motley (vafer, varius) is not far
removed from deceit.
Up to this point the metaphors of the language are
irreproachable ; we will now observe certain features of
utilitarian ethics. To think is to count ( putare, reputare).'
1 Puture has itself acquired the meaning of “ to calculate "
through a metaphor. Putare rationes, “ to audit accounts. ” Varro
and Festus said putare, purum facere. It was the common expres
sion for the pruning of trees and vines : putare vitom , arbores.
The word, in its strict sense, is preserved in Old French : poder
 134 SEMANTICS

The valuation or weighing of moneys lends its name to

all kinds of cstimation or cstccm (æstimare, existimare,
pendere ). To delibcratc is also to weigh (dcliberare)."
Anything which can be bought cheap is despicable
(vilis); ? the value which we attach to objects arises
from their rarity (carus, caritas).
It is unnecessary to continue. We see at a glance
the nature of these examples. They are like the
sayings of some peasant endowed with good sense and
honesty, but not without a certain rustic cunning. They
must be placed below the proverbs, as proverbs denote
a more protracted experience, a larger faculty of
combination.
Here is another metaphor belonging to the same
order of idcas.
Among the old Romans all superfluous expense was
a breach of custom, a departure from upright principles.
Hence the word luxus, a word borrowed from the
language of surgery . Cato, giving a receipt for sprains
and fractures, said : Ad luxum aut ad fracturam alliga,
sanum fiet (De re rustica, 160 ). The word, like so many
1
medical terms, is perhaps of Greek origin : docós,
" crosswise " ; do fów , " to dislocate." We have made

pouer (" pouer et tailler la vigne " in Olivier de Serres) ; pod, “ to

prune,” in the dialect of the French-speaking Cantons of Switzer
land. This poder, “ to clean," has passed into German : butzen,
putzen (den Baum , den Strauch, die Hecke putzen ) ; then they said :
den Bart, die Haare putzen ; finally the word acquired the meaning
of dress, apparel (alie Putzmacherin, “ the dressmaker ' ).
I From librır, " scalcs."
; From the same root as venum, " a sale."
 METAPHOR 135

luxation out of it. There were no doubt many kinds

of disturbance comprised under this word. Occultiores
in luxus et malum otium resolutus, says Tacitus when
speaking of Tiberius.

We know how much trouble the ancients took in

classifying metaphors, and in labelling them according
to their kind and species. They rightly declared the
number to be immense. This number is even greater
than was supposed, for they were by no means cognisant
of all. Exstinguere had already taken the sense of
“ extinguish ” ; yet the flame is here compared to a dart
or lance whose point is broken . Erudire passed for
the proper word meaning " to instruct" ; yet it is
borrowed from the branch of a trec which has bcen
hewn into shape. In the time of Virgil, the word
tranquillitas, when applied to the soul, no longer pro
duced the effect of a figurative expression, although it
implies a comparison with the transparency of sky or 1

water . Sometimes the recollection of the metaphor is

so completely obliterated that mistakes arise. Cicero
was astonished that peasants should have had the idea
of giving the name of pearl (gemma) to the buds of trees ;
but the exact contrary is the truth, since it was pearls
which, thanks to an imagination not lacking in grace,
received their names from buds about to open .
· Quintilian, viii. 6. Arsène Darmesteter attempted a classi
fication, for which we refer our readers to the Preface, as yet
unpublished, of his etymological Dictionary.
• Memoires de la Société de Linguistique, v. 346.
8 Nam gemmare vites, luxuriem esse in herbis, lætas esse segetes
126 SEMANTICS

When the Science of Language bestows on the meaning

of words a part of that attention which it now fixes too
exclusively on the letter, it will be able to create for the
various languages a curious and instructive list showing
the contingent of metaphors furnished by each class of
citizens, by each professional body. The weaver gave
to the Latin language the words which mean “ to
begin ” : ordiri, exordium , primordia. Ordiri, was to
arrange the threads of the warp for making the woof.
Cicero, who was still conscious of the image, makes one
of his interlocutors say, not unintentionally : pertexe
Antoni quod exorsus es. Plautus had already said in
the same way :
“Neque exordiri primum, unde occipias, habes,
Neque ad detexundam telam certos terminos."
The word ordo, with its long series of varied and
important meanings—in politics, in war, in administra
tion, in the arts — is itself a gift from the humble trade
of weaver.
The auspices were of such great importance that it is
not surprising to discover traces of them in the common
language : the adjective propitius, which marked the
forward Aight ; the adjective sinister, which marked
etiam rustici dicunt ( De Or., iii. 38). Lætus, which Cicero regards
as a metaphor, is also the real word (“fat harvests").
| The vocable is probably much anterior to the Latin language.
We find in Hesychius this gloss : repðiós úpavons.
Or ., ii. 33. It is curious to note that the verb ordiri has
survived in French in its exact primitive sense, ourdir. The weaver
had furnished it : the weaver preserved it.
s From a root pet,which reappears in the Greek métouai, " to fly."
 METAPHOR 127

the fatal portents ; the verbs aucupari, "to watch " ;

augurare, "to conjecture " ; autumare, " to affirm ,"
which all three contain the substantive avis ; the adverb
extemplo, first used for presages arising in the interior of
the celestial templum ; the verb contemplari, borrowed
from the ordinary occupation of the augurs — all these
bear unanimous witness to this fact.
The language of law has been no less fertile. I will
cite merely the curious word rivalis, which originally
designated neighbouring proprietors using the same
stream of water, and which has become the name for all
kinds of rivalry .
The special characteristics of different nations
alrcady begin to show in some of the old metaphors.
Thus the Greeks, to express the idea of " resource,
expedient," use mrópos. “ What remedy for my woes ? "
exclaims one of the characters of Euripides. Τις αν πόρος
κακών γένοιτο.2 The word πόρος , which properly desig
nates a passage, and in particular a sea - passage, obvi
ously comes from a people which has been acquainted
from early days with the υγρά κέλευθα. An impossible
transaction was called άπορον πράγμα.. The revenues
of a state were called tópou. Even in modern Greek
èutropéw means “ to be able ."
Sometimes a whole historical perspective is discovered
in a metaphor. The Greek romance writer Longus, in
the history of Daphnis and Chloe, speaks of a wolf
There was in Rome a Lex rivalicia (Festus, p. 340 ) which
regulated the relations between rivales.
Alcestis, v. 213 . 3 Cf. Booropos, “ the Bosphorus."
 128 SEMANTICS

trap, a gin contrived in the ground. But the wolf

does not allow itself to be caught : αισθάνεται γάρ γης
σεσοφισμένης. This σοφίζω presupposes Protagoras,
Socrates, Plato , and a long past of philosophical
discussions.
The word influence, of which so much use is made at
the present day, takes us back to the superstitions of
ancient astrologers. It was supposed that a certain
Auid which reacted on men and on things escaped from
the stars. Boileau still used the word in its primitive
sense, when in his Art Poétique he spoke of the secret
influence exercised by the sky on the poet at his birth.
The Italian word influensa makes allusion to an
analogous belief.
All languages might in this way make their museum
of metaphors. In German, the verb einwirken, often
employed in the most abstract sense, answers to the
And similarly the Latin exprimere,
Latin intexere .
which reappears so frequently in this book, is borrowed
from the fine arts, since it suggests an impress : this
alone might teach us, were we not already aware of the
fact, that the ancients were acquainted with the art of
repoussé. So many obsolete customs are perpetuated
in an expression which has become commonplacc : in
saying of some great personage that he is invested with
a title or dignity, no one at the present day thinks of
the investiture.1

* How many expressions we owe to the theatre : to play a part

in an affair, to make a scene, a person behind the scenes, a tragedy
which took place yesterday, a quick change, a silent part, etc.
 METAPHOR 139

Language reserves one satisfaction for the observer,

all the more lively because it is not sought after : the
satisfaction, namely, of feeling a metaphor, whose value
has not hitherto been understood, suddenly open and
reveal itself. Thus wc establish a secret harmony
between our own thought and the ancient inheritance
of speech .
Nothing shows so clearly thc power which even at
the present day, with our languages long since fixed,
individual action continues to exercise. Such an image,
born in a capable mind, spreads into common property.
It ceases then to be an image, and becomes a current
appellation. There is the same difference between the
tropes of Language and the metaphors of poets as
between a product in common use and a recent conquest
of science. The writer avoids figures of specch which
have become common : he prefers to create new ones.
So Language is transformed . It is this truth that has
been sometimes forgotten by our etymologists, who are
always prone to supposé a so-called verbal root, as
though imagination had ever been at a loss to transfer
a word ready made from one order of ideas to
another.

A special kind of Metaphor, extremely frequent in all

languages, comes from the communication between our
organs of sense, which permit us to transport the

The very name of person — persona, --which Cicero already used

As we do, is a theatrical word, signifying " mask. ”
к
130 SEMANTICS

sensations of sight into the domain of hearing, or the

ideas of touch into the domain of taste. We speak of
" a warm reception ," " a broad style ," " a bitter reproach ,"
" a black grief,” with the certainty of being understood
by everybody. Modern criticism , which uses and abuses
this kind of transposition, only develops what is to be
found in the germ in the simplest language. A deep
sound, a high note were originally images.
The people transfer adjectives whose idea is borrowed
from man to inanimate objects : it says " a dumb waiter,"
" a blind alley," as the Greeks already said kwpòv
Béos (surdum jaculum ), for a shaft that miscarries, and
Médaiva pwvý (vox atrı) for a hoarse voice. The Hindoos
call andha -kupa, “ blind well,” a well the opening of
which is hidden by plants. Sometimes it is difficult to
tell exactly from what organ of the body these expres
sions came : for example, it was long considered
doubtful whether the adjective clarus came from sight
or from hearing. Without the words acies, acus, acutus,
acer, we should not know that acid (the French aigre)
did not always belong to the sense of taste.

The Homeric language did not lack words to

express the idea of “ meditating, preparing." But that
did not prevent the poet from creating Buorodopeuw
which signifies literally " intus ædificare."
"Εσθλ ’ αγορεύοντες, κακά δε φρεσι βυσσοδόμενον.
" Holding brave discourse, they builded evil in the
depths of their hcart."
 METAPHOR 131

And elsewhere :
' Αλλ' ακέων κίνησε κάρη, κακά βυσσοδομεύων .
“ He shook his head in silence, building evil within . "
For the same idea, Homer has also the verb unxavów,
which has passed from Greek into Latin ?
It is difficult to recognise the most ancient metaphors.
The state of things which suggested them having dis.
appeared, we are confronted with a root of colourless
signification. This cxplains to us how the Hindoo
grammarians, in drawing up their lists, could inscribe
so many roots meaning " to think, know, feel.” Were
it possible for us to go back farther into the past of
humanity, we should doubtless find Metaphor every
where present as in better known languages.
Before leaving this subject, which is infinite, we wish
to mention one more point.
Metaphors are not chained to the language which
gave them birth. When they are true and striking, they
travel from idiom to idiom and become the patrimony
of the human race . It is, therefore, for the historian
to make a distinction between the images which, being
perfectly simple, are found independently in a thou
1 Od ., xvii ., 66 , 465.— It is to be noted that is exactly the same
expression as the Latin industrius (from indu and struere ). Some
thing of the ancient pejorative sense has remained in the expres
sion : de industria .
2 Not always in an evil sense : ώναξ Παιάν, εξευρε μηχανάν τιν'
' Aguhtu kak (Euripides, Alc., 221 ). “ Find, Apollo, some help
for the woes of Admetus.” A man without resources, an impossible
thing, were called dutxavos.
132 SEMANTICS

sand places, and these which, invented but once in one

particular language, have been subsequently transmitted,
borrowed and adapted. Metaphors are translated, as
may be seen by such examples as decide and entscheiden,
discover and entdecken, comprehend and begreifen, succumb
and unterliegen, confirm and bestätigen . The difficulty
is to know in cach case whether there has been a loan,
and who is the borrower. Among the old nations of
Europe there exists a common fund of Metaphor which
arises from a certain unity of culture. Nations which
have attained rather late to the same degree of civilisa
tion are not slow to translate this stock of metaphorical
expressions and make it their own. It would hardly be
just to reproach them for it, since thcy do but use the
same right as their elders, a right from which there is no
reason to exclude them. I am thinking at this moment
of thc Greck nation which is accused of doing what
every nation of Europe has done in its turn. I will
give a single cxamplc. To express " I am not in
accord with you ," thc Grocks say : dyù deu oudwr @. Is
not this the same as the German : Ich stimme nicht mit
Ihnen überein ? Or merely : Es stimmt nicht ? Was
this to bc forbidden to them because we chose to create
the word symphony ? For that matter, the Greek has
1 On these imitations, examples of which are to be found in all
languages, see L. Duvau, in the Mémoires de la Société de Linguis.
tique, viii. 190 . An interesting specimen is the French com
pagnon, which has its prototype in the Gothic gahlaiba ( from hlaifs,
" pain ,” bread).
? See the imitations of Latin by old Irish, Journal de Kuhn, XXX.
255, article by Zimmer .
 METAPHOR 133

all the appearance of being in this case the original, and

we the imitators, since as early as in the Egyptian
papyri of the time of the Ptolemys we find oúupwvov in
speaking of a harmony made between two parties.
The law of metaphors is the same as for all signs.
A metaphor having once become the name of an object,
may again, starting from this second stage, bc cmployed
metaphorically, and so on . The result of this is that
the modern languages are more complicated study for
the philologist than the ancient. But for the child who
learns to spcak them the complication does not exist :
the last meaning, the meaning farthest removed from
the original, is often the first learnt. What is called
argot, or slang, is largely composed of metaphors, more
or less vaguely indicated, yet it is a language which can
be learnt as quickly as any other.
 CHAPTER XIII

ABSTRACT WORDS AND CONCRETION OF MEANING

What is to be understood by Concretion of meaning – Examples
drawn from various languages.

THE richness of our languages in abstract words is con

siderable. We shall later on have to investigate the origin
of this wealth, and the manner in which it has been the
most active instrument of progress. For the moment,
we wish to study a fact which I shall call, for lack of a
better term, Concretion. This is what it is : an abstract
word, instead of keeping its abstract sense, instead of
remaining the exponent of an action, a quality, or a state,
becomes the name of a material object. This fact is
very frequent; sometimes the modified word preserves
both meanings, sometimes, the abstract idea being for
gotten, the material signification alone survives.
This phenomenon goes as far back as the history of
our languages, and continues under our eyes. I shall
begin by examples drawn from ancient languages.
A very simple suffix which served to form nouns of
1 From the Latin concretio.
134
 . CONCRETION OF MEANING 135

action was the feminine suffix -ti ( nominative -ti- s),which

we find in Greek in the shape of -86-8 in words like
yéveous, “ birth " ; yvôois, " knowledge " ; xoñous, " custom " ;
Kplous, “ decision" ; TTWOLS, " a fall," etc. It is the suffix
seen in the Latin ves-tis, “ the action of clothing one
self.” But from this general meaning it came to mean
the object which served that purpose, and vestis was the
name of the garment. If vestis is feminine, it dates
from thc timewhen it was an abstract noun.
Let us take another example borrowed from alimenta
tion. The Latin suffix -tu - s produces abstract substan
tives like cantus, adspectus, gemitus, conatus, cultus.
Among thesc substantives appears fructus, "the action
of enjoyment," from fruor. Plautus still used it in its
strict sense. But this abstract noun has solidified to
the point of designating the fruits of the earth and of
trees, to such a point indeed that when we say " to live
on the fruits of one's labours," we seem to employ the
word in a metaphorical sense.
The suffix which gave in Latin thc nouns in otas, like
dignitas, cupiditas, in Grcek the nouns in -rns, like
dikalórns, " justice " ; piórns, “ friendship,” served to form
nouns expressing a quality or a state. But we find it
already becoming opaque in certain Latin words : civitas
meant originally the quality of citizenship ; then the
same word was taken to designate citizens as a whole ;
it ended by signifying " the city." Facultas, formed from
the adjective facilis or facul, denoted the possibility of
doing ; but facultates became a synonym of riches.
1 Casina, iv. 4, 16. Scio, sed meusfructus est prior.
136 SEMANTICS

The same suffix existed in Sanscrit and in Zend, under

the form täti or tåt. Already in the Vedas, dēva - tät
designates not only the divine quality or nature, but the
gods as a whole (as when we say Christianity)."
Legio meant originally “ levy ” : it is formed like
internecio, obsidio. Then it became the name of an
absolutely fixed military unit, " a legion.” To denote
the idea of a “ levy," new words such as delectus had to
be formed .
A similar change took place in the case of classis,
which is the Greek klous, Dorian kous, and which
became the Roman name for the fleet, after having first
designated the army in general. The primitive meaning
was “ roll - call.” 2
Regio, formed like legio, signifies "direction." Rectă
regione, “ in a straight line." E regione, “ in front.”
Deflectere de rectă regione, " to leave the direct way ."
But this meaning has given place to a far more material
sense : regio came to signify a country or the quarter of
a town.
The Latin suffix -tion, which became of such great
importance, and which is related to the preceding one,
formed abstract nouns such as lectio, admiratio. But

from the earliest times, Concretion began to make itself

felt. Portio was originally the act of sharing ; then it
became the name of the portion . Mansio was the act
· Rig Veda, iii. 19, 4 ; à vaha devatītim , “ bring us the gods.”
% It is curious to note that classe has returned to its ancient
signification in French military language.
3 From a root por, “ to assign,” which appears in the Greek
Itopov, “ I have procured ” ; aétpwtai, " it has been attributed .”
 CONCRETION OF MEANING 137

of stopping ; in Cicero it is opposed to discessus. It was

used afterwards for relays established at intervals along
the roads, and it finally produced the French maison,
the English " mansion.” 1
Our readers will now begin to see why so many
material objects are feminine ; from being abstract they
have become concrete, but without changing their
gender. Must we believe that our ancestors had a
faculty of abstraction which has gradually diminished
in their descendants ? That would, I think, bc a great
illusion. We will return later to this question of abstract
nouns, which contains, in part, the secret of the richness
of our languages. It suffices for the moment to remem
ber that, as Language is a work of collaboration, every
abstract word is in danger of changing its sense , when ,
by passing from mouth to mouth, it goes forth from the
inventor to the mass .
The history of religions, of institutions, and even of
sciences might furnish us with the proof of this. These
abstractions of Language, abandoned from the first
moment to the popular mind, were with still more
reason exposed to the same fate.
Modern languages abound in examples of a like
change of meaning ; we find in all professions abstract
We say in the same way habitations, constructions. Homer
even said of Ulysses at the time that he was about to build himself
a ship : eð eidus TEKTOO UVdwr, “ skilled in construction." .
9 There exist indications which allow us to believe that the Latin
nouns in -tus, like exercitus, amictus, were first of all feminine. We
find in Ennius : non metus ulla tenet. Cf. Greek feminines like
# partús, " action , " DENKTÚS, " enchantment."
138 SEMANTICS

nouns which have become the names of some tangible

object. The musician understands by overture the
orchestral piece which precedes an opera , the merchant
supplies the novelties of the season, the financier calls in
his credits, and so on. The stages of this transformation
can be easily observed in the case of certain substantives.
La Bruyère, in the portrait of Distrait, said : “Il écrit
une seconde lettre, et après les avoir cachetées toutes
deux, il se trompe à l'adresse. " Here adresse was still
taken in the sense of directio. In the seventeenth century,
economies, alms, charity, had not yet coagulated into
material objects as at the present day."
There is herc a mine of surprises for the etymologist.
We find in thc Venetian dialect of the Middle Ages a
word rità which has the meaning of “ descent.” Whence
comes this ritd, which, even by its inflection, leads the
reader astray ? Certain comparisons which cannot be
called in question have proved that we are here con
cerned with the word heredità , which by shedding its
abstract signification came to designate the heirs instead
of the heritage . Something of the same kind occurred
in the case of the German Kind, which mcans “ child ,"
but which first meant “ the race," as is shown by the
English mankind, “ human race. ”
Although the infinitive resists this change to a greater degree,
we yet observe that a certain number of French infinitives, such as
devoir, plaisir, loisir, have not escaped.
Rajna, in the Comptes Rendus de l'Académie des Lincei, 1891 ,
p . 336.
 CHAPTER XIV

POLYSEMIA
What Polysemia means - Why it is a sign of civilisation — Why
it causes no confusion - A new acceptation is equivalent to a
new word - Indirect Polysemia.
We have just seen some of the reasons why words
change their meaning. To be sure, they are not the
only causes, since Language, besides obeying its own
laws, is subjected to the rebound of outward events,
which evade all classification. But without prosecuting
this inquiry, which would be endless, we wish at this
point to make an essential observation .
The new meaning of a word, whatever it may be, does
not make an end of the old . They exist alongside of
one another. The same term can be employed alter
nately in the strict or in the metaphorical sense, in the
restricted or in the expanded sense, in the abstract or in
the concrete sense. In proportion as a new significa
tion is given to a word, it appears to multiply and
produce fresh examples, similar in form , but differing in
valuc.
139
140 SEMANTICS

We shall call this phenomenon of multiplication Poly

semia . All the languages of civilised nations have their
part in it. The more meanings a term has accumulated,
the more it may be supposed to represent the various
sides of intellectual and social activity. It is said that
Frederick II saw in the multiplicity of meanings onc
of the superiorities of the French language. He implied,
no doubt, that these words of manifold meanings were
the sign of a more advanced culture.
We must imagine Language as a vast catalogue in
which are recorded all the products of human intelligence.
Often , under the name of one exhibitor, the cataloguc
rcfcrs us to different classes.
Ict us give some examples of this Polyscmia.
Key, borrowed from the mechanical arts, belongs also
to music. Root, which comes to us from agriculture, is
equally connected with mathematics and philology.
Base, a term of architecture, has its place also in che
mistry and in military science. Act belongs at the
same time to the thcatrical and to the legal vocabulary.
And so in other cases. It was in no way different among
the ancient languages. Lúvratis, in a book of grammar,
means syntax, and in the account of a war the order of
battle. Mélos, which is the name for the members of
the human body, is also a term used in prosody and in
music. The substantive å oplopós, derived from the
verb åpoplów, " to assign limits, to define," mcant on the
onc hand the material delimitation of a territory, and
on the other hand the definition of an object or idca .
1 From monós, " numerous," and onucior, " signification ."
 POLYSEMIA 141

In this latter senise it supplied medicine and philosophy

with the word aphorism : in the former sense, there
still remains Monte Aphorismo, a spur of Pen
telicus. In the time of thc Roman empire, the sub
stantive étionula , followed by a proper noun, meant the
voyage of the sovereign across his states. We find, for
example, in a Syrian inscription : επιδημία θεού 'Αδριανού.
But in medical language the same word, followed by
the name of an illness, signified a contagious disease
prevalent in a certain country , an epidemic. Súpıyć, in
modern Greek, designates, in accordance with the con
text, a flute, a fistula, a syringe, or a tunncl.

It will be asked, how it is that these meanings do not

thwart cach other ; but we must remember that cach
time the words are placed in surroundings which prede- /
termine their import. When we see the doctor at the bed
of a patient, or when we enter a chemist's shop, the word
prescription assumes a colour which in no way suggests
a legal terminology. If we sce the word ascension
printed on the door of a religious building, it does not
remind us of aronauts, of mountain climbs, or of the
height of the stars. We are not even troubled to suppress
the other meanings of the word : these meanings do not
exist for us, they do not cross the threshold of our
consciousness, It is bound to be so, since happily
the association of ideas is for most men based on
essentials of things, and not on the sound.
What wc say of ourselves is no less truc of the listener.
He is in the same situation ; his thought follows, accom
 142 SEMANTICS

panies, or precedes our own. He speaks inwardly at

the same time that we do : he is therefore no more ex
posed than we are to being troubled by collateral signi
fications which are dormant in the depths of his mind.

A new acceptation is equivalent to a new word. This

is proved by the precept - by no means artificial, but
confirmed by general opinion that a word used suc
cessively in two different mcanings must be repeated.
Morcover, it is allowable to makc a word rhyme with
itself, if the two mcanings are sufficicntly removed from
one another.
It would not therefore be accurate to treat words as
signs which disappear at one stroke. A word may have
long fallen into oblivion in its proper meaning, and yet
have survived in an indirect sense. Danger, in its
strict sense, which is " power,” exists no longer, but it
continues to be used as a synonym ofperil.3

Sometimes, after a more or less lengthy sojourn in

some particular department of a language, a word is
inscribed twice over in the general catalogue, but spelt
differently. It is in this way that the French get the
desseins (designs, purposes) of God and the dessins
1 Victor Egger, La Parole Intérieure.--.“ What we call hearing
often comprises a beginning of silent articulation, of feeble, faintly
indicated movements, in the vocal mechanism .” (Ribot.)
2 " Les accommodements nefont rien en ce point ;
Les affronts a l'honneur ne se réparentpoint." - CORNEILLE.
3 " You stand within his danger " (Shakespeare). Etre au
danger ( au pouvoir) de ses ennemis, tirer quelqu'un du danger de
mort, used to be said in French. It is the Low- Latin dominiarium ,
 POLYSEMIA 143

(designs, drawings) of Raphael ; the chambre des Comptes

(audit-room ), and the Contes (tales) of the Queen of
Navarre. These differences exist among all nations and
in all languages ; half -knowledge triumphs over them ,
although at bottom they are not to be wondered at, and
sometimes even are not without a certain value. It is
difficult to draw up a rulc on the subject. Nevertheless
I will propound onc : Respect distinctions which are
ancient and made in good faith ; abstain from crcating
fresh oncs of set purposc.
So truc is it that the bifurcation of meaning can make
two or even several words out of a single word , that the
grammatical changes which modify the onc leave the
other untouched. The Latin verb legere changes itse
into i in the compounds : eligere, colligere. But when it
means “ to read,” it keeps its e : perlegere, relegere. An
author of the seventeenth century ? draws attention to the
fact that the French bon has for its comparative meilleur,
except when , taken in a bad sense , it means “ foolish ,
simple,” as in this example : “ Vous vous étonnez, dites
vous, qu'il ait été assez bon pour croire toutes ces choses : et
moi,je vous trouve encore bien plus bon de vous imaginer
qu'il les ait crues .” Distinctions of this kind exist
everywhere. A German author observes that roth makes
in the comparative röther, except when it has reference
to political colour, in which case it makes rother. Rather
than turn observations of this nature into ridicule, we
1 It is so to a certain extent in the case of words like prope
nouns, such as Smith, Smyth , and Smythe, which , starting from
a single type, are found in the Directory with their special
orthography. * Nicolas Andry .
144 SEMANTICS

should do well to inquire into their cause, which lies in

the fact that grammatical rules are kept up by usage,
and that words belonging in their diverted meaning to a
later cpoch, evade the rule. The French are accustomed
to make of ciel the plural cieux : “ Celui qui régne dans
les cieux, jusqu'au haut des cieux." But they say of a
paintcr that he takes pains with his cicls, not at all for
the pleasure of making a futile distinction, but because
art criticism only crcated its language in the eighteenth
century.

We have not yet exhausted this chapter on Polysemia.

There is an indirect or secondary Polysemia, which it is
as well not to confuse with the other, although the two
are very generally confounded. A few examples will
show wherein they differ.
In Latin, truncus designates the trunk of a tree ; it
also means " mutilated, incomplete." But it would be a
mistake to pass straight from one meaning to another ;
there is an intermediary, which must not be omitted.
From truncus, " trunk of a tree," there came truncare,
" to cut down, to pollard a trec," and this truncare pro
duced the adjective truncus, which has but a remoter
relationship with the older form .
Another example is the Latin examen , which means
at the same time “ swarm ” and “ examination ." In
order to know the reason of this Polysemia, we must
turn to the verb exigere, which means sometimes “ to
lead out" and sometimes " to weigh.” Suetonius reports
that Caesar had a taste for pearls, and that he liked to
 POLYSEMIA 145

weigh them in his hand : sua manu exigere pondus. It is,

therefore, only through the verbs from which they are
derived that the two meanings are united ."
A word may thus be brought, by a long or short series
of intermediaries, to mcan almost the opposite to that
which it first signified.
Maturus meant " matutinal.” Lux matura was the
light of dawn. Ætas matura significd youth. Faba
matura, the carly bean, in opposition to faba serotina.
An early winter, matura hiems. Hence arose the verb
maturare, “ to hasten ,” which Virgil uses with fugam .?
Applied to the produce of nature, maturare took the
meaning of ripening, and, as ripening implies time, the
adjective maturus, under the influence of the verb,
ended by becoming an epithet signifying “ wise, well
thought out.” Maturum consilium , “ a well -matured
plan." Centurionum maturi, " the oldest of the cen
turions ” (Suetonius). This meaning is therefore almost
opposite to that which was accepted for maturus, in
the beginning. A dictionary which bracketed together
the two mcanings might give colour to the opinion
maintained a few ycars ago by a philologist, that Lan
guage began by the identity of contrarics.
· An example in French of this indirect Polysemia is grenadier,
which designates alternately a soldier and a species of tree. To
find the point of junction, we must go back to the " grenade "
(grenade or pomegranate). It is by this sham Polysemiathat the
spirit of words is especially fed .
· Maturate fugam , regique haec dicite vestro (Æn., l. 146).
Maturandum Annibal ratus, ne prævenirent Romani (Livy,
xxiv . 12).
L
 CHAPTER XV

A SPECIAL CAUSE OF POLYSEMIA

Why an expression can be mutilated, without losing anything

of its signification - Abridgment, the cause of irregularities in the
development of meanings - Expressions known as “pregnant."
A VERY frcqucnt causc of Polyscmia, which cvades
foresight and classification, is abridgment. It happens,
for example, that of two words primitively associated,
the one is suppressed. At this sudden removal the
remaining term secms to change its meaning abruptly.
In this casc, it would not be accurate to say that there
was cither Expansion or Restriction. The event which
has come to pass is of a different nature : like an heir
who becomes the sole owner of a property which had
previously been a joint possession, the last survivor
succeeds to an entire idiom and absorbs its signification.
This fact descrves a moment's attention, for nothing
better demonstrates the real nature of Language.
When two words have been habitually used in con
junction, the one may be suppressed with no disadvan
tage whatever to the idiom of which it forms part :
146
 A SPECIAL CAUSE OF POLYSEMIA 147

sometimes even the expression gains thereby in energy.

This is owing to the fact that when once the meaning
of the two words has combined, they thenceforward
constitute only a single sign. Now a sign can be cut,
clipped, reduced by half ; provided it is recognisable, it
still fulfils its original function. We can imagine the
strange accumulations of mcanings which must arise,
since there is nothing to prevent the suppression of the
most essential portion . It is of no use to inake cate
gories, according as the first or the second word has
been removed, according as the adjective or the substan
tive is the survivor. The only rule which holds good is
this : thc portion which survives takes the place of the
wholc ; thc sign, though mutilated, remains adequate to
the objcct.
Examples of this practice are innumerable : the
articles in our dictionaries would not display their
present length, if the verbs had not absorbed into them
selves the meaning of a complement which could thence
forward be omitted, if the adjectives had not been
enriched by the value of an implied substantive, or if
whole phrases had not been compressed into a single
word .
Many apparent incongruities vanish in the light of
this simple truth . As modern languages are generally
more charged with mcaning than ancient ( for the very
simple reason that the experience of the human race is
longer), we will begin by borrowing a few cxamples
from them . It is true that when these facts suggest
themselves to us in the present, they scem hardly worthy
148 SEMANTICS

of our attention. Yet what we find in the past, though

more difficult to recognise, is of the same nature.

Everybody knows that the House means “ the House

of Commons " (French, la Chambre = la Chambre des
Députés) ; that when we speak of members of the
Cabinet, it is the Cabinet of Ministers which is implied.
We should at once be puzzled by this word minister,
were we not aware that at Rome, in the time of the
Empire, minister significd " scrvant of the prince."
Prince again takes us back to a yet earlier abridg
ment, princeps senatus, "first of the senate. " It is in
this way that from age to age words assume the
meaning of lost companions. Without this kind of
intussusception Language would soon take unduc
developments.
It has been supposed that absolute power especially
favoured the multiplication of these phenomena, since
the idea of a sovereign placed all that concerned or
approached him on a uniquc footing. Thus at Versailles
le lever was the reception of the king ; avoir la plume
mcant to imitate the writing of the king and keep up the
correspondence in his stcad and place. But there is in
this nothing but a fact which recurs at all times and at
every stage of socicty. At a certain epoch of the French
Revolution , suspected citizens were décrétés : it seemed
unnecessary to add d'accusation . In the legal language
of France, instruire means to inquire into a contention ,
a law-suit. In the language of tuition, instruire les
enfants means to supply children with necessary know
 A SPECIAL CAUSE OF POLYSEMIA 149

ledge. In a regiment, donner le mot means donner le

mot d'ordre, "to give the word of command. " In Rome,
eris confessus was a man who admitted a debt ; the full
idiom would have been æris alieni.
In every situation, and in every profession , there is a
certain idea so present in the memory that it seems
unnecessary to state it in speech. The epithet which
serves to specify the idea is alone expressed. Hence
the quantity of adjectives which, in time, take their
place among the substantives. The geomctrician speaks
of the perpendicular, thc oblique, the diagonal. The
French writing-master speaks of la ronde (round hand),
of l'anglaise (a small hand), of la båtarde (a sloping
hand). These abridgments are so well known that it is
needless to dwell on them. But we must note how
faithfully, in French, the gender of the implied substan
tive is preserved : à la Française, d l'étourdie, de plus
belle, à droite is still said, although the substantive, which
is mode,façon, manière, main, long since ceased to be
expressed.
The Roman family, familia, was composed of children
and of slaves ; hence arose the two adjectives, liberi
and famuli. Both have been substantives from time
immcmorial. .
In Greece, the brother born of the same parents was
κασίγνητος .. The brother on the father's side only,
" Most of the problems relating to gender must be solved in this
way. Oriens, occidens are masculine because of the implied sol.
Prosa is feminine on account of oratio. Ovile is neuter because of
stabulum . We are here speaking, be it well understood, of sub
stantives of secondary formation only.
150 SEMANTICS

ομόπατρος or oπατρος. The brother on the mother's side

only, adeados. With all these words, Opátwp had to be
understood, which, having become useless, had dropped
out of ordinary language, though it survived in the
language of politics.
There is no doubt that if we could go farther back
than the Indo -European period many of the substantives
of that period would be revealed to us as adjectives.

It is obvious that a large field is opened to Polysemia

by these suppressions. The adjective novellus (English
novel) is one of the diminutives so common in the
familiar language of Rome. Novella, then, was said
speaking of young vines, leaving vites to be understood.
But Roman legists, speaking of constitutions given to
the empire after the codification of Justinian, also used
the word Novella, with leges understood. These ex
amples are so frcqucnt that it is unnecessary to multiply
them : we know how sorely double meanings have been
misused by the punster.
Words designating an article of daily use like leaf,
card, board, table, owe their Polysemia to the suppression
of the determinative. It would be wrong to attribute
this variety of meaning to the word itself : the meaning
made its entry subsequently, by abridgment of the idiom.
Etymology might in such a case become the most mis
leading of guides, were not knowledge of things added
to knowledge of words.

Old - fashioned philologists, who had noted a certain

 A SPECIAL CAUSE OF POLYSEMIA 191

number of facts of this kind , invented an original

denomination in order to describe them . When the
verb absorbed into itself the signification of its comple
ment, they said it was pregnant. The expression is
pretty though inaccurate, for to place gestation after a
divided existence is to defy the usual order of things,
and to do violence at once to chronology and to natural
history. Be that as it may, this absorption is extremely
frequent, especially in the language of divers professions
and estates. The meaning of the complement enters
then in some degree into the verb, and gives it a wholly
characteristic signification. We know, in the language
of devotion, the meaning of a practising Christian , or in
French of un malade qui est administre. What is more
common than the verb déposer ? But when we speak
of a witness who deposes, everybody understands that
there is question of information given to justice. To
heave can be said of any object that is raised or pro
jected : but, in nautical language, the order to heave to
has a special signification .
In the presence of a hearer of some information, it is
natural to suppress what is self -evident. In the sixteenth
century , the expression “ a woman possessed ” gave rise to
no uncertainty : it meant a woman possessed by the
devil. When in the reports of the law courts the
French newspapers announce une affaire de inæurs, the
reader understands that it refers to an offence against
morals.

Sometimes suppression changes the meaning of the

152 SEMANTICS

surviving word to its own advantage. We have a

characteristic example of this in the word points."
It is commonly believed that, in the eyes of the
Grecks, the poet was "the creator " and the poem " a
creation.” This sounds very fine and places the poet
very high. But the truth is rather different. After a
first epoch, that of the indol, when poets were their own
interpreters, came another period in which a distinction
was made between the author of the verses and the
singer or actor who merely reproduced them in public.
Then μελών ποιητής, or επών ποιητής was said in opposition
to ραψωδός or-υποκριτής. So by abbreviation, ποιητής, when
odcs or dramas were in qucstion, significd the author of
the lincs, exactly as when, at the end of a play, the
public now calls for the “ author.” But this twofold
function faded gradually from the memory. The poet,
still keeping the same name, though he had no more
need for a spokesman, scemed to owe his title to some
higher conception : and it is surrounded with this halo
of nobility that the name appears to us to -day.
We owe the Latin expression defunctus, for designat
ing the dead, to an idiom whose simplicity was not
devoid of beauty. It must be completed by vita, that is
to say, " one who has accomplished life," a difficult and
serious function . Defunctorum memoria, is the memory
of those who, having served their time in the army of
the living, have received their discharge. By a similar
sentiment migrare, in Grégoire de Tours, means “ to
" See on this word an article by M. Weil in the Annuaire de
l'Association pour l'encouragement des études grecques, 1884.
 A SPECIAL CAUSE OF POLYSEMIA 153

dic ." Ad dominum or a sæculo is understood . We will,

here transcribe the reflections of M. Max Bonnet. “ All
fixed expressions have this in common : that the words,
by dint of being placed together, react to some degree
on each other, and each acquire part of the signification
of the other. . . . It may happen also that one of the
two, by itself alone, arouses in the mind of the reader
the idea usually expressed by both."
A few examples of expressions in which abridgment
has brought about a remarkable change of signification
in much uscd words will end this chapter.
When a Frenchman says : entendre un orateur, entendre
un discours, he uses entendre in the sense of “ to hcar."
But in reality it means " to apply." Intendere stands
for animum intendere. This change of meaning is
moreover an ancient one. We find already in Grégoire
de Tours : Quos sæpe conspicit et intendit .
Defendere meant originally " to remove " ; defendere
ignem a tectis, defendere hostes ab urbe. It is by an
abbreviation that defendere urbem , defendere domos were
said. Mactare meant “ to enrich, to amplify” ; by an
abbreviation mactare deos bove became mactare bovem ,
“ to sacrifice an ox . " Adolere meant " to increase , to
· Le latin de Grégoire de Tours, p. 255.
* The regular construction required the dative. The French
still say : “Il ne veut entendre à rien . Je ne sais auquel en
tendre . "
3 The idiom which is condemned by grammars : fixer un but,
fixer une personne is of absolutely the same nature. But it made
the mistake of coming at an epoch when language no longer lends
itself to the same extent to these abridgments.
154 SEMANTICS

enrich " ; by an abbreviation adolere aram ture became

adolere tus, " to burn incense.”
Thus Language, wherever it is closely examined, shows
an idea which remains intact while its expression is
restricted and abridged. In spite of the somersaults
to which these ellipses subject the history of words, we
must recognise in them the normal and legitimate work
of the mind .
 CHAPTER XVI
COMPOUND NOUNS

Importance of meaning – The order of the terms— Why Latin

forms less compounds than Grcek - Limits of Composition in
Greek - Sanscrit compounds - Compounds have never more than
two terms.

The Composition of nouns forms an attractive chapter

of Indo- European philology, for it shows us more clearly
than any other the part played by the genius of the
nations ; it even reveals to us individual action , so that
grammar already to some extent borders upon literary
criticism. Moreover, since the Indian theory has cleared
the ground and provisionally marked the divisions, this
question has become the object of numberless investi
gations."
It is the semantic side which has so far been forgotten
" A bibliographical list will be found in the Studien of Curtius,
v. p. 4, et vii. p . 1 , also an enumeration of more recent works in
Brugmann, Grundriss, ii., p. 21. We will also mention two French
books, both of them important: Meunier, Les Compose's syntutiques
en grec, en latin, en français (Durand, 1872) ; Ars. Darmesteter,
Traité de la formation des noms composés (2nd edition, 1894).
155
156 SEMANTICS

in these studies. You might think, from reading the

works cited, that questions of accentuation, of the con
necting vowel, of the order of terms, were all -important.
I fear that the essential, which is to know the meaning,
has been forgotten ; for it is the meaning, and nothing
else, which makes the compound, and which finally
determines its form .
It is a primordial condition that, in spite of the
presence of two terms, the compound should make the
impression of a simplc idca on the mind. 'Akprodus
designatcs, not a city inore or less lofty, but the fortress,
the citadel ; dodóuntis is synonymous with the French
adjective ruse ; tohútpotos corresponds exactly to the
Latin versutus.
Such is the necessary and at the same time the
sufficing condition. Thus brother-in-law, step - sister,
grand -father, although having nothing to distinguish
them outwardly, arc compounds, bccausc thc mind,
instead of dwelling successively on both terms, perccives
nothing but the whole .
An attempt has been made to distinguish these com
pounds from compounds such as åxpóriodis, by describing
them as juxtaposed. But the line of demarcation is
visible to the grammarian alone. Words like aquæduc
tus, terræmotus, legislator, jurisconsultus, fideicommissum
were also considered as juxtaposed, because the first
term bears the mark of an inflection ; but for the Latin
language thcy were compounds ; it is this fact indeed
which explains the phonetical and grammatical peculi
arities which appear in some of them, such as crucifixus,
 COMPOUND NOUNS 157

manifestus, triumvir. Crucifixus has shortened its first i

Manifestus has disfigured the ablative mann . Triumvir
has acquired and fixed a genitive plural, which owed its
existence to idioms like lis trium virum . As soon as
the mind unites in one single idea two notions previously
distinct, all kinds of reductions or petrifications of the
first term become possible. But these are accessory
facts, the presence or absence of which makes no change
to the essence of things. True Composition has its
criterion in the mind.?

Philologists have laboriously discussed the order of

terms, which is not the same in all languages. This is
attaching considerable weight to a question of secondary
importance. The order of terms within the compounds,
is generally determined by the habitual order of words
in the phrase. Legislator, which is in juxtaposition, is
constructed according to the usual method of the Latin
language. Signifer, which is a compound, is constructed
in thc same way as the two component words would
be constructed in the regular course of speech. The
advantage of this order is that it allows the principal
portion, which comes last, to assume the inflection of
either the nominative, the accusative, or any other case
according to the general construction.of the phrase.
1 Festus, past participle of fendo, "to knock . " Res manifesta is
a thing that can be touched with the finger.
9 These considerations should be decisive in a discussion con
cerning the orthography of nouns such as the French arc-en - ciel,
chef- d'æuvre, cul-de-sac, etc. There is no doubt that the decision
ought always to be in favour of unification .
158 SEMANTICS

But we know that Greek often abandons this order :

the attempts which have been made to explain com
pounds such as pilótevos according to the Sanscrit
type have been far from convincing. It has not been
sufficiently remembered that we are here entering upon
a domain in which the special originality of cach nation
has a frcer play. It is impossible for the individual to
crcatc at will a new inflcction whcther of noun or of
verb, bccausc thc clements out of which grammatical
inflcctions were formed have long since dropped out of
circulation : but it is not forbidden to individual initiative
to attempt after its own fashion the joining together of
compounds, cach part of which conveys a meaning, and
forms a word by itself. The Greek custom of choos
ing for proper names compounds such as Θεόδωρος,,
Νικόστρατος , Δεώκριτος, and then of reversing the order,
so as to forum Δωρόθεος, Στρατονίκη , Κριτόλαος, may
have contributed to the habit of freely handling these
words. We hcrc scc a self-conscious liberty making its
appearance in the language.

The question has been raised why Latin forms fewer

compounds than Greek , and the reason given is a
lack of “plastic power," a metaphor which begs the
question, and is entirely devoid of meaning. Certainly
poets were not without a desire to imitate the com
pounds of the Greek language. Attempts of this kind
are not lacking. Why then have these compounds a
borrowed look ? Why were the Latins themselves the
first to laugh at thcm ? It was no doubt because the
 COMPOUND NOUNS 159

popular mind has to be prepared for the creations of

poets by the language of every day. On the other
hand, ancient compounds such as princeps, pauper,
simplex, were already too much restricted and contracted
by pronunciation, and had already lost too much of their
transpicuity, to serve as initiation or as guide.'
It is with reference to compound nouns that Lucretius,
having to find an equivalent for the Greek ομοιομέρεια,,
makes his complaint on the poverty of thc Latin
language, patrii sermonis egestas. Quintilian has an
analogous observation : Res tota magis Græcos decet, nobis
minus succedit. Yet it must not be thought that Latin
possesses no compounds : were they to be all collected
together, the list would be long. The language of the
calendar alonc offers a certain choicc, like armilustrium ,
regifugium , fordicidia, ctc. Law shows an cqual number :
judex, manceps, justitium , ctc. What the Latin language
docs lack are such fine cpithcts of pure ornament as
αργυρότοξος, βωτιάνειρα, κερδαλεόφρων, which are so abun
dant in Greek poetry. One feels that the model of epic
poetry was wanting.

Though multiplying compounds of this. kind, Greek

seems at the same time to have fixed a limit for itself.
It created them to designate a permanent quality, a
constant action, but not to indicate a transient fact,
or an accidental attributc. Achilles, for example, was
" If English had no compounds save those like world (for
wer-old, “ age of man "), or lord (for hlāf-ward , one who dis.
tributes bread"), the English language would have preserved the
use of compounds to no greater extent than French.
160 SEMANTICS

called ωκύπους ; but he was not called βλητόπους or

Tpwrónous to show that he had just been wounded in
the foot. Briareus of the hundred arms was called
&Katóyxelp ; but Greek would not have allowed a com
pound étatóxelp, " with arms spread out,” or 4c0bxeup,
“ with a stone in his hand .” 1 It relegated to the phrase
and to the verb the care of marking these transitory
conditions. We know that this was not the case in
Sanscrit : there it constantly happened that a compound,
all packed with momentary circumstances, absorbed into
itself the whole movement of the phrase, which was
thereafter left with nothing more to say. But Compo
sition is for Sanscrit a sort of alternative course which
permits it to cvade syntax almost entirely.
It was in this way that from krõdhas, " anger," and
ģita, " vanquished ,” a compound gita -krõdhas was formed,
" one who has vanquished or mastered his anger." Out
of prăpta, " obtained ,” and ģīvika, “ provision ,” arose
präpta -givika, "one who has the necessaries of life."
From kāma, " desire," and tjaktum , infinitive of the verb
tjag, " to leave," came tjaktu -kāma, “having the desire
to go away."
Words like the above are by no means uncommon
in Sanscrit. This language introduces into an epithet
circumstances which are foreign to the subject, as for
instance the hour of the day or the number of people
present. From mätri, “ mother," and šaštha, “ sixth , "
In Sanscrit, grāva-hasta , from gravan, “ stone, ” and hasta,
“ hand ,” is an epithet applied to the priest who crushed the soma.
Cf. F. Justi, Zusammensetzung der Nomina.
 COMPOUND NOUNS 161

Sanscrit makes måtrt-šaštha, an epithet given to the

five brothers Pandavas accompanied by their mother.
This is translated by “ having their mother for sixth
(companion).” It is from asthi, “bone,” and bhūjas,
comparative of bhiri, “ much ," that Sanscrit makes
asthi-bhijas, which signifies " composed chiefly of bones,
being nothing but skin and bones." From daca, " ten ,"
and avara , “ inferior,” it makes daça -avara, an epithet for
the assemblage of ten persons at least. This is a real
abuse, which has extended the faculty of Composition
beyond due limits, and which has had the counter -effect
of starving all other means of expression.
might be supposed, no doubt, that Hindoo gram
marians, faithful to their systematic vicws, have some
times interpreted as compounds, and treated as such,
small phrases where words are placed end to end,
according to a fairly loose construction, in which neither
rules of agreement nor rules of subordination are found.
This is a suspicion which we cannot avoid when we see
the extraordinary explanations to which commentators
have recourse . We find, for example, that in a narrative
nihçvāsa - paramă (sighing a great deal) is translated by
regarding sighs as the supreme thing," while cintā -parā
(very pensive) is rendered " holding meditation for the
chief good.” One wonders if these are not artificial
interpretations, and if there is not hidden behind the
so - called compounds a less strictly regulated condition
of Language . An examination of the modern languages
1 To go back to the examples quoted above, the following inter:
pretation would be quite comprehensible : " the five brothers,
M
162 SEMANTICS

of India, the usages of which come through Sanscrit,

will help to solve these questions.
I have made this digression for the purpose of showing
that the different parts of a language are mutually
dependent, and that by developing one of them unduly,
we tend to diminish the power of some other. I should
add that modern German, which makes great use of
Composition, is to a certain extent exposed to a like
danger, not indecd in the works of Goethe and Schiller,
or of other writers of their rank, but in the ordinary
language of daily use, specimens of which may be found
in the last page of any newspaper."
I said above that the genius of the nations begins to
appear in this department of grammar.
To the Greck language belong those peculiar and
embarrassing compounds, the first member of which
ends in σι : φιλησίμολπος , " a Iover of songs, ” τερψίχορος,
" onc who cnjoys dancing," Avoltovos, " onc who rests
from fatigue," pouolußporos, “ destroyer of men ," álcoloikos,
" onc who destroys the house," 'Apkeolaos, " a defender
of the peoples," åreElkakos, " one who represses evil,”
owolmodes, “ one who saves the city ,” etc. There has

been no lack of explanations to account for their first

Pandavas, the mother for sixth . ” And similarly with the others.
They say in French : “Il vient, les cheveux hérissés, le visage en
feu," without its being possible to explain these fragments of a
phrase from the point of view of French syntax.
I Präsidentschaftswahlkampf. -- Postdampfersubventionsvorlage.
- Vierwaldstätterseeschraubendampfschiffgesellschaft.— Das ein
jährigfreiwillige Berechtigungswesen. – Heute verschied Frau
Chef-redacteurs-wittwe der Allgemeinen Zeitung,
 COMPOUND NOUNS 163

term ; but this is not the place to discuss them . We

think that the origin was a rather emphatic phrase, such
as the popular imagination is quite capable of inventing,
“ the Salvation of the City," for instance, or “the Bul
wark of the Peoplc." It is in any case the fact that no
similar compound is found elsewhere. The Latin poets
indced attempted something of the same kind. Versi
color recalls dueryixpoos, fluxipedus is an imitation of
dkcolteados. But these forms were never acclimatiscd in
Latin. The Greeks, on the contrary, even at the present
day form compounds of this kind : alleElképauvos signifies
" lightning conductor, ” and årefißpóxlov “ umbrella.”
There is a certain pleasure in collecting Greek creations
of this kind : daxéovuos, “ that bites into the heart,"
édémodis, " taker of towns,” xalpékakos, “ who rejoices in
eyil,” ¿De opítwp, " who pretends to be an orator,"
dococopos, “ who thinks himself wise,” palvounpls, “ who
shows her thighs" (speaking of women of Sparta ),
åußodoypa ,“ who puts off old age " (surname of Aphrodite
among the Spartans).
I will mention yet one more form which has been
especially developed in Germanic languages.
German contains a certain number of compounds
such as himmel-blau, " sky -blue," schnee-weiss, " snow
white," stock -fest, “firm as a stump,” in which the first
term serves as specimen of the quality denoted by
the second. Modern language has worked freely upon
this model ; and compounds of this sort are very numerous.
We will only quote : thurm -hoch, "high as a tower,”
blei-schwer, “heavy as lead," eis-kalt, "ice-cold," felsen
164 SEMANTICS

fest, “ firm as a rock ," leichen -bleich, " pale as a corpse,"

etc. Some of these terms of comparison have passed
out of the words which gave rise to them into others
with which thcy have no concern, and in which, whether
intentionally or not, they producc an odd cffcct. It is
in this way that, in imitation of stock -fest, "firm as a
stump," the Germans said, stock -taub, " deaf as a post,
stone-deaf,” stock -blind, “ completely blind ,” stock- finster,
“ pitch - dark ." From having said stein -hart, “ hard as a
stonc," they said stein -alt, “ old as the stones," stein -müd,
" dead -tired," stein -reich , "very rich." 1
The languages which prefer Derivation to Composi
tion are formed of a less tractable material ; they lend
themselves less easily to the creation of new words, for
which they must not only choose a suffix, but prepare the
first part of the word . Thus French, to obtain deriva
tives of frère, makes use of Latin (fraternel, fraternité ).
It is clcar that languages which habitually employ com
pounds, and in which the very suffixes were originally
independent words, have no need to fight against
difficulties of this kind. I will mention one example
only. The traveller Bleek , speaking of the clicks of the
tonguc in common use ainong the Hottentots, employs,
to designate certain exceptional dialects which are
devoid of them, the compound clickless. Neither French ,
nor any of the Romance languages could compete with
| Instead of saying : Es schreit zum Himmel, “ it cries to
heaven ," German, by an ellipse, whose boldness is cloaked by
custom, can say : Es ist himmelschreiend. Therc has no doubt
been an amalgamation with compounds like himmelklar, himmel
weit, “ clear as the day,” “ far as the sky.”
 COMPOUND NOUNS 165

English in this particular. But it is not by chance that

the idea of “ purity,” whence came the Della Cruscan
and the French Academy, should have been hatched
among the nations which make use of derivatives.
But we must not imagine that a nation is ever pre
vented from forming the new words of which it is in
need . If French harks back to Latin, it is because
it has in a manner grown up under its eye and is
attracted to it by a time-honoured habit, grown stronger
from century to century. Had this great reservoir
failed, the genius of the people would have looked else
where. The homogeneity of certain languages such as
Lithuanian, arises from their self-dependence. Custom
and greater conveniencc -- that is what we find : we must
speak neither of compulsion nor of an incvitable law.
In closing the chapter, I will recapitulate the principle
that governs the matter.
Whatever the length of a compound, it never com
✓ priscs more than two terms. This rule is not an
arbitrary onc : it arises from the nature of the human
mind, which associates ideas in couples. It may chance
that each of these two terms is in itself a compound.
Thus in the Aristophanic word στρεψοδικοπανουργία , the
second term πανουργία, is a derivative of πανούργος, which
is formed from tây and <pyov, and on the other hand
otpeyodikos itself contains two words. But it is clear .
that each of the two parts counts for one element only.
The important question in such a case , is to make the
division in the right place : that is the chief difficulty of
languages which are too lavish of Composition.
 CHAPTER XVII

ARTICULATED GROUPS

Examples of Articulated Groups - Their use.

LIKE the various parts of a mechanism , which we are
so accustomed to see fitting into one another that we
do not think of imagining them apart, Language con
tains words which have been so long united by usage,
that, for our intelligence, they no longer exist separately.
These I call Articulated Groups. Their importance in
Syntax is very great. It will be sufficient to give as
examples the French expressions parce que, pourvu que,
quoique, attendu que, afin que, etc. There is no language
without a certain number of these. They were thus
adjusted by the hand of our ancestors, and provided
for posterior ages as prop or lever. These Articulated
Groups are for the ordinary speech of daily life what
formularies are in law or in administration.
The most of men make use of them without having
cver given them a thought. They have become so
thoroughly incrusted in our mind that they determine
166
 ARTICULATED GROUPS 167

its movements. We only properly realise them when we

compare our own with a foreign language. Whenever
two different populations come into contact, the mistakes
and errors committed on both sides reveal their
presence .
If the lettered classes were to disappear, the Articulated
Groups would form solid blocks ; and it is the blocks,not
the component parts, which would survive to furnish the
elements of a future language. Everybody knows that
the word, in its isolated state, has no very clear existence
in the popular consciousness, and that it is apt to join
on to that which precedes or follows it. Telegraph
offices, where words are counted one by one, must have
gathered on this subject an ample harvest of observa
tions. For purposes of interrogation the French make
use of the group est - ce que ; for expressing doubt, of the
group peut- être que ; for explaining the motives of an
action, of the group c'est que ; all of which seem to us now
to be expressions of single growth . In modern Greek
the future is indicated by means of the particle Oa
followed by subjunctives : a ném , " he will say: " This
particle da is nothing but the ainalgam of the group
Oémet iva, " he wishes that." 2 These facts should make
us cautious on the subject of ancient particles, so short,
1 M. Hugo Schuchardt has studied the language spoken by the
Slavs and by the Germans of Austria, from this point of view.
He has attempted to catalogue and enumerate the mistakes caused
on both sides by an ill-timed recollection of the mother tongue.
They are practically the same mistakes which are made at school,
and upon which our professors adjudicate.
? Bend is still found for ea, in the Epirot dialect.
168 SEMANTICS

but so often full of meaning, which Pott compared to

those light substances one pinch of which is sufficient to
change the taste and flavour of a dish."
Not only do these Articulated Groups preserve the
entire signification of the elements of which they are
composed ; they also acquire a certain value which
does not properly belong to them, but which results from
the position which they habitually occupy in the phrase.
I take as example the French word cependant, in which we
now suppose ourselves to detect an opposition. Nothing
in this word denotes opposition. But as it often happens
that two concomitant facts are enumerated in order that
they should be opposed, the adversative idea has little
by little crept in. In the same way we imagine that we
feel a sense of opposition in the Latin conjunctions,
quamvis, quanquam, etsi, etiamsi, licet, etc. All these
words are merely affirmatives; some of them even
exaggerate the affirmative element, allowing it to be
emphasised to any extent in order to bring out all the
more the fact which is held in reserve, and which is
coming to limit or contradict the first proposition . The
hearer, prepared by custom , so clearly foresees this
second assertion, that no sooner does the first appear,
than he is conscious of the antithesis.

See, for example, the ingenious analysis of the Latin particle

an, by James Darmesteter, in the Mémoires de la Société de
Linguistique, vol. v .
Quamvis sis molestus, nunquam te esse confitebor malum "
(Cicero, Tusc. II. 25,61 . It is pain which is in question). “ Be thou
as importunate as thou choosest, I shall never admit that thou art
an evil.”
 ARTICULATED GROUPS 169

These expressions, having passed into the condition of

an indissoluble group, preserve grammatical forms which
no longer exist in the ordinary language. Thus the
Latin duntaxat, contains the aorist of the subjunctive
of the verb tango analogous to aúon, 1 (En. An ancient
neuter substantive regum , signifying " direction ,” is con
tained in the adverb , crgo, for e rego, " in a straight linc,
in consequence.” In the German nur, we have a small
proposition : ne wære, "werc it not.” The modern
Greek ås, which expresses an invitation (ås dalúowuer, às
eloénowot), represents the ancient imperative άφες,,
“ permit .”
The more closely we examine Language, the more new
semantic stratifications we find. This long labour was
necessary in order that a somewhat condensed train of
reasoning should be communicated to others without
digression or obscurity. All men can now profit by this
labour: it is so easy to handle these Articulated Groups,
that one is tempted to believe that they have always
existed . The child learns to manage them, as he learns
to use the inheritance of his fathers. Yet we have but
to look at the more backward nations to see not only
that they have more difficulty in making themselves
intelligible, but also that, not finding any support for
their thought, they have to make greater efforts in order
to preserve it in the mind, and to remain its masters.
These Articulated Groups, which have been the instru
ments of the syntax, and by the help of which the period
is unrolled, may be transferred by imitation from one
i Cf. e regione.
170 SEMANTICS

language to another. We are even tempted to believe

that the form of the period was invented once and for
all : when we read a Latin Senatus Consultum, or one
of those Epistole addressed to the provinces by the
Roman Emperors, we recognise in them the same
arrangement as in the edicts of our parliament and the
decrees of our kings. The most immaterial part of our
languages is never lost. Phonetics and morphology are
right to distinguish that which comes through conscious
imitation from that which comes through popular
tradition : between these two elements there can be no
fusion . But in Semantics this distinction is useless.
Even if interrupted at certain moments, the chain of
progress can always be linked together again.
 CHAPTER XVIII

HOW NAMES ARE GIVEN TO THINGS

Names given to things are necessarily incomplete and inac

curate - Opinions of the philosophers of Greece and of India
Advantages of phonetic change - Proper names.
We have reserved for the end of this second part the
question which is generally asked at the outset of any
study on Language : How did men go to work to give
names to things ? What we have seen in the preceding
chapters provides an answer.
Onc conclusion is to be drawn from all that has gone
before : it is an undoubted fact that Language designates
things in an incomplete and inaccurate manner. In
complete : since we have not exhausted all that can be
said of the sun when we have declared it to be shining,
or of the horse when we say that it trots. Inaccurate :
since we cannot say of the sun that it shines when it has
set, or of the horse that it trots when it is at rest, or when
wounded or dead.
Substantives are signs attached to things : they con
tain exactly that amount of truth which can be contained
171
172 SEMANTICS

by a name, an amount which is of necessity small in

proportion to the reality of the object. That which is
most adequate to its object is the abstract noun, since it
represents a simple operation of the mind : when I use
the two words compressibility, immortality, all that is to
be found in the idea is to be found also in the word .
But if I take a real entity, an object existing in nature,
it will be impossible for language to introduce into the
word all the ideas which this entity or object awakens in
the mind. Language is therefore compelled to choose.
Out of all the ideas it can choose one only ; it thus
creates a name which is not long in becoming a mere
sign.
For this name to be accepted it must no doubt origin
ally possess some true and striking characteristic on one
side or another ; it must satisfy the minds of those to
whom it is first submitted. But this condition is im
perative only at the outset. Once accepted , it rids itself
rapidly of its etymological signification ; otherwise this
signification might become an embarrassment. Many
objects are inaccurately named, whether through the
ignorance of the original authors, or by some intervening
change which disturbs the harmony between the sign
and the thing significd. Nevertheless words answer the
samc purposc as though they were of faultless accuracy.
No one drcams of revising them. They are accepted by
a tacit consent of which we are not even conscious.
The reader will here recognise the subject of many
discussions in Grecce and India . The debate begins for
us in the Cratylus of Plato. Socrates alternately supports
 HOW NAMES ARE GIVEN TO THINGS 173

either of two opinions : first, that there exists for each

thing a name by nature belonging to it ; next that the
propriety of a name is wholly dependent on the consent
of men . This discussion lasted as long as there were
schools of grammar in Greece and Rome. It is not
so well known that the same question occupied the
schools of the Brahmans. “ If the grass is called trina,
from its quality of pricking ( trt), why docs not this name
apply to everything that pricks, as for instance a needle
or a lance ? And on the other hand, if a column is called
sthūnā because it stands upright ( sthā ), why is it not also
called that which supports, or that which fits in ? ” 1
Whether from a more or less rational belief in the
necessary truth of Language, or from respect for ancestral
wisdom , it has been the unfailing habit, at every epoch
and among all nations, to refer to words for information
concerning the nature of things. Reference was not
always made to the mother tongue, which was both too
well known and in too close proximity, but rather to
some more ancient language. This conviction of the
ορθότης ονομάτων is universal.. Yet a little reflection
might have shown that it is scarcely reasonable to expect
lessons in physics or in metaphysics from Language, a
work of improvisation, in which the most ignorant man
hab often the largest share, and on which accidental Х
events have set their mark. Yet this conviction has
been a freak of every epoch. I will say nothing of the
ancients, nor of the learned men of the Middle Ages, but
we see even the head of the sensualist school of the
1
1 Jáska, Nirukta, at the beginning .
174 SEMANTICS

cighteenth century, Condillac, a prey to the same

illusion. He has been discussing the qualitics or appear
anccs of bodics. “ As soon," he says, as qualities dis
tinguish bodies and form their manner of being, there is
in the bodies something which these qualities modify,
which is their support or subject, which we represent to
ourselves as being below , and which for this reason we
call substance, from substare, to be bclow ." The ancestor
of thc idcologists hcrc reasons likc a pupil of purc
scholasticism .
How should Language teach us about substance and
quality ? It can but give us the echo of our own
thought: it registers faithfully our prejudices and our
mistakes. At times it may astonish us, like a child, by
thc frankness of its answers or thc ingenuousness of its
representations ; or it may furnish us with valuable
picccs of historical information, of which it is the in
voluntary depositary ;' but to take it for teacher and for
master would be to misjudge its character.
Do words created by scholars and men of science
possess a greater accuracy ? We must not count upon it.
In the scventeenth century, Van Helmont, with a vague
recollection of the Low Country gest, " spirit ," calls bodies
which are neither solid nor liquid by the name of gas.
That is as vague and incomplete as spiritus in Latin.or
7
· If all the monuments of pottery and sculpture had perished, the
words effigies, figura, fingere, would still show us that the Romans
were not strangers to the plastic arts. The single substantive
invidia would tell us that the superstition of the jettatura existed
in Rome. Such is the nature of the information furnished by
Language.
 HOW NAMES ARE GIVEN TO THINGS 175

yuxh in Grcek . A French chemist, having discovered a

new metal, and being moved by a fccling of patriotism,
called it gallium ; a German scientist, no less patriotic,
retorted with germanium . Now these designations tell
us as little about the essence of things as do the names
of Mercury or Jupiter given to plancts, or those of
ampère and volt recently given to quantitics in
clectricity.
Everybody knows that learncd names have been given
by mistakc. Christopher Columbus called the inhabit
ants of the New World Indians. A French department
owes to a mistaken reading the name of Calvados."
We can thereforc sum up after this fashion :
The more thc word is detached from its origin, the
morc is it at the scrvice of thought : in accordance with
the experiments which we make, it is restricted or cx
panded, specificd or gcncraliscd. It accompanies the
objcct which it serves as label through all the events of
history, rising in dignity or descending in the scale of
opinion. At times even it passes over to the very
opposite of its initial acccptation ; and is better adapted
to its many parts thc more completely it becomes a
mere sign. Phonetic change, so far from doing harm, is
favourable to it, since it conceals its former relations with
other words which have remained nearer to the initial
meaning or have gone off in different directions. But
1 It is well known that Calvados is for Salvador. The mistake
arose through a map of the diocese of Bayeux of 1650, which bears
these words : Rocher du Salvador. Without a misreading of the
name the rock would never have had such good fortune.
li
 1

HOW NAMES ARE GIVEN TO THINGS 177

special care with which they are preserved, and they

have this characteristic in common with certain religious
or administrative words.
The difference then between proper and common
names is wholly intellectual. If names were classified
according to the variety of ideas which they evoke,
proper names should head the list : being the most indi
vidual, they are the most significative of all. An ad
jective like augustus, in becoming the name of Octavius,
acquires a wealth of ideas once foreign to it. On the
other hand, we have only to compare the word Cæsar,
taken as designating the adversary of Pompey, and the
German word Kaiser, which signifies “ emperor," to see
how greatly a name loses in significance in becoming a
common name. Whence we may conclude that, from a
semantic point of view, proper names are the most
important of all.

N
៖
 .
PART III

HOW SYNTAX IS FORMED

 CHAPTER XIX

THE PARTS OF SPEECH

What we are to understand by the parts of speech - How they

exist in the mind - Are they innate or acquired ? -Do they all date
from the same period ?
HAVE the parts of speech, such as substantive, adjec
tive, pronoun, adverb, always existed, or are they a
gradual acquisition ? This question is not identical
with thc problem of the origin of Language, since there
are certain languages which, even at the present day,
do not distinguish between the parts of speech, and it is
quite possible that our own idioms may have passed
through a like stage. The question therefore concerns
facts of comparatively recent date, for which observation
should not be declared d priori impossible.
Not only is it not impossible, but the means of in
formation supplied by the history of the Indo -European
languages go back far enough to allow us to see several
of these parts of speech formed under our eyes. Let
us begin with the most modern.
One of the most recent is the adverb . Words such
181
182 SEMANTICS

as οίκοι , πέδοι , χαμαι, εύ , κακώς, ούτως , humi, domi , recte,

valde, primum , rursum , hic, illic, are substantives, adjec
tives, or pronouns regularly declined. But when a word
has ceased to be in an immediate and necessary relation
to the rest of the phrase, when it serves to determine
more fully some other term, without, however, being
indispensable, it is ready to take the value of an adverb.
If it ccases in the smallest degrec to be perfectly clcar
in its construction, if above all any appearance of irregu
larity be in the least degree visible, it is classified apart.
Not that we are to imagine anything pre-established
and innate in the mind. But, since our Indo- European
languages are so formed as outwardly to distinguish
words according to the part which they play in the
phrase, the mind is accustomed to certain inflections
usually encountered in this role of a loose and super
fluous complement, and converts them into adverbial
inflections. Such is the origin of the inflections ws in
Greek , 7 and ter in Latin.
The first deposit of this kind was doubtless formed by
a few words which we may be allowed to consider earlier
than the invention of our grammatical mechanism, and
which, by the singularity of their aspect, and by the
absence of inflection, claimed a place in a class apart."
The recent date of adverbs is proved by the fact that
the Indo -European languages are not in agreement as
to choice of infections. Greek has nothing similar to
the Latin adverbs in tim or in e ; nor has Latin anything
Such are (to enumerate them under their Greek form ) åró, repl,
dal, apó, dvi, etc.
 THE PARTS OF SPEECH 183

like the Greek adverbs in dov, onv, as, Oev, Ba. This dis
agreement, which does not exist in conjugation or
declension, indicates a less ancient formation .
And yet it may be said that the adverb does exist
in our intelligence as a separate part of speech. In
French, a special inflection , which is an ancient substan
tive diverted to this use, serves as its exponent, but even
without this inflection we recognise the adverb by the
part which it plays in the phrase : Il faut parler haut.
-Des voix qui ne chantent pas juste.

The preposition is even more modern than the adverb. t+

At the time that our languages separated, there was not
one single genuine preposition. We have indicated
above the origin of this part of speech. A time came
for all our languages, when the cases, not clear or precise
enough in themselves, were escorted by an adverb . So
it is that the ablative, which by itself alone marks the
idea of removal, was accompanied by ab or ex . The
accusative, which marks the place to which motion is
directed, was accompanied by in or by ad. Now these
words ab, ex, in, ad, were originally adverbs of place, as
may still be seen in most cases by referring to their most
ancient form and usage. But the habit of seeing them
joined to a certain case suggested the relation of cause
to effect ; the little word which used to be a mere
accompaniment of the accusative or ablative now
appeared to govern them. Thenceforward it did in fact
govern them ; from adverb it became preposition.
The preposition has been so firmly impressed on our
 184 SEMANTICS

minds as a word which must be followed by an objective

case , that it is difficult to understand a preposition
4
employed alone ; it calls for, it awaits its “ complement.”
In the time of Plautus and of Terence, pre could still
be used as adverb . But a little later it is only found
followed by an ablative. The Romance languages, faith
ful in this to the Latin tradition, inherited ancient pre
positions, formed new ones, and applied themselves to
separating the preposition from the adverb as clearly as
possible ; the distinction between dans and dedans,
between sous and dessous, etc., to which Corneille had not
yet attained, has become a rule of modern French .

The agreement which exists on this point between the

diverse languages of Europe (since everywhere we see
prepositions forming themselves after the same manner)
proves that, given the general plan of their grammar,
the creation of the proposition was a foregone conclu
sion. From the moment that inflections needed the co
operation of a word to specify them, this word perforce
appeared after a certain time the cause of the infections.
It is interesting to see how this part of speech has
borrowed words from all corners of the horizon. In
French, participles like excepté, passé, hormis, vu, durant,
pendant, adjectives like sauf, substantives like chez, all
do duty as prepositions. In Latin, penes, secundum , had
already the same fate.?

" Plautus, Amph ., i. 3, 45. Abi pra , Sosia ; jam ego sequor.
Terence, Eun., v. 2, 69. I præ i sequor.
* We find in Plautus presente testibus, and in Terence presente
 THE PARTS OF SPEECH 185

The oldest prepositions have a tendency to rid them

selves of their mcaning in order to become mere gram
matical instruments. In English the particle to is often
made to precede the infinitive, simply to show that it is
an infinitive.
It is the presence of these words, apparently empty
of meaning, which has made the creation of Language
appcar a superhuman achievement.

Something of the same kind happened to the conjunc

tions. If we consider a word as devoid of meaning as
is the French conjunction que, it is difficult to conceive
how the mind could first create so abstract a sign, and
then cause it to be accepted. But the further we go
back the casier we find the explanation. The conjunc
tion que resumes its place among the pronouns. The
subjunctive which it now appears to govern preceded it.
By an illusion similar to that which we have just noted
in the case of prepositions, the mind creates between the
two words a relation of cause to effect ; a relation which
has become actual, since in the matter of languages the
errors of the people gradually become truths.
The history of the Latin conjunctions ut, ne, quominus,
quin, etc., shows us a like set of facts. These words
originally possessed a full signification, but the signifi
cation was gradually lost in the movement of the phrase,
to which thenceforward they served as a mere hinge.
The pronominal origin of such ancient conjunctions
nobis. These may be called prepositional formations which have
been arrested half-way.
 186 SEMANTICS

as ws, ut, fits them to assume successively a signification

either of time or cause. But the same fact may be
observed in conjunctions derived from substantives.
We will take an example from German.
The German word weil, " because," is an ancient sub
stantive dragged into the class of conjunctions. The form
used to be die wile, die weile, " as long as. " Luther uses
it in this way, and Gocthe too, who loved the language
of the people, often employed it. But the word was
transferred from the conception of time to the conception
of cause, as happened also with the Latin quoniam . At
the present day weil gives the impression of an abstract
word indicating the motive of an action.
Since thesc three parts of speech - adverb, preposition
3
and conjunction - have not always cxisted, but have
been formed by slow elaboration , at a comparatively
recent period, it is not rash to imagine a similar process
at a more ancient epoch for substantives, adjectives and
verbs. Not that the idea of an object, of a quality, of
an action , awaited the birth of the Indo -European
languages; no language exists without words to repre
sent objects of nature, such as man, stone, mountain , or
qualities of objects, such as large, small, high, low , far,near,
or the most obvious actions, such as to walk, run ,eat, drink,
speak. But that is not what we call the class of substan
tive, adjective and verb . The class of substantives
includes nouns representing mere conceptions of the
mind ; these nouns are treated in identically the same
way as the others. The class of adjectives includes
 THE PARTS OF SPEECH 187

words which do not correspond to any quality, as when

the Greeks said, Tpiraios Xev, "he came the third day ,"
or the Latins, nocturnus obambulat. The class of verbs
presupposes a system of persons, tenses and moods.
Taken thus, these parts of speech are not contempor
aneous with the first awakening of intelligence. They
have been formed by slow degrees, like the adverb and
preposition, although too remotely to enable us to trace
their evolution .

The species of word which must first have been dis

tinguished from all the others is, in my opinion, the
pronoun . I believe this to be more primitive than the
substantive, because it demands less invention, and
because it is more instinctive, more easily cxplained by
gesture. We must not therefore allow oursclves to be
led astray by the appellation “ pronoun " ( pro nomine),
which we derive from the Latins, who in their turn .
translated the Greek avtwrpia. The error has survived
till our own day . In my opinion pronouns are, on the
contrary, the most ancient portion of Language. How
could the me have ever existed without a designation by
which to express itself ?
From another point of view pronouns constitute the
most versatile part of Language, since they are never
definitely attached to one entity but are perpetually
travelling. There are as many I's as individuals who
| Even Reisig says that pronouns are an invention of convenience
(eine Erfindung der Bequemlichkeit) for replacing either a sub
stantive or an adjective.
 188 SEMANTICS

speak. There are as many thou's as individuals to be

addressed . There are as many he's and it's as there are
real or imaginary objects in the world. This versatility
arises from the fact that they do not contain any descrip
tive element. So that a language composed entirely of
pronouns would rescmblc the random cries of a child, or
the gesticulations of a deaf-mute. The need of another
d element, out of which arose the substantive, adjective
and verb, was therefore evident. But it is no less true
that the pronoun takes its place at the base and origin
of languages; it was no doubt through the pronoun ,
which opposed itself to other kinds of words, that the
distinction was first made between the parts of speech.
 CHAPTER XX
TRANSITIVE FORCE

Whence arises our idea of a Transitive Force resident in cer .

tain words — Verbs which change their meaning in becoming transi.
tive - Transitive Force is that which gives unity and cohesion to a
phrase - The ancient grammatical mechanism is despoiled of its
original value.

As the stones of a building, having been long and

accurately joined together, end by forming one single
whole, so certain words which are approximate in
meaning, adhere to and support each other. We are
accustomed to see them thus bracketed together, and
by dint of an illusion, of which the study of Language
produces other examples, we imagine some hidden force
which holds them together, and keeps them in due
subordination. Thus there arises in our minds the idea of
a “ Transitive Force ” resident in certain kinds of words.
We are all aware of the difference between the verbs
+
known as neuter and the verbs known as transitive : the
first are self-sufficing, and express an action which in
itself constitutes a complete meaning (such as to run ,
189
 190 SEMANTICS

to walk, to sleep) ; while the others require to be followed

by what has been called a complement. The question
has been raised : which of these two sets of verbs is the
more ancient ? In my own mind, the answer admits of
no hesitation : not only are the neuter verbs the more
ancient, but there must have been a period in which only
t ncuter verbs existed. I belicve, in fact, that words were
crcated to possess a full signification of their own, and
not to serve a Syntax which was not yet in existence.
Some of these verbs were frequently associated with
words which determined their bearing, and directed their
action on to a particular object; the mind therefore
became accustomed to the companionship, and even
expected what seemed to it an obligatory addition, a
necessary direction. By a transfer of ideas, analogous
examples of which are to be found outside philology, our
intelligence bclicved that it felt in the very words them
selves that which is merely the result of our habit of
thought : from that moment arose verbs which demanded
to be followed by a complement. The transitive verb
was born .
A double conscquence has resulted from this fact :
(1) The meaning of the verb has been modified : (2) the
significative value of case - inflections has been lessened.
· It has been agreed to reserve the name of transitive verbs for
those verbs alone which are constructed with the accusative. In
a large sense, one might also call transitive those verbs which,
like μιμνήσκω, χρήσθαι, are constructed with the genitive or dative.
It is not the choice of any particular case that is of importance; but
the close connection which has been established by the mind, to
such a degree that the verb would seem incomplete without its
accompaniment.
 TRANSITIVE FORCE 191

We will first give a few examples of words having

changed their meaning.
The root pat expresses a rapid movement like that of
a body which falls or of a bird which Aies. It furnished
in Greek πίπτω, « to fall,” πέτομαι and ίπταμαι, « to Hy.”
In Latin, it supplied petulans, impetus, acipiter, præpes,
propitius. But, having become transitive, the verb petere
denoted the impulse towards some goal ( petere loca cali.
diora, petere solem ), until at last it signified any kind
of search : petere consulatum , honores. Hence petitio,
appetitus.
This succession of meanings is so natural that we find
it again in other languages.
The Greek έκνέομαι, nearly related to ήκω and to ικάνω ,
means “ to go.” But, constructed with the accusative,
it assumes the mcaning of “ to pray." It is enough to
quote these words of Æschylus ( Perse, 216) :
Bedus dè apootporais lxvovnévn .
Imploring the gods with sacrifices .

It has supplied, in this acceptation, the derivative IkeTIS,

" suppliant," whence IKETEÚw , " to implore."
In Sanscrit, the verb jā, the ordinary meaning of
which is "to go," acquires the mcaning of “ to pray," if it
is followed by an accusative. The Vedic tat tvā jāmi
( literally "te hoc adeo " ) is interpreted by tat två jāce,
)

“ I ask you this.”

We will now turn to an association of ideas which is

the counterpart of the preceding one.
The verbs which signify " to retire," assume, when
 192 SEMANTI
CS

they become transitive, the meaning of " to cede, to

abandon . "
Cedo means properly “ to retire ” : this is the meaning
which it has preserved in recedo, discedo, decedo. Cedere
alicui therefore meant “ to retire in favour of some
one, to yield one's place." The idea of yielding one's
place having later become the symbol of all kinds of
concession, cedo assumed the meaning of “to yield ."
Then by a fresh advance, it came to be constructed with
the accusative, and signified " to cede, to grant.” Cedere
multa multis de jure suo. - Cedere possessionem . - Cedere
victoriain .
The same succession of meanings is to be found in
Greek. Είκω means to retire . Είκειν θυράων , κλισμοίο ,
Tolémov, " to retire through the door, from a throne,
from the war. " The Scholiasts rendered it by útoxwpéw ,
παραχωρέω..
Later the Greeks said : είκειν οργή , θυμώ , ανάγκη, "to
yield to anger, to passion, to necessity ."
But cikw , being constructed with the accusative, as
sumed the further signification of “ to leave, to abandon ."
Nestor, giving advice to his son about a chariot race,
tells him that in turning the goal he is to excite the off
horse by his cries, and to give it the reins :
τον δεξιον ίππον
κίνσαι όμοκλήσας, είξαι τί οι ηνία χερσίν .
This succession of ideas is so natural that one may

' Inversely, obstare has in French attained to themeaning of

enlever, “ to take away." The original form of usage was : “ Oter
laretraite d quelqu'un, lui ôter lesmoyens de vivre."
 TRANSITIVE FORCE 193

expect to find it again in other languages. The Germans,

for example, for “ to retire from an affair ,” say von
einem Geschäft abtreten . The verb is, in this case , neuter,
and possesses its original signification. But they also say,
making this same verb transitive : Jemanden einen Acker,
ein Recht, ein Land abtreten, “ to cede to some one a
field, a right, a territory.” ? In English the verb forego
or forgo means in a like manner “ to retire from ” and
“ to cede.”

It is a long way from “ to stand upright" to " to com

prehcnd, to know ." Yet this is the change which has
come about for the root sta, not once, but in at least thrcc
cascs.

We have the Greek lotnus, which, in combination with

dal, gives énlotapal “ to know ," whence &Tlothun, “ clever.
ness, science ."
We have, on the other hand, the German stehen , which
produced verstehen, “ to understand," whence Verstand,
“ intelligence .” Already in Middle High-German verstan,
and in Old High -German firstån, signified "to com
prehend.”
Finally, in English we have stand, whence understand,
which was preceded by the Anglo-Saxon forstandan
(with same meaning).
Jacob Grimm, in his Dictionary, has inverted the order of
things. He considers the transitive sense as the more ancient of
the two. He translates by deculcare, and gives as first example :
den Absats vom Schuh, den Schuh vom luss abtrcten . In the ex
pression : ein Land abtreten, " to cede a territory," he discovers a
figure of speech : mit dem Fusce von sich abtreten. The metaphor
would at the least be a strange one.
194 SEMANTICS

Fully to realise this change, we must remember that

the first arts were not taught by books : they were prac
tical arts, in which the first necessity was to learn the
proper attitude and position. Such was the art of hurl
ing the javelin, of wielding the club, or again the art of
striking fire from the fint, or of horse -taming. On the
other hand, we must take into consideration the fact
that driotauai is a verb of middle form , that is to say, a
reflexive verb : it means literally “ to hold oneself.”
Verstehen in German is still often a reflexive verb. It
is quite usual to say : sich auf etwas verstehen ; er
versteht sich auf Astronomie, auf Literatur, auf Politik.
We see henceforward how a verb which signifies
“ to hold oneself ” can acquire the meaning of “ to
know " : er versteht sich auf das Speerwerfen, auf das
Pferdebändigen .
Homer ( II., xv. 282) employs the participle émotduevos
with the dative :

Toioi d' Umur' dgydpeve obas, 'Avdpaluovos viós,

Αιτωλών όχ’ άριστος, επιστάμενος μεν άκοντι,
Εσθλός δ' εν σταδίη.
lis autem contionatus est Thoas, Andræmonis filius,
Ætolorum longe præstantissimus, peritus quidem jaculi,
Strenuus etiam in stataria .

The commentators choose to imagine an implied word

udpvardal. But this is unnecessary ; it could be trans
lated into German without ellipse : sich auf den Wurf
spiess verstehend.
There was thenceforward but one step to saying, as
we find already in Homer : ανήρ φόρμιγγος επιστάμενος
 TRANSITIVE FORCE 195

και αοιδής, or again επιστάμενοι πολέμοιο. Finally , we

already have επίσταμαι with the accusative : πολλά
δ ' επίστατο έργα.
The history of the two Germanic verbs is just the
same. The German says, using the accusative : Verstehst
du mich ? —Keiner hat die Sache verstanden . And in
English : Do you understand me 2 - Who has understood
the apologue ?
These three examples show in the clearest manner
that Transitive Force does not limit itself to the
establishment of a link between the verb and the com
plement : it transforms the meaning of the verb.
There is a historical conclusion to be drawn from
these facts.
When we survey the lists of “roots drawn up by
the Hindoo grammarians and adopted, subject to rectifi
cation , by modern science, we note that most of them
already possess a transitive meaning. This would prove,
were proof needed, the antiquity of Syntax. But we
should risk a wide deviation from the truth , were we to
believe that the meaning attributed to these roots is the
original and initial meaning. Many, assuming a transi
tive value, must have changed their acceptation. The
examples which we have just given abundantly prove this.
Those who believed that the root man mcant from the
first "to think ," or the root budh, “ to know ," would make
the same mistake as if " to ask, to pray " were set down
in a Latin historical dictionary as the original meaning
of petere.
196 SEMANTICS

Wc pass now to the second consequence, the deterior

ating of the significative value of the case - inflections.
It is interesting to note how Transitive Force gradu
ally enters into conflict with the original value of the
cases, or -- to speak without metaphor - how the force
of habit results in a certain case being in course of time
considered as pre-eminently the complementary case.
The Latins first said with the accusative : petimusurbem ,
because the accusative marks the goal towards which
motion is directed. But helped by Analogy, they also
said : linquimus urbem , fugimus urbem , so that the
accusative, from being a local case, became a gram
matical case. Nothing could be more destructive to the
original value of the inflections.
Sequor means literally, “ I attach myself ” : it corre
sponds to the Greek énopar, which takes the dative. Yet
the Latins said : Sequi feras, sequi virtutem . Meditor
means " I exert myself ” : it corresponds to the Greek
uelet@mai, of which it is a more or less exact copy. Yet
they said : meditari versus, meditari artem citharædicam ."
When once the type of transitive verb had been
adopted, it multiplied rapidly. Verbs like dolere, flere,
tremere, which by nature would seem as if they ought to
remain without complement, were commonly declined
with the accusative : Tuam vicem doleo. — Flebunt Ger
1 Meditor, meditatio are terms of school or gymnasium come
from Greece to Italy : they represent the Greek μελετών, μελέτη,
Modéroua. A military exercise was called meditatio campestris ; an
oratorical exercise, meditatio rhetorica. Virgil uses the word as a
ncuter verb and in its proper sense, when he says : meditantem in
prælia taurum .
 TRANSITIVE FORCE 197

manicum etiam ignoti.-- Te Stygii tremuere lacus. Thus

the spirit of imitation can go far. Amo having become
a transitive verb, ardeo, pereo, depereo, demorior also
became transitive. We find in the comic poets : Is
amore illam deperit.

All ancient languages did not in this respect arrive at

the same point. Greek preserved for a longer time than
Latin a sense of the value of the cases. Thus a certain
number of Greek verbs take the genitive.
It is on account of the partitive idea expressed by the
genitive that we find it used with verbs signifying " to
eat, to drink .” It is the same in French : they say, “ boire
du vin ," and not " boire le vin." Ilvew otvov, udatos,
yálaktos is the usual construction. For a like reason ,
the genitive is used with verbs signifying " to taste,
to touch, to take, to obtain .” ! When Thetis, imploring
Zeus, touches his chin, the poet says : kal Maße xeipi
yevelov. Again for the same reason , the genitive is
employed with verbs signifying " to desire," such as
ζεσθαι, ορέγεσθαι, επιθυμεϊν.2 Hector is seized with a desire
to embrace his child :
“Ως είπών ου παιδδς ορέξατο φαίδιμος Έκτωρ .
The verbs which express the activity of organs, such
as “ to hear, see, know, remember," complete this series.
There is, in fact, a difference between the effective and
1 Θιγγάνειν, ψαύειν, τυγχάνειν.
? This is what has been misunderstood by many excellent gram
marians, who have preferred to assume an ellipse. Thus Kühner
(3 415) explains επιθυμώ της σοφίας by επιθυμώ επιθυμίαν της σοφίας.

$
198 SEMANTICS

direct possession, expressed by the accusative, and the

more or less superficial attack, expressed by the genitive,
which is suitable to these verbs of intellectual signification.
Latin has preserved one single specimen of this kind,
memini, which takes the genitive, as iſ to prove that this
construction was not always alien to the languages of
Italy. But memini itself is joined to a complement in the
accusative: Suam quisque homo rem meminit, says Plautus.
And Virgil : Numeros memini, si verba tenerem .
Latin, while levelling its syntax, preserved the memory
of a condition at once more ancient and more akin
to Greek. The verbs signifying " to desire, to like,"
ended by treading the usual path, that is to say, they
are followed by the accusative ; but the adjectives
or participles derived from these verbs remained
faithful to the ancient construction. Cupidus fame,
amans laudis, were still used with the genitive, although
cupere, amare had long ceased to be employed in this
way.
In like manner the construction with the genitive is
preserved in Sanscrit. It is even continued in some
cases in modern German : Iss des Brodes. Geniesse dieser
Freude. Wir pflegen der Ruhe.
The ancient grammatical mechanism, then, has not
been abolished : but it has been despoiled of its original
value for the benefit of a new order. The phrase, in this
new period of Language, is composed of words which are,
on the one hand, governing, on the other hand, governed.
Syntax confiscates for its own profit the individual
meaning of inflections. This, to borrow from Germanic
 TRANSITIVE FORCE 199

mythology , might be called " the twilight of in

flections. ”
Must we regard this adaptation to new usages as
decadence or as progress ? The question may seem an
idle one, since every epoch forms for itself the language
which it needs. But if an answer must be made I would
say that we should consider it as an advance. If it be
in the nature of all arts to transform themselves, how
should the most necessary of them all, which is made
to accompany thought in its every step, have failed to .
transform the material left to it by the childhood of
humanity ? The advance is obvious to all eyes. The
words which were, so to speak, shut up in themselves,
are gradually linked with the other words of the phrase.
And the phrase itself, though composed of small pieces
which are immovable and related, appears now a work
of art possessing its centre, its lateral parts, and its
dependencies, now an army on the march, with all its
subdivisions in connection with and in support of one
another.
 CHAPTER XXI

CONTAGION

Examples of Contagion-Negative words in French - The English

but — The active past participle — The Latin conjunction si.
I HAVE already proposed to call by the name of
Contagion a phenomenon which makes its appearance
fairly often, and which has the effect of communicating
to a word the meaning of its surroundings. It is obvious
that this Contagion is nothing but a special form of the
association of ideas.
French supplies an example, which is well-known, but
so much to the point that I feel bound to recall it.
We all know what happened in the case of the words
pas, point, rien, plus, aucun, personne, jamais. They
a served to reinforce the only genuinc negative, to wit ne.
Je n'avance pas ( passum ).— Je n'en vois point (punctum ).
-Je ne sais rien ( rem ).- Je n'en connais aucun (aliquem
unum ).— Je n'en veux plus ( plus ).- 11 n'est personne
(persona ) qui l'ignore.— Je ne l'oublierai jamais ( jam
magis ).
These words, by their association with the word ne,
200
 CONTAGION 201

became themselves negatives, and to such good purpose

that they can dispense with their companion. Qui va
la ? Personne. - Pas d'argent, pas de Suisse. — Sans la
connaissance de soi-même, point de solide vertu . - Son
style est toujours ingénieux, jamais recherché.
In Semantics it is of interest to consult alternately,
on the subject of these words, a modern and an historical
dictionary. This comparison is like taking the sound
ings of the mind. The two answers obtained are con
tradictory, but, on reflection , though mutually opposed,
are both of them reasonable and legitimate.
The French Academy, in its dictionary of usages,
places the negative sense before all the others.
“ Aucun ,'” says the edition of 1878, " adj. Nul, pas un.
-Rien . Néant, nulle chose.”
For which the Academy must not be blamed . It was
part of its plan to explain words according to the
impression which they make to -day. It was, moreover,
the same impression which they made in the seventeenth
century :
" Laissez faire, ils ne sontpas au bout,
J'y vendrai ma chemise, et je veux rien ou tout."
Racine (Plaideurs).
And even in the eighteenth :
“ Car de rien fait -il tout saillir,
Lui qui a rien ne peut faiblir."
Let us now turn to Littré :
" Aucun ,' quelqu'un . - Rion ,' quclque chose,
We see what a distance lies between the original
meaning and the meaning produced by the long sojourn
202 SEMANTICS

in negative phrases. It must indeed be added, that it is

not only through negative, but also through inter
rogative phrases, that the change came about: De tous
ceux qui se disaient mes amis, aucun m'a - t - il secouru ?
Auriez -vous jamais cru 2 - Avons -nous rien négligé ?
There are cases in which the meaning remains half
way between the two acceptations : Il m'est défendu
de rien dire.- Je doute qu'aucun homme d'honneur y
consente .
It is not therefore direct contact nor the actual neigh
bourhood of the negation which is the cause of the
change. The contagious action has been produced by
the general meaning of the phrase.
Something of the same kind exists in English.
The English but, which comes from the Anglo-Saxon
bitan ( = be-utan ), means properly “ outside." 1 When
it means “ only ,” it stands for ne but. The negation was
finally suppressed. "We have here but five loaves and
two fishes " (Matt. xiv. 17). Such is the text of the
Authorised Version. But the Anglo-Saxon Gospel
says: "We nabbad (ne habbad ) her buton fif hlafas und
twegen fiscas." In the course of time the negation
became superfluous, the particle but having taken the
meaning upon itself.

Contagion supplies, I think, the true explanation of a

fact connected with the French language which has
greatly occupied grammarians—the change from past
· Dutch buiten. Hence, in opposition to Binnenzee, “ the inner
sea,” Buitensee, “ the outer sea." - Storm , Philologie Anglaise, p. 8.
 CONTAGION * 203

participle passive into participle active. In these

phrases : " J'ai reçu de mauvaises nouvelles, j'ai pris la

route la plus directe,” reçu , pris, have to-day the active
sense which they owe to the proximity of the verb
avoir. The proof that they possess this active sense is
that in telegraphic language I should say : " Reçu de
mauvaises nouvelles. - Pris la ligne directe. "
Herein lies, if I mistake not, the reason for that rule
of non-agreement which has been so doubtfully and so
variously explained. The truth is that the participle
has, by Contagion, become active. It has incorporated
itself with its auxiliary. But as time has been necessary
for the operation of this change, as the ancient turns of
phrase were long in dying, and as the slightest deroga
tion from the ordinary course provides them with a
pretext for continuing to exist, the change in question
has only been accomplished in the case of the construc
tion most frequently in use, that which we are accus
tomed to regard as the normal construction. Everywhere
else Language is faithful to the ancient grammar.

I will give one more example of the force of

Contagion .
Whence comes the conditional idea which is sug
gested in French, as in Latin, by the conjunction si ?
To explain this we must go back many centuries.
The Latin particle si was primitively an adverb signi
fying “ in this way, in this manner.” The conditional
idea came to it by the proximity of the subjunctive or
of the optative. The old formula of the invocations
204 SEMANTICS

and prayers : Si hac, Dii, faritis, derives its hypothetical

signification from the verb. The meaning was at first
the same as though it had been : Sic, Dii, hæc faritis.?
The second proposition comes later to enunciate a
second fact, the consequence of the first : Ædem vobis
constituam . The mind grasped a link between these
two propositions, and as on both sides the action is
presented as contingent, it very naturally introduced
into the first word the idea of a supposition or condition.
Already in the afore -mentioned formula, when em
ployed by contemporaries of Æmilius Paullus, si was
a conjunction. To such a degree had it become one, so
markedly had it assumed the conditional idea, that it
could be followed by an indicative. Si id facis, hodie
postremum me vides.3
Conjunctions of the same kind in other languages
have an analogous origin. When looked at closely,
these little words are nothing but pronominal adverbs,
in which there is nothing to announce a supposition or
condition.

1 In more modern language, si hæc, Dii, feceritis.

• The adverb sic is nothing but si accompanied by the enclitic
which we find in nunc, tunc.
s French went even farther. The conditional, after si, would
seem a pleonasm.
 CHAPTER XXII

ON CERTAIN GRAMMATICAL INSTRUMENTS

The relative pronoun - The article — The verb substantive
Auxiliary verbs.
ONCE the idea of a phrase forming a whole has
impressed itself on the minds of men, they feel a desire
to complete it, and to give it the instruments which it
needs. But since the popular intelligence, as we have
seen , confines itself, without creating anything, to adapt
ing what has been furnished by anterior centuries to
new uses, a certain number of words are converted for
the requirements of Syntax.
A first conversion - the most important of all — is that
which has given us the relative pronoun.
A certain pronoun, in no way distinguished exter
nally from others, acquires, by the use to which it is put,
a force of union which allows it to weld together two
propositions. This is expressed in grammatical language
as follows: from demonstrative it becomes relative or
anaphorical
A somewhat advanced syntax is needed for this
205
206 SEMANTICS

conversion to take place : in the divers Indo -Euro

pean languages the choice of the relative pronoun
came about late and not always in the same way. To
verify this, we have only to compare the Latin qui with
the Sanscritjas and the Greek ős. The Greek language
in the time of Homer, and even later, in the time of
Sappho and of Alcman, had not yet made its definite
choice. It hesitated long between the pronouns ja, ta ,
and sva .?
One naturally wonders at what epoch so necessary a
means of expression began to exist. With regard to
this, we must make a distinction between the idea of the
relative, and the final adoption of any particular pro
noun. The idea of the relative is probably anterior to
the separation of our languages, since we find every
where a certain pattern of phrase, always the same,
which presupposes the presence of a relative pronoun .
Proverbs and popular adages are inclined to this turn of
phrase :
Quod ætas vitium posuit, id ætas auferet. Quod aliis
vitio vertis, id ne ipse admiseris. Qui pro innocente
dicit, is satis est eloquens. Cui plus licet quam par est, is
plus vult quam licet. Quam quisque norit artem , in hac
se exerceat.

i In the Homeric language, to is the ordinary anaphorical pro

noun. Ex .: El més ris Beds doon, Tol oupardy eupùr Yxovo i.- 'Anna
συ μεν χαλκόν τι άλις χρυσόν τε δέδεξo, Δώρα , τα του δώσουσι πατήρ και
πότνια μήτηρ. Etc.
• The very generally admitted identification of 8s with jas is not
certain : from the form Foto preserved in an inscription of Locris,
one is led to suppose that 8s corresponds to svas.
 ON CERTAIN GRAMMATICAL INSTRUMENTS 207
This type of phrase reappears in Sanscrit : 1
“ Whose the mind, his the power.” Jasja buddhis,
tasja balam .
“ Who loves, fears. " Jasja snehas, tasja bhajam .
" For whom the gods prepare death , from him do they
wrest the mind. " Jasmäi devās prajacchanti parābhavam ,
tasja buddhim apakarśanti.
" As is a man towards others, so must one be towards
him ." Jasmin jatha vartatejas, tasmin tatka vartitavjam .
" What thou givest, in that lies thy wealth . ” Jad
dadasi, tad te vittam .
“ As do the great, so do the rest of mankind . " Jad
acarati çrešthas, tad itaras ganas.
The same construction is already of current use in
the Vedas : " Quod sacrificium protegis, id ad deos per
venit. " Jam jagnam paribhur asi, sa deve u gacchati.
“ Qui nos lacesset, procul eum amovete." Jo nah prt
tanjad, apa tam dhatam .
We shall be asked the reason why the relative pro
position is thus launched before the chief one. I believe
that we have here a semantic fact of which examples
are to be found in other families of languages. By
the action of the mind an interrogation must be
established , with the result that the two propositions
See Boehtlingk, Indische Sprüche. By not referring to Indian
ists, we have simplified the quotations and suppressed the effects
of sandhi.
* The type of these constructions has been preserved in our
proverbs : Qui aime bien , châtie bien . Spare the rod, spoil the
child. Whom the gods love, dieyoung. Etc.
208 SEMANTICS

form the question and answer. This is probably the

reason why a large proportion of the Indo - European
languages make the one pronoun fulfil both the
interrogative and relative functions.
To appreciate to its full extent the importance of the
relative pronoun, we must remember to how many deri
vatives it gives birth : first, words like qualis, quantus,
quot; then the conjunctions, quod, quia, quum, quoniam .
In Greek : ώς, ότε , ή, ου, όθεν, ηνίκα , ότι, also derivatives
like őros, olos. In Sanscrit, derivatives such as jādriça,
jāvant, to which must be joined the most important con
junctions, jad, jadi, jatra, jada, jatha. The creation of
a relative pronoun is therefore one of the capital events
of the history of Language ; without a word of this kind,
every idea possessed of any force, of any completeness,
was impossible. But this creation was obtained by the
slow conversion of one of those numerous pronouns
which served to accompany a gesture in space. So
we here find the human mind patiently forging the
instrument of which it is in need .

The same can be said of the little word which the

Greeks, comparing it with the articulations of the body,
called åpOpov, and which we call the article.
We know that the article is an ancient demonstrative
· For additional detail, see, in the Studien of Curtius, the articles
of Windisch in vol. ii. and of Jolly in vol. vi. See also Delbrück,
Grundriss, ß 222, s. and the thesis of Ch. Baron, Le pronom relatif
et la conjonction en grec. Essai de syntaxe historique. Paris,
Picard, 1891 .
 ON CERTAIN GRAMMATICAL INSTRUMENTS 209

pronoun . But the signification of this demonstrative

pronoun has been to a certain degree converted. It
has been confiscated for the benefit of the syntax .
We can take as example the French article le, which
represents the Latin ille. This last served to designate
objects or persons : Magnus ille Alexander 1-Ita ille
farit Jupiter ! But in course of time the demonstra
tive sign became a mere grammatical indication : " La
personne dont je ťai parlé hier . - Les pays que nous
avons traversés." Here the article only figures as the
antecedent of the relative pronoun. It has become a
grammatical instrument."
The usefulness of the article can be felt more than
explained. Latin , from being deprived of it, is often
weighed down and hampered in its course . Greek , on
the contrary, which early felt its need, owes to it in part
its suppleness. The conformity of the French language
to the Greek, noted by Henri Estienne, arises in some
degree from this. I need only recall these turns of
expression : οι πάλαι σοφοί ... εν τω μεταξύ χρόνω ..
των νυν οι τότε διέφερον. . Or these: ορεγόμενοι του
πρώτος έκαστος γίγνεσθαι, etc.
It has happened indeed that the article has finally
been inserted where it brought no appreciable help. It
might be said that the languages in which it is of the
greatest use are those which remain free to employ or to
· Definitions of the grammarians : " An article is a word placed
before a substantive to indicate whether it is masculine or feminine."
- “ An article is a word placed before a noun, to show whether this
noun is used in a particular or in a general sense. ” Etc.
P
210 SEMANTICS

omit in accordance with the meaning. There is no

doubt that French, within the last two centuries, has
extended its use beyond due limits, so that it has
become less useful in proportion as it became more
indispensable.

We must also make mention of the verb to be, which

was declared by the Scholasticism of the Middle Ages to
be a mere “ copula " ; a fact which shows the impression
now inade on the mind by this verb, after reaching the
climax of its evolution. Yet it undoubtedly began by
some concrete signification : others have followed in
the path, such as fuo, exsto, evado. If they have not
attained to the same degree of colourlessness, this is
attributable to a difference of age, not of nature.
Something of the same kind took place in the case of
the verb to have. When I say , “ This man has lost all
that he had," I employ twice over the same verb to
have ; and this without objections from any one, to such
a degree has the auxiliary verb been converted by
change of use into a word of a distinct species.

Thus Language deducts from the hereditary stock a

certain number of expressions of which it makes gram
matical instruments. One who has only known them
in this last capacity, has difficulty in imagining that
there was a time when these words had their proper
signification. An author of the eighteenth century
draws attention to the fact that in this expression : Il
a été ordonné ... " it has been decreed ,” three words out
 ON CERTAIN GRAMMATICAL INSTRUMENTS 211

of the four serve simply for the construction of the

sentence. The number of these words increases slowly
with the lapse of centuries, since, on the one hand,
Specialisation of Function ' tends to create new ones,
while, on the other hand, Transitive Force mixes
them more and more, as a necessary element, with
the structure of the phrase. This is the reason why
etymology, when confronted by a modern language, and
deprived of carlier documents to enlighten and to guide
it, gocs so hopelessly astray.
1 Scc above, p. 11 .
 CHAPTER XXIII

THE ORDER OF WORDS

Why strictness of construction is in inverse proportion to rich

ness of grammar - Whence comes the Order of French construction
-Advantages of a fixed Order - Comparison with the modern
languages of India .
AMONG the different methods of expression of which
our languages makc use, the Order of words, that is to
say, a certain fixity in the construction of the phrase-a
fixity which of itself often determines the meaning of
vocables — is the method which was thought of last. It
is because this method is in fact a shade less material
than the others. In this phrase, “ The Japanese have
vanquished the Chinese," the position alone indicates
which is the object and which the complement ; change
the Order, while keeping the words, and the contrary
assertion is the result. This may be compared with the
Arabic system of numeration, in which each number, in
addition to its proper value, has a value of position.'
This circumstance alone might teach us that we are
· Jespersen, Progress in Language, p. 80.
212
 THE ORDER OF WORDS 213

in the presence of the work of centurica, As a matter

of fact the ancicnt languages, so superior from other
points of view, show nothing at all approaching to it.
Here a question occurs to us which often presents
itself in the history of languages and of human affairs
in general. Is it the loss of inflections that has resulted,
by way of compensation and makeshift, in the growing
strictness of construction, or is it rather that a more
regular construction has rendered the inflections useless ?
The answer is one which is generally the right solution
for dilemmas of this kind : both . In proportion as these
inflections disintegrated, the necessity of a fixed Order
made itself additionally felt ; and, on the other hand, the
habit of this fixed Order completed the downfall of the
inflections. It may be imagined that official documents,
such as charts, diplomas, public or private acts, contracts
of all kinds, in which it was most important to avoid all
possibility of misunderstanding, were the first to introduce
the habit of a uniform construction as they were also the
most persevering (and this implies no contradiction ) in
their attempts to retain the inflections. The two
methods, employed simultaneously , were bound to
converge towards the same goal. This explains the
preservation of the declension with two cases for certain
names of relationship , such as fils and fil, enfes and
enfant; for certain titles such as cuens and conte, ber and
baron, and for certain proper names such as Jacques and
Jacque, Hugues and Hugon . While these differences of
inflection were omitted at last, the Order of words only
· gained in strength.
214 SEMANTICS

The question of the Order of words is never raised but

another follows in its wake : is it an advantage or a
drawback to have a fixed and invariable construction ?
It has been the fashion to praise the freedom of Latin
and Greek, which allow the word towards which atten
tion is to be drawn and light directed, to be pushed to
the front or rcscrvcd till thc cnd of thc scntcncc. But,
to be just, it must be recognised that thosc languages
which are the most bound to a certain Order are not for
that reason absolutely fettered. Perhaps even inversion
produces the greater effect for having more completely
broken through the ordinary custom.
It is in any case certain that a predetermined Order is a
relief, if not for the writer or the speaker, at all events for
him who reads or listens. If we read an ode of Horace,
in which the adjective is often far away from its substan
tive, or a speech of Cicero, in which the crucial word
only comes at the end of a whole long period, we feel
that in French things are made casier for us. It is
probable that the manner of declamation contributed
towards the understanding of the phrase ; perhaps even,
in the public market-place, those words announced from
afar, so long awaited , were the only ones which reached
the ears of the audience. On the other hand, the tend
ency of all literatures is to exaggerate, to extend beyond
due limits, to push to an extreme the resources of
cxpression which are supplied to them by the ordinary
language of daily life : it may therefore be imagined that
the ingeniously distorted construction of the Greek and
Latin lyrics is up to a certain point a trick of style. The
 THE ORDER OF WORDS 215

speech of ordinary conversation, such as we find it in

the comic poets and in familiar letters, is not nearly so
tortuous.

Since the Order of words becomes stricter in proportion

as the grammatical resources diminish, every disturbance
of the construction involved the risk of changing the
meaning. Wc are acquainted with thosc sccrct locks,
the mechanism of which only works when the different
parts are disposed according to a preconcerted arrange
ment. Our modern languages are in a like case. Modify
the Order : either the meaning will be modified also, or
we shall cease to understand .
It is especially in ready-made expressions, which
sometimes preserve the signs of a more ancient grammar,
that this Order must be maintained : a delicate test and
touchstone by which the foreigner of imperfect education
is revealed .
The expression, "Logical Order," has been used with
reference to the French phrase. There is here a certain
exaggeration . It is the moment to recall the remark of
an English writer, that it is with this as with the anti
podes : every nation is tempted to think that it alone
places its words in their true position. One can easily,
without being wanting in logic, conceive a different
Order. In the primitive plan of our languages, the verb ?
was followed by its subject ( Slowpe, Oldwoi). Without
leaving the French language, we find propositions which
place the subject at the end."
1 " Les arbres qu'avait abattus le vent .” “ L'homme de qui
dépendait notre sort." Etc.
216 SEMANTICS

Rivarol especially, in his Discours on the universality

of the French language, allowed himself to be led into
eulogies, which are at once vague and excessive :
" French , by a unique privilege, has alone remained
faithful to the direct order, as though formed wholly
of rcason . .. In vain do passions overthrow and strive
to seduce us into following the order of sensations ; the
French syntax is incorruptible. From this arises that
admirable lucidity, eternal basis of our language. When
the French language translates an author, it in reality
explains him ....
It is not the French language in abstracto which
ought to have been praised, but the persevering effort
of its writers during the last three centuries, to render
the liberties of its syntax proportionate to the resources
of expression of which the language has command. In
this they have displayed a singular honesty. They
understood that, in writing, clearness was one of the
forms of probity. Those who, under pretence of progress,
or in imitation of foreign literatures, desire at the present
day to free themselves from these ancient rules, should
first give to their language the means of doing without
them .

This is the place to recall the hypothesis propounded

on the subject of monosyllabic languages like Chinese,
in which the rules of construction alone constitute almost
the entire grammar. It has been conjectured that mono
syllabism represented not a primitive condition, but, on the
contrary , the old age of a language, in which everything
T
 THE ORDER OF WORDS 217

was worn out and laid bare. It is possible, indeed, that

it might prove necessary to reverse in this way the series
of linguistic periods. It would then have to be supposed
that our languages, in abandoning more and more their
grammatical mechanism, were doomed to arrive some
day at a more or less similar condition. It is true that
literary tradition would, in case of need, be a safeguard
for them, a safeguard which was lacking to the Middle
Kingdom, since Chinese writing preserves the thought
without transmitting the language.
It will not be without use to add here, by way of
counterpart, the fate which has befallen the idioms
derived from Sanscrit. To the ancient cases of the
Sanscrit declension there were joined words which had
the same meaning as our prepositions, έν, πρός , παρά, επί,
etc., but which, in mingling with the preceding substan
tive, soon produced the impression of case -inflections.
From this resulted declensions of a wholly new aspect.
It is in this way that we find locatives ending in majjhe,
majjhi, mahi, mai, which represents for us the Sanscrit
word madhje, " in the middle ." Another locative ends in
thāni, thāi : this must be regarded as the Sanscrit sub
stantive sthānē, coming from sthānam , " the place." A
third locative ends in păsë, pasi : this is the Sanscrit
pārçvë, “ at the side."
The dative is in a like manner represented by inflec
tions of great variety. It may be by kāchē, kahi, khē,
which is the Sanscrit word kakše, " at the side." It may
be also by lidhë, lajē, laē, lai, lē, which is the Sanscrit
labdhë, “ for the good of." It may be by athim , which is
218 SEMANTICS

the 'Sanscrit arthi, " in the interest of." It may be by

kägi, which is the Sanscrit kärjē, " for the benefit of."
It may be by bâtī, vātī, which is the Sanscrit vārtlo, “ in
favour of.” 1
Here then we are confronted with the spectacle of
a language which, instead of attaining to simplicity,
like the Romance languages, by acquiring distinct
exponents, has only succeeded in creating fresh con
fusion .
1
Hoernle, A Comparative Grammar of the Gandian Languages.
London, Trübner, 1880, p. 224, s.
 CHAPTER XXIV

THE LOGIC OF LANGUAGE

The nature of the Logic of Language - How the popular mind

proceeds.
LANGUAGE possesses its Logic. But it is a special ㅈ*
Logic, in some sort professional, which is not to be con
founded with the science usually called by this name.
Logic properly so called forbids, for instance, the union in
one judgment of contradictory terms, such as saying of a
square that it is long. Now, Language has no manner of
objection to this : it even allows us to say, if we wish,
that a circle is square. But, on the other hand, it makes
certain prohibitions which are a matter of indifference to
logic ; it forbids, for example, a verb in the singular to
be joined with a plural subject, or an adjective to be in
a different gender to its substantive. These are technical
rules, at the same time both narrower and wider than the
rules of the art of thinking.
There have often been attempts to find a sort of logical
framework behind the rules of grammar, but Language
is too rich and not sufficiently rectilinear to lend itself
219
220 SEMANTICS

to this presentment. It overflows the bounds of logic

on every side. Moreover, its categories do not coincide
with those of reasoning : having a method of procedure
peculiar to itself, it at times constitutes grammatical
groups which are not reducible to any abstract con
ception.
Those who seek for the fundamental notion expressed
by the subjunctive, and who expect to find this funda
mental notion by comparing all the usages of the sub
junctive, and so eliciting their common clcmcnt, arc, I
do not hesitate to say , on a wrong track. Thcy can only
attain to an extremely vaguc and general idca, such
as the people at large would find hard to conceive, and
such as we have no reason to attribute to primitive ages.
Yet it is the method habitually pursued by those who
propose to explain to us the essential idea of a mode,
case, conjunction, preposition .

Popular Logic does not proceed in this way. It ad

vances, so to speak, by stages. Starting from a very
circumscribed and definite point, it pushes straight ahead ,
and arrives, in all ignorance, at a stage at which by the
nature of things — I mean by the general tenour of the
speech — a change is produced. Thenceforward there
exists a halting -place from which there can be a fresh
advance at a different angle, without any interruption
of the original direction . This already furnishes two
meanings. Then the same things recur in a third stage,
which supplies a third orientation. And so on . In
all this procedure there is no generalisation, but a march
 THE LOGIC OF LANGUAGE 221

in a broken line, in which every stopping-place, present

ing the idea in a different incidence, becomes in its turn
the starting -point of a new line of advance.
In order to verify this we will glance through a
chapter of syntax, with apologies to our readers for the
dryness of the details, and for the school memories which
they cannot fail to evoke. But it is a question of rectify
ing a prevalent misconception, and of showing once for
all, on well-defined ground, how the rules of grammar
join on to one another.
We choose as examples the rules concerning the
accusative, on account of their apparent complication.
What is the fundamental idea of the accusative ? It
will be remembered that our text-books distinguish the
accusative of the direct object, that which marks
duration, that which marks distance and extent, that
which indicates the goal. The variety is considerable.
One of our leading philologists, renouncing the hope
of finding the essential idea, declares that he is tempted
to apply to the accusative what the Hindoo gram
marians say of the genitive : to wit, that it is admissible
on all occasions when other cases cannot properly be
employed. The search for the primary idea does not,
however, seem to us so difficult.
If we can anywhere find the accusative employed alone,
with no accompaniment, there is a chance of our being
enlightened as to the original signification. Latin has,
as a matter of fact, a usage in which the accusative is
self - sufficient.
 222 SEMANTICS

It is in official language, which varies at the slowest

rate, and preserves its archaisms for the longest period
of time, that we find this usage. The following is the
beginning of the inscription on a milestone of Southern
Italy : 1

“ HINCE SUNT NOUCERIAM MEILIA CAPUAM XXCIII

MURANYM , LXXUI COSENTIAM CXXIII
VALENTIAM CLXXX . "

The accusatives Nouceriam , Capuam , Muranum , Cosen

tiam , Valentiam , accompanied each time by a number,
mark the distance from the milestone to these towns.
The accusative is therefore here used as case for the
+ goal towards which motion is directed.
This usage has been preserved in poetic language :
“ Hac iter Elysium ," says the priestess of Virgil. We find
the same turn of phrase in certain expressions : Malam
crucem , " go to the devil.”
We have taken Latin as example ; but the same use
of the accusative exists in Sanscrit. “ (Come) on the
carth, O God, with all the Immortals !” Deva, kšim ,
viquibhir amritibhih .
From the moment that the accusative, by itself alone,
expresses direction towards a place, it is not surprising
that it should have been joined to verbs signifying "to
go ” : Language here unites two words, the association
of which was fully indicated. Thus was born a first
syntactical usage .
1 Corpus Inscriptionum Latinarum , 1 , No. 551 .
· Æneid , vi. 542.
 THE LOGIC OF LANGUAGE 223

“ Ibitis Italiam , portusque intrare licebit

Ad nos hinc alii sitientes Afros.
Italiam fato, profugus Laviniaque venit
Litora . ” 1

In Greek, examples are numerous :

κνίση δ'ουρανόν εκε. ??
έβαν νέας αμφιελίσσας..'2
πέμψομέν νιν Ελλάδα. •

Instead ofdesignating the place, the accusative can also

serve to mark a more or less abstract goal. Such is the
meaning of the expression venum ire, “ to go for sale, to
be sold,” pessum ire (for perversum ire), " to hurt oneself,
to fall," suppetias accurrere, " to rush to the rescue," etc.
We here find, after the rule eo Romam, another text -book
rule : eo lusum , “ I am going to play .” Lusum is the
accusative of a verbal substantive which has been drawn
into the mechanism of the conjugation. The Latin
grammarians, without understanding it, disguised it under
the ludicrous name of " supine." This also is the way in
which arose : conveniunt spectatum ludos, " they come to
sce the games."
We shall call this first use of the accusative the
accusative of direction.

We have so far been confined to the first stage. The

1 Examples are rarer among the prose -writers. Yet we find in
Cicero : Ægyptum profugisse, Africam ire, . . . Rediens Cam
panian . But in general the names of countries are preceded by
a preposition : perhaps we should here assume the part of copyist
or editor, who could easily add an in or ad which seemed to him
necessary.
2 Iliad , i. 317 . 3 Ibid ., iii. 162. • Euripides, Tr ., 883.
324 SEMANTICS

accusative is used in its proper sense and with its original

import.
The second stage is denoted by constructions such
as invenire viam , attingere metam . Here the point of
view changes : the accusative seems to be governed by
the verb. We showed in a preceding chapter by the
example ofpetere and a few others, how the verbs, from
being neuter, became transitive. In this way another
type of accusative has been gradually impressed on the
minds of men : the accusative of the direct object. Lan
guage, with its special Logic, having said cupere divitias,
said also temnere divitias; having said sequi honores, said
also fugere honores. The primordial idea of the accu
sative was bound to be effaced in presence of this diver
sity : a grammatical accusative took the place of the
local accusative.
We have seen further back 2 that this change was but
slowly effected . Thus the Greek verbs which are con
structed with the genitive, such as ακούω, επιθυμώ, τυγχάνω,
witness to a condition of language in which the proper
value of the case is still distinctly felt. It is only through
the lapse of time that there is established in the minds
of men a sort of levelling expressed by the rule : The
active verbs require to be followed by the accusative.
A few scholars, preoccupied with the essence of things
have wished to establish a special category of verbs in

" See p. 189. It must be added that most languages have, through
an instinct of order and clearness, effected a differentiation , allotting
to some the exclusive function of neuter verbs, employing others
exclusively as transitives. · See above, p. 197.
 THE LOGIC OF LANGUAGE 225

which the accusative should mark the result of the

action, as in : Deus creavit mundum , scribo epistulam ,
Themistocles extruxit muros. But these verbs, which are
distinguished from others by their meaning, do not in
any way differ from them in usage ; it is quite as easy
to say : Xerxes evertit muros, mandata neglexit.
The relationship between the accusative of the direct
object and the accusative of direction is no longer felt ;
so that there is no reason why a verb should not take
both at the same time. When, in Homer, the soothsayer
Helenus invites his mother Hecuba to lead the Trojan
women to the sanctuary of Athene,
ξυνάγουσα γεραιάς
Νηον 'Αθηναίης,
these two accusatives in no way hamper each other. It
is the same when Sarpedon, accusing Paris, complains
of the evils which he has caused to the Trojans :
και δη κακά πολλά ιοργαν
Τρώας.
Herodotus, retailing what he had learnt about educa
tion among the Persians, said that they brought up their
children to do three things only : to ride, to draw the
bow , and to tell the truth. Iaidevovoi tous naidas (accu
sative of the direct object) Tpla rouva ( accusative of
direction ), ιππεύειν , τοξεύειν και αληθίζεσθαι. The same
construction is to be found in Latin : Catilina juven
tutem multis modis mala facinora edocebat.
1 The accusative of the direct object is that one of the two which,
the construction being reversed and the verb transposed into the
passive, becomes the subject of the phrase.
Q
226 SEMANTICS

Once in possession of this construction , Language re

verses it, as a mathematician would reverse an algebraic
equation : it transposes it into the passive. Rogatus sen
tentiam , edoctus litteras, id jubeor, didáokouai tùy Lovolanu,
κρύπτομαι τούτο το πράγμα : all of these are constructions
which we should have difficulty in understanding were it
not for the special Logic of which we have spoken.

If we wish to understand thc third usage of the accu

sative, which is to mark duration, we must return to the
original signification. Space and time being, for the
Logic of Language, two quite similar things , it cxprcsses
in the same way up to what epoch an action has been
continued, and up to what place a movement has been
prolonged ; on both hands, the accusative marks direction.
Demosthenes, recalling that the power of the Thebans
had lasted from the battle of Leuctra till recent times,
expressed himself thus : ίσχυσαν δέ τι και Θηβαίοι τους
τελευταίους τουτουσι χρόνους μετά την εν Λεύκτροις μάχην..
To convey that Mithridates was in the twenty -third year
of his reign, Cicero said : Mithridates annum jam tertium
et vicesimum regnat.
Thus was formed what grammarians call the accusative
of duration : Vejorum urbs decem æstates hiemesque con
tinuas circumsessa Flamini Diali noctem unam extra
urbem manere nefas est. We find in Lysias, to indicate
that a man has been dead for three years : τέθνηκε ταύτα
· This can be verified by an examination of adverbs of place, like
hic, ubi, inde, which served equally to express ideas of place and
of time.
 THE LOGIC OF LANGUAGE 227

opla 79. Latin also says no less strangely : Puer decem

annos natus.
It sometimes happened, as was bound to be the case,
that the accusative of duration was confounded with the
accusative of the direct object. When the French say ,
les années qu'il a vécu, it is difficult to know exactly
in what light to consider this construction. The same
fact is to be met with in the ancient languages . There
may be a difference of opinion with regard to some of
these cases, and the hesitations of French orthography
are well known, but with the exception of those particular
cases for which it is difficult to formulate a rule, the
existence of an accusative of duration is beyond doubt :
it forms the third stage of this history,

There only remain to be explained expressions such

as decem pedes latus, or os umerosque deo similis. But
we do not wish to prolong too technical a study ; we
have said enough to show how popular Logic goes to
work.
This Logic, we repeat, rests entirely on Analogy,
Analogy being the method of reasoning of children and
of the multitude. An expression is given ; they form
another somewhat like it. This one in its turn produces
a third, slightly different, which again provokes imita
In Sanscrit : çatam grva caradas, “' may you live a hundred
years ! ". In Greek : Iva uiva mévwv, “ remaining a month.” Th7
aðplov métrovoav el Biboeta (Euripides, Alc., 784) " (no one knows)
whether he will live to- morrow." The ancient languages appear to
range these constructions under the category of the accusative
of the direct object. But French is more preoccupied with the
essence of things, which demands the accusative of duration .
.

228 SEMANTICS

tions, the first and the second not having in the meantime
ceased to be productive. Language can go far in this
way. To one who learns Language by usage this is in
no way surprising, since he does not dream of bringing
together or mutually comparing such different appli
cations, But if any one, finding these applications
enumerated singly in a book, tries to discover some
connecting idea, some first idea common to all, he runs
the risk of being lost in the most shadowy of abstractions.
What we must do is to retrace the road over which
Language has travelled, striving to recognise the turnings,
and never forgetting that, Language being the work of
the people, we must, to understand it, put off the logician
and become one with the people ourselves.
 CHAPTER XXV

THE SUBJECTIVE ELEMENT

What we are to understand by the Subjective Element - Its
connection with speech - The Subjective Element is the most
ancient part of Language.
If it be true, as has been sometimes maintained, that
Language is a drama in which words figure as actors,
and in which the grammatical ordering reproduces the
movements of the characters, we must at least correct
this comparison by the addition of a special circum
stance : the impresario frequently intervenes in the
action , for the purpose of contributing his reflections
and his personal feeling, not after the fashion of Hamlet,
who, though interrupting his comedians, remains a
stranger to the piece, but as we do ourselves in a dream ,
when we are at the same time the interested spectator
and the author of the events. This intervention is what
I propose to call the Subjective side of Language.
This Subjective side is represented : ( 1 ) by words or
paragraphs ; (2) by grammatical forms; (3) by the general
plan of our languages.
229
230 SEMANTICS

I take as example a miscellaneous fact of the most

ordinary kind. " An accident took place yesterday on
the line from Paris to Havre, which interrupted traffic
for three hours, but happily caused no loss of life." It
is clear that the word printed in italics does not refer
to the accident, but that it expresses the sentiment
of the narrator. We receive, however, no shock from
this confusion, because it is absolutely in conformity
with the nature of Language.
Many adverbs, adjectives, or sentences which we
interpolate in the same manner, are reflections or appre
ciations of the narrator. I will cite in the first instance
the expressions which mark the greater or less degree of
certainty or of confidencc felt by the spcaker, such as no
doubt, perhaps, probably, assuredly, ctc. All languages
possess a store of advcrbs of this kind : the farther we
go back into the past, the more of them do we find.
Greek is fully provided with them : I need only recall
the variety of particles with which the prose of Plato
is sown, and which serve to qualify the impressions or
the intentions of the speakers. They may be compared
to gestures made in passing, or to glances of intelligence
cast in the direction of the audience.
A genuine logical analysis should, to justify its name,
carefully distinguish between these two elements. If
when speaking of a traveller I say, " He has at this very
moment no doubt arrived," no doubt docs not refer to the
traveller, but to me. Logical analysis, as practised in
schools, has been at times cmbarrasscd by this Subjective
1 "Ή, μήν, τοί, που, ίσως, δή, τάχα, σχίδον, άρα, νύν, etc.
 THE SUBJECTIVE ELEMENT 331

Element: it has not realised that all writing of at all a

vivid character may take the form of a dialogue with the
reader. Such also is the effect produced by certain
pronouns when cast into the middle of a story , in which
the audience seems suddenly to be made by the narrator
a party to the proceedings. La Fontaine was partial to
these :
"Il vous prend sa cognée : il vous tranche la tête. "
They have been called " expletives," and as a matter
of fact do not form part of the narrative ; neverthe
less they correspond to the primary intention of the
narrative.
For want of having taken this Subjective Element into
consideration, certain words of the ancient languages
have been misunderstood. One of the greatest of our
contemporary philologists, treating of the Latin adverb
oppido, refuses to believe that it is the ablative of an
adjective signifying " solid, firm , sure. " ? He asks how
this meaning can be reconciled with phrases like gido
interiä, oppido occidimus. But this is because we must
postulate the part of the Subjective Element We say
in the same way : I am crtainly lost ; or in German :
Ich bin sicherlich verloren , idioms in which were we
to adhere rigidly to the text, there would be a sort of
contradiction in terms.
The same thing took place also in the case of the
German adverb fast, which now signifies * a'ss bei
which formerly denoted an idea of fixity or of cortisode
It was usual to say : vaste ruafcn , " to call lacr *
* Cl. the Greek latebas, sosa *
232 SEMANTICS

swiveln , "to doubt strongly .” “ I have prayed for him

long and earnestly." " Ich habe lange und fast fiir ihn
gebeten " ( Luther). If it is now taken in the sense of
“ almost," it is because it represents a phrase like ich
glaube fast, ich sage fast, " I believe firmly." The same
thing happened to ungefähr, which acquires its true
signification if completed : “ without fear of mistake."
So in Latin pæne, ferme, mean “ ncarly," although the
first is closely related to penitus, and the second is a
double of firme; but we must re-establish the complete
idioms : pæne opinor, firme credam ."
The web of Language is being continually embroidered
with these words. If I chance to formulate a syllogism ,
the conjunctions which mark the different clauses of my
argument relate to the Subjective Element. They appeal
to the understanding, they call it to testify to the truth
and concatenation of facts. They are not therefore of
the same order as the words which I use to expound the
actual facts themselves.

But Language does not stop there. The mingling of

the two elements is so intimate that an important part
of grammar derives its origin therefrom .
It is in the verb that this mingling is most visible.
My readers will guess that I am speaking of moods.
The Greek grammarians understood this thoroughly :
they said that moods served to mark dispositions of the
soul, OlaOégels Yuxas. As a matter of fact, an idiom
like Deol dołev contains two very distinct things: the idea
i On pæne see Mém . de la Soc. de Ling., V. p. 433.
 THE SUBJECTIVE ELEMENT 233

of succour lent by the gods, and the idea of a desire

expressed by the speaker. These two ideas have in
a way entered reciprocally into each other, since the same
word which marks the action of the gods marks also
the desire of the speaker. The simple word in Homer,
re / valns, “ utinam moriaris ! " besides expressing the idea
of dying, expresses also the wish of him who gives vent
to this imprecation. In this lies no doubt the original
meaning of the optative.
But the optative is not the only mood of this kind.
The subjunctive mingles also with the idea of action an
element drawn from the διαθέσεις ψυχής. . It is true that
it approaches very near to the meaning of the optative.
According to the most recent investigations, it would
seem that, in the Vedas, the optative was the mood
preferred for certain verbs, and the subjunctive for others,
although there was no very decided shade of meaning to
distinguish them . This abundance of forms shows how
important a place was attributed by Language to the
Subjective Element. The languages which, like Greek,
preserved both moods, attempted to differentiate them .
But the generality of languages, somewhat encumbered
by this excess of riches, have fused together the optative
and the subjunctive.

The Latin future is so near the subjunctive and the

optative, that in certain persons it is confused with them .
Inveniam , experiar are, ad libitum , either futures or
1 Delbrück, Altindische Syntar, $ 172 ; Whitney, Indische Gram
matik, $ 572.
234 SEMANTICS

subjunctives. There is in this a true insight into the

nature of things. To announce what will happen , is in
reality in most human affairs merely an expression of
our wishes or of our doubts. We canunderstand that
in carly days these shades of meaning used to be con
founded. Examples abound which show that there was
no precise limit between the future and the subjunctive.
Thus the difference between tenses and moods is obliter
ated in the eyes of the historian of Language . Those
who, in our own days, have started the extraordinary
idea that the optative was invented to be the mood of
the unreal (de l'irréel, der Nichtwirklichkeit ), credited the
ancients with the same power of conception that one
admires in the creators of Algebra. But in those distant
ages Language had less exalted aspirations and more
practical aims.

The Subjective Element is not absent from the gram

mar of our modern languages.
French, to express a wish, uses the subjunctive : Dieu
vous entende !-puissiez -vous réussir ! Certain logicians,
to justify the use of the subjunctive, have imagined an
ellipse : Je désire que Dieu vous entende. - Je souhaite
que vous puissiez réussir. In reality, so little has
French given up this Subjunctive Element, that it has
provided itself with new forms to express it. If it
wishes to indicate an action with a reservation of doubt,
1 Ουκ ίσσεται, ουδέ γένηται. -00 πω ίδoν, ουδέ ίδωμαι. - Ει δέ κε μη
Ouwow, dow Bé Kev atrds črwual, etc. Cf. Tobler, Uebergang zwischen
Tempus und Modus, in the Zeitschrift für Völkerpsychologie, II .
p. 32. See also Mém . de la Soc. de Ling., VI. 409.
 THE SUBJECTIVE ELEMENT 235

it has turns of phrase like the following : Vous series

d'avis que ... Nous serions donc amenés à cette con
clusion. In these phrases it is not a condition that
.

the verb expresses, but a fact which is considered as

uncertain. The conditional has therefore inherited from
some of them the most delicate usages of the subjunctive
and optative.
Indirect speech, with its varied and complicated rules,
is like a transposition of the action into another key.
What written language obtains now -a -days by means of
inverted commas, used to be indicated in spoken language
by the different formations of the verb. The subjunctive
and optative there found their natural place, since a
certain element of doubt was necessarily spread over
the whole discourse.

We have still to speak of the mood in which the

Subjective Element is to be seen in greatest strength :
the imperative mood. What characterises the imperative,
is that to the idea of action it unites the idea of the will
of the speaker. It is true that in most forms of the impera
tive it would be vain to seek for syllables which specially
indicate this will. It is the tone of voice, the expression
of countenance, the attitude of body which are charged
to convey it. One must not forget the existence of these
elements, which, though not indicated in writing, are
none the less an essential part of Language. The im
perative shares certain of its forms with the indicative :
yet there is no reason to regard these as borrowed from
the indicative. I am inclined to think, on the contrary,
236 SEMANTICS

that the imperative is the first in date, and that, inversely

to what we are taught, when there is identity , it is the
indicative which is the borrower. Perhaps those brief
forms, such as tol, " come ! " dós, "give," otíte, “stop ! "
are older than any other part of the conjugation.
We have made an allusion to the duality of human
personality. There is in the Sanscrit and Zend conjuga
tion a grammatical form in which this duality appears
openly ; I mean the first person singular of the impera
tive, like bravāni, “ let me invoke,” stavāni, “ let me
celebrate ." However strange may scem to us a forın of
command in which the speaker gives orders to himself,
there is nothing in it that is not in conformity with the
nature of Language. This first person says briefly what
is expressed in other languages in a more or less round
about way. French employs the plural. The shepherds
of Virgil call upon one another in the second person :
“ Insere nunc, Mclibæe, piros; pone ordine vites / "
We can now understand why it has always been so
difficult to give a true and complete definition of the verb.
The ancients succeeded best in this. The moderns, in
defining the verb as “ a word which expresses a state
or an action ," leave out a great part of its contents,-the
larger and more characteristic part.
1 It has been wondered whether this first person in ni is ancient,
or a comparatively recent acquisition. Its presence in Zend, in
which it has, in the middle voice, a corresponding form in nē, leads
one to believe that it is ancient. It should be an archaic fragment,
which, being no longer attached to anything, disappeared later on
from almost every usage.
 THE SUBJECTIVE -ELEMENT 237

If from the moods and tenses we pass to the per

sons of the verb, we find facts of a yet more striking
character.
Man is so far from considering the world from the stand
point of disinterested observer, that we find, on the
contrary , that the part which he has appropriated to him
self in Language is a wholly disproportionate one. Out
of three persons of the verb, there is one which he keeps
absolutely to himself (that which is by agreement called
the first). In this way he already places himself in
opposition to the universe. As to the second, it does
not take us very far from ourselves, since the whole
existence of the second person depends on its being
addressed by the first. It may therefore be said, that
the third person alone represents the objective portion
of Language.
Here also it is permissible to suppose that the Subject
ive Element is the oldest. Philologists who have tried
to analyse verbal inflections should have an inkling of
this : while the third person is easily explainable, the
first and second persons are those which oppose the
greatest difficulties to etymological analysis.
An analogous observation may be made with regard
to pronouns. The pronoun me was not sufficient : a
special pronoun was also needed to indicate that the me
takes part in a collective action. This is the meaning of
the pronoun us, which means I and they, I and you, etc.
But that is not yet enough : in many languages a special
number has been needed to indicate that the me counts
for half in an action in which two people take part.
238 SEMANTICS

This is the origin and real cause of the dual number

in the conjugation.
We begin to see from what point of view man ordered
his language. Speech was not made for purposes of
description, of narration, of disinterested considerations.
To express a desire, to intimate an order, to denote a
taking possession of persons or of things — these were
the first uses of Language. For many men they are
still practically the only ones. If we descended one or
more stages, and sought for the beginnings of human
language in the language of animals, we should find that
among these the Subjective Element reigns supreme, that
it alone obtains expression or comprehension, that it
exhausts their faculty of understanding and the entire
subject -matter of their thoughts.
We are not therefore concerned with an accessory ,
a sort of excrescence, but, on the contrary, with an
essential part, and with what is no doubt the primordial
foundation to which the remainder has been successively
added .
 CHAPTER XXVI

LANGUAGE THE EDUCATOR OF THE HUMAN RACE

Part played by Language in the operations of the mind

Wherein lies the superiority of the Indo -European languages
The place which the Science of Language should hold among
sciences.
THERE is no reason to fear that the importance of
Language in the education of the human race will ever
be depreciated. We can in this matter trust to the
natural instinct of mothers ; their first impulse is to talk
to the child, their first joy to hear it speak. Then come
masters of all degrees and of all kinds, whose various
arts each presuppose Language, even if not absolutely
one with it. In every country, in antiquity as in our
own days, in China and in India as in Athens and in
Rome, Language supplies both the instrument and the
matter of the first lessons.
This universal agreement has its reasons ; it is not
difficult to understand the great influence which is exer
cised over the mind by Language, if we reflect that we
none of us receive it whole and intact, but are each
obliged to build it afresh for ourselves. Here is an
239
240 SEMANTICS

apprenticeship which, though escaping notice and ignored

even by those who take part in it, forms nevertheless a
kind of training -school of humanity. If it be true that
the best instruction is that which gives us the most to
do by ourselves, what more profitable study can be
imagined for a child ?
The greatest attention is needed for the mere recogni
tion of the word, since it is a question of disengaging
it from what precedes and from what follows, of dis
tinguishing the permanent element from the variable
elements, and of understanding that the permanent
element is, to some degree, confided to us, in order that
we should in our turn wield it, and submit it to the
same variations. On what occasions, under what cir
cumstances, after what models ? For the most part we
are told nothing : on us lies the burden of discovery.
The most simple phrase invites us to analyse its thought
and to note the special contributions of each word. The
adjective, the verb, are the first abstractions understood
by the child. The pronouns me and you, my and your,
which, in changing mouths, transfer themselves from one
to the other, contain its first lesson in psychology,
Teaching mounts a step higher in proportion as
mankind advances in this apprenticeship.
Let us imagine the effort which must have been
required in their own days to speak the ancient
languages with even moderate accuracy. It was neces
sary in the case of the various declensions to establish
different series, in which certain inflections corresponded
to without resembling each other, and in which other
LANGUAGE, EDUCATOR OF THE HUMAN RACE 241
inflections were similar, but had to be kept apart. An
analogous classification was necessary for the persons,
tenses, and moods .? Herein lies a whole chapter of
inner life which began afresh with every individual. The
people was therefore possessed of an unwritten grammar,
into which no doubt various mistakes and errors crept,
but which, since these languages have for centuries
been transmitted from generation to generation, must
nevertheless have been endowed with a certain fixity.
When we consider the trouble which these ancient
languages cost us even now, we are somewhat sur
prised .But we must remember that the teaching
of a mother tongue has the advantage of being con
tinued at all times and in all places, that it is spurred
by the stimulus of necessity, that it has to deal with
minds in the freshness of youth, and finally, that it
possesses the unique characteristic of linking words to
things, and not the words of one language to the words
of another language. The same circumstances occur
in the case of all mother - tongues ; everywhere the
mind of the child triumphs over them. I do not mean
indeed that the course of time may not produce such
difficulties as to disconcert the new generations. But
then, as we have seen the popular mind extricates itself
from these in the most simple manner ; it evades the
difficulty by means of analogy, of unification , and of
suppression. The people being in this matter both
· H. Paul, Principien der Sprachgeschichte, 2nd edit., p. 24. See
also the studies of Steinthal and Lazarus in their journal.
* See above, Chapters I., VI. and VIII .
R
 242 SEMANTICS

pupil and master, whatever it changes, unifies, or

V abrogates, becomes the rule of the future.
Our modern languages, though less encumbered with
a formal apparatus, are yet by no means free. The
complication has, morcover, been diverted on to another
point. The question is to know how to employ words
which are almost wholly devoid of meaning, words so
abstract and " scrvile," that one may all one's life ignore
their existence, while duly placing them in their proper
position. Here we may notice an action of the mind
which has passed to a condition of instinct, similar to
that which guides the fingers of the lạcc-worker when
she moves her distaff without looking at it.
Were all the usages of our prepositions to be enumer
ated and explained, they would form a whole volume.
Littré's dictionary gives no less than twelve columns 1
to the word à alonc. Yet the people finds its way
with no difficulty in this seeming chaos. Not, as we
have sccn, owing to a more or less clear notion of the
import of the word : no more than the philologists,
would it be able to give a definition that fitted every
usage. It allows itself to be directed by a certain
number of expressions which the memory retains and
uses as models. In this way the turns of phrase are
preserved and propagated. Invention always builds on
a foundation which cxists already.

1 “ By an evil chance, the alphabetical order forced me at the

very outset of my labours to treat of the preposition d, that most
tiresome of all words, and one with which I did not succeed to my
satisfaction.”—Littré, Comment j'ai fait mon Dictionnaire.
LANGUAGE, EDUCATOR OF THE HUMAN RACE 243
Every one has at times admired the unpremeditated
turns of phrase in popular language. Besides the
pleasure which is felt at any discovery, these encounters
have the added advantage of showing us the roads by
which the mind has travelled. It is especially on
occasions when some passion warms the soul and
intensifies its power, that we can best observe these
improvisations of the moment.
The human intellect obtains from Language, for its
daily operations, the same services that it obtains from
numbers for its calculations. One conscquence of the
infirmity of our understanding, an infirmity familiar to
all philosophers, is that it is easier for us to work upon
the signs of the ideas than upon the ideas themselves .?
Before the invention of writing men counted by means
of pebbles. The idea certainly had to come first, but
this idea is vacillating, fugitive, difficult to transmit;
once incorporated in a sign, we are more sure of possess
ing it, of handling it at will, and of communicating it to
others. Such is the scrvicc rendered by Language ; it
makes thought objective.
Having been from the very first associated with the
conception, it was not long before the words took the
· In this lies the answer to the question why the intelligence of
animals remains stationary. They have not arrived at the point
of voluntarily incorporating their thought in a sign : their whole
ulterior development has remained thenceforward arrested in its
very first steps. The idiot child does not speak , but it is not the
organs of speech which are lacking to it. The inward labour of
observation and of classification which allows the idea to be affixed
to the sign is beyond its strength .
244 SEMANTICS

place of the conception ; we compare, we connect, we

oppose the signs, not the ideas. It is true that behind
these signs there survives a partial memory, a fragmentary
recollection of the idea which it represents, and we in
wardly believe that we could, if we would, recall the idea
in its pristine clearness." But it is no less true that in
the case of rather complicated operations, of operations
to be effected rapidly, the signs are enough for us. Not
words alone, but also those unions of words which we
have called Articulated Groups , are necessary. All
this gocs to form Language. It at the same time
renders the idcas inanagcable, and furnishes the frame
work of reasoning.
Thinkers have turned this into an occasion for re
proach. “ Every word represents indeed a portion of
actuality , but a portion roughly hewn , as though
humanity had carved according to its own convenience
and its own wants, instead of following the articulations
of the reality .” Let us suppose for a moment that this
reproach is well -founded. Even then how slight it is in
comparison with the immense service rendered to the
mass of mankind ! However imperfect, Language yet
outstrips most of us : we need time to overtake it. How
few would be capable of setting to work on their own
account to produce these carvings ! We have seen,
moreover, that the outlines are not so fixed and unyield
ing but that they can be bent or enlarged for insertion
in new classifications. On the contrary, a philosophical
1 Taine, De l'Intelligence, bk. i. chap. iii.
* See above, p. 166. 3 Bergson.
LANGUAGE, EDUCATOR OF THE HUMAN RACE 245

language such as has often been planned, a language

resulting from a system, in which each word should
romain for cver delimited by its definition, and in which
the affinity of words should be traced from the real or
imagined connection of idcas,—such a language may
indeed be suitable for a few special sciences, like chem
istry, but if applied to human thought, with its fluctua
tions and its progressions, could not fail to becomc, in
the course of time, a hindrance and a restraint. In
proportion as the experience of the human racc increases,
Language, thanks to its clasticity, acquires new mcaning.

Were it necessary to say wherein lies the superiority

of the Indo -European languages, I should turn neither
to the grammatical mechanism , nor to the compounds
nor even to the syntax : I believe it to lie in another
direction . It consists of the facility of these languages
from the most ancient periods of which we have any
knowledge, in creating abstract nouns. If we cxamine +
the suffixes which serve for this usage, we shall be
astonished at their number and varicty. They are by
no means peculiar to such and such a language, but are
to be found alike in Latin, Greck, Sanscrit, Zend, and
in every branch of the family. They are therefore of
early date ; and indeed of a date so early, that, to
borrow the denominations of another science which
marks the epochs by the monuments which are left of
it, we might speak of a period of suffixes, a period which
necessarily presupposes a certain power of abstraction
and reflection. It is the presence of these nouns in great
246 SEMANTICS

number, added to the possibility of making others after

the same pattern , which has rendered the Indo -European
languages so well fitted to all the operations of the
mind. Even at the present day we still make use of
the same means of expression, to which hardly anything
has been added by posterior ages. If we cared to
cxaminc thc mcthods used by modern literaturc to
renew the resources and the character of its style, wc
should be convinced that it still has recourse to thosc
same abstractions, the first specimens of which were
contemporaneous with the Vedas and with Homer.
It is not on this account necessary to imagine trans
cendent intelligences. There are divers degrees of
abstraction to be distinguished. The onc in question
is derived more from mythology than from mctaphysics,
Thc samc order of thought is there when the people
talks of a reigning fashion or disease,' or of electricity
running along a wire. The abstractions created by the
popular mind assume for it a sort of existence. The
world is full of thcsc entities. The form of phrase, in
which all subjects are represented as active, is a still
surviving witness of this state of mind. Language and
mythology issued from onc and thc samc conception.
This, as we have already said, is the explanation of
the fact that the majority of abstract nouns are feminine ;
' It is easy to imagine how useful these suffixes have been for
the language of philosophy. Greek , in combining the two pronouns
Toobs and noios with an abstract suffix, makes roobons, “ quantity,"
Toidons, “ quality." In thc same way, in Latin, qualitas, quantitas.
In Sanscrit, the pronoun tat, “ this," gives, in combination with the
abstract suffix tvam , the substantive tattvam , “ reality.”
LANGUAGE, EDUCATOR OF THE HUMAN RACE 247

they are of the same sex as the innumerable divinities

which peopled sky, earth , and water. Even at the
present day — so great is the continuity of things
those who discuss Matter, Force, Substance, perpetuate
more or less this ancient condition of mind.

Accustomed as we are to Language,we do not casily 7

realise the accumulation of intellectual work which it
represents. But, to be convinced of this, we have only
to take a page of any book , and to suppress all the
words which, not corresponding to any objective reality,
must be the result of an operation of the mind. Of a
page so treated hardly anything will remain . The
peasant who speaks of times and of scasons, the
merchant who advertises his assortment of warcs, the
child who brings his notes of conduct or progress, all
move in a world of abstractions. The words number,
form , distance, situation , are so many concepts of the
mind. Language is a translation of reality, a transposi
tion in which objects appear already generalised and
classified by the work of thought.

Are there in Europe any languages which are more

favourable than others to intellectual progress ? With
the exception of slight differences we can answer in the
negative. They are all (or nearly all) sprung from the
same origin , built on the same plan, drawing from the
same sources. They have been more or less nourished
on the same models, perfected by the same education.
They are therefore capable of expressing the same
 248 SEMANTICS

things, although within the limits of this close relation

ship we can already note certain special aptitudes. But
if we wished to realise the help which Language gives to
the mind, as well as the particular turn which it imposes,
we should have to compare some language of Central
Africa or some indigenous dialect of America. In
Brazilian, the one word tuba signifies: ( 1 ) he has a
father; (2) his father ; (3) he is a father. Actually, tuba
mcans “ hc the father." It is as the speech of a child.
Even languages provided with a rich literature do not
always form a sufficient support for the mind. In
Chinese, this phrase : šin hi thien may be translated :
( 1 ) the saint aspires to hcaven ; (2) he is a saint to aspire
to heaven ; ( 3 ) he is a saint, who aspires to heaven.
Chinesc merely says : Saint aspire heaven .' The service
rendered us by our languages is to impose upon us a
form that compels precision.

tLanguage
Language has been called an organism , a hollow ,
deceptive word too frecly lavished at the present day,
and used every time that we want to dispense with
the troublc of sccking for true causes. Since certain
illustrious philologists have declared that man counted
for nothing in the evolution of Language, that he was
incapable of modifying anything, or of adding anything,
and that one might as well try to change the laws of
the circulation of the blood; since others have compared
this cvolution to the trajectory of a shell or to the orbit
of a planct ; since this is to-day currently accepted as a
Misteli , in Techmer's Zeitschrift, vol . ii.
 LANGUAGE, EDUCATOR OF THE HUMAN RACE 249
truth and passed on from book to book : it has seemed
to me useful to have it out with these assertions, and
once for all to make an end of this phantasmagoria.
Our forefathers of the school of Condillac, those
ideologists who for fifty years served as target to a
certain school of criticism, were less far from the truth
when they said, in simple and honest fashion, that words ģa
arc signs. Where they went wrong was when they
referred everything to a reasoning reason, and when
they took Latin for the type of all Language. Words
are signs : they have no more existence than the signals
of the semaphore, or than the dots and dashes of Morse
telegraphy. To say that Language is an organism
is but to darkcn counsel and to sow a sced of error in
the minds of men. It might be said, with an equal
degree of truth, that writing is also an organism, since
we see it evolving throughout the ages, without any
one in particular having a very perceptible influence
on its development. It might be said that song, religion,
law, all the component parts of human life, are cach an
organism.
If we take nature in its widest sense, it evidently
comprises man and the productions of man. The
history of morals, of customs, of habitation, of dress,
of the arts, as well as social and political history, will,, y
together with Language, form part of natural history.
But if we admit a difference between the historical
sciences and the natural sciences, if we consider man as
furnishing the material for a separate chapter in our study
of the universe, Language, which is the work of man ,
250 SEMANTICS

cannot remain on the other side, and the Science of

Language, by a necessary consequence, will form part of
the historical sciences. If, on the other hand, on account
of phonetics, which study the sounds of language pro
duced by the larynx and the mouth, it were necessary
to transfer the Science of Language back to the natural
sciences, all the rest, man and all his works, must in
evitably accompany it, since human productions, of
whatever kind, come after all from the organs of man
kind and are directed to thosc organs.
With still more reason will Semantics belong to the
class of historical investigations. There is not a single
change of mcaning, a single modification of grammar, a
single peculiarity of syntax which should not be counted
as a small historical event. Will it be said that there is
no free will in this domain, because I am not free to
change the meaning of words, nor to construct a phrase
according to a grammar of my own ? We have shown
that this limitation of freedom arises from the desire
for comprehension, that is to say, it is of the same kind
as the other laws which govern our social life. To
speak here of natural law is to create confusion .
I have come to the end of my labours. Warned

by example, I have avoided comparisons drawn from

botany, from physiology, from gcology, with the same
zeal which others have shown in seeking them out. The
cxposition of my argument is in consequence more
abstract but, I think I am justified in saying, more true.
I have no wish to be unjust towards the theory which,
with a certain show of brilliancy, has classed the Science
LANGUAGE, EDUCATOR OF THE HUMAN RACE 251
of Language among the sciences of nature. At a time
when these sciences rightly enjoyed the favour of the
public, it was a clever and politic act. . It also imposed
on philologists the duty of prosecuting their observa
tions with a redoubled accuracy. Finally this idea con
tained just the amount of paradox necessary to excite
curiosity. If the phrase had been : regular development,
constant progress, nobody would have cared. But at
the words : blind laws, astronomical precision , general
attention was aroused.
I think, however, that I am right in saying that the
history of Language, when referred to intellectual laws,
is not only more true, but also more interesting : it
cannot be a matter of indifference to us to note, above
the sceming chance which governs the destiny of the
words and forms of Language, the appearance of laws
corresponding in each case to an advance of the mind.
For the philosopher, for the historian, for every man
who watches the progress of humanity, there is a certain
pleasure in verifying this ascent of the intelligence which
is perceptible in the slow renewal of languages.
*
END OF SEMANTICS.
 WHAT IS MEANT BY PURITY OF
LANGUAGE ?

#
 WHAT IS MEANT BY PURITY OF
LANGUAGE ??

THERE appeared a few years ago, under some such

title as the above, a work by the Swedish professor, Mr.
Adolphe Noreen , which instantly attracted attention by
the independence of its views. When translated into
German, it gave rise to argument and discussion, the
inevitable fate of writings which forsake the ordinary
path. We propose in our turn to pass judgment upon
it ; but we are glad to be able to say beforehand, that on
the most essential points we are in agreement with the
author.
Mr. Noreen is professor of Scandinavian Philology at
. the University of Upsala. Familiar with all the methods
and all the results of modern philology, his long -estab
lished reputation as a scholar can only add to the autho
rity of his opinions and of his decisions. These opinions
we will now recapitulate for our readers, but without
feeling bound to adhere strictly to the original, and
1 A. Noreen, Om Sprakriktighet, 2nd edition. Upsala : W.
Schultz, 1888. A German translation, by Arwid Johannson, was
published in the Indogermanische Forschungen, v. I.
255
256 SEMANTICS

reserving to ourselves the right to substitute for his

examples other examples drawn from our own history.
Let us begin by saying that there must be an element
of truth in this idea of purity, since so many minds, in
ancicnt as in modern days, have been interested in the
subject. But it is not casy to justify in the eyes of
reason what fceling would dictate to us on this ques
tion. No sooner is an attempt mąde to formulate
principles than there arises a difference of opinion, and a
consequent beginning of uncertainty. Artists and poets
speak of it by the light of instinct alone ; philologists,
wishing to illuminate it, introduce at the same time their
own systems. Let us see if it be possible, by discarding
all foregonc conclusions, to attain to some degree of
clearness.

The first point to examine concerns words of foreign

cxtraction .

The road is harassed by many prejudices. The chief

of all these, or, to speak like Bacon, the chief " idol," the
one from which all the others are derived , is the deter
mination to consider purity of Language as something
akin to purity of race. For all who look at matters in
this light, the introduction of a foreign word is a con
tamination ; an English or German word introduced into
French is a blot on the national language. It is not in
France that this point of view is most frequently to be
met with. The Germans have for a century past raised
barrier upon barrier to stop the immigration of French
words. It would be impossible to give the number of
 WHAT IS PURITY OF LANGUAGE ? 257

manifestoes hurled, since the days of Adelung, against

foreign words , or the number of societies which have
proposed to repel the invasion . Do foreign words merit
such a degree of animadversion ? Are there not dis
tinctions to be made, a modus vivendi to be adopted ?
Are all foreign words to be cqually condemned ?

When an art, a science, a fashion, a game, comes to us

from abroad, it usually brings in its train its own voca
bulary. It takes less time to appropriate terms than to
invent them purposely for the expression of ideas or
objects which have already been named. A certain kind
of music having come to us in the seventeenth century
from Italy, our musical language is full of Italian words.
Which of us, in speaking of an adagio, of a sonata, re
members the exotic origin of these appellations ? Rabid
lovers of purity should not forget that the like has taken
place in every epoch ; and,since they invoke the classical
tradition, we may remind them that in this matter the
ancients did exactly the same thing. As the Romans
received their writing from thc Greeks, cverything which
refers to the art of writing is Greek, beginning with
scribere and litteræ . And not these alone. Whether it is
a question of science, of law, of ritual, of military tactics,
of weights and measures, of constructions, of clothing, or
of works of art, we find throughout the Latin ianguage
traces of Greek and of Greek names. If we could go
1 One of the last of this kind is that of Professor Herman Riegel :
Ein Hauptstück von unserer Muttersprache. Mahnruf an alle
national gesinnten Deutschen, 1884 . S
258 SEMANTICS

farther back still, we should no doubt find that many

technical terms which we have always believed to be
Greek were born far from the soil of Hellas. They
would lcad us to Egypt and to Chaldæa. Borrowings
therefore arc a characteristic of cyery epoch. They are
as old as civilisation. Objects useful in daily life, the
instruments of sciences and of arts, as well as the abstract
conceptions which enhance the dignity of man, are not
invented twice over, but are propagated from one people
to another, to become the common property of humanity.
It appears, therefore, legitimate to preserve their name.
Words being, in their way, historical documents, it would
seem hardly to the point to wish of deliberate purpose
to suppress the witness that they bear.
The defenders of purity are not entirely blind to these
considerations. But they suggest that—if borrowing be
an absolute necessity-nations should have recourse
rather to a sister language ; that French , for instance,
should turn to Italian or to Spanish, while English should
borrow from Danish or from Dutch. One would more
readily admit these congeneric words, just as it is
Leibnitz who speaks) one more willingly gives admission
to foreigners who in their customs and manner of living
come near to our own habits. The advice is excellent,

but not always easy to follow , since if we can only take

the necessaries of life where we find them, so also we
can only take the words from those to whom they
belong. Many terms of parliamentary life are English,
because England supplied the first model of the con
stitutional system. On the other hand, if the English
 WHAT IS PURITY OF LANGUAGE ? 259

language uses French words to designate a variety of

things which relate to the elegancies of life, it is because
the things themselves have come from France.
At least, we are told, words must be modificd so that
they become unrecognisable, and that the transaction
does not strike the eye. In this respect, we could in
former days safely trust to popular usage : the foreigner
was promptly clothed in a dress which prevented it from
attracting notice. But now matters have somewhat
changed. The majority of the loans are effected, not
through the medium of conversation, but chiefly by the
intermediary of written language : foreign words are
brought before our eyes in newspapers or books before
becoming familiar to our ears. It is therefore more
difficult for important modifications to take place. There
is, moreover, something repugnant to our modern ideas in
an arbitrary alteration : when the French are willing to
take back the names of their ancient heroes of the Round
Table in the form under which the pronunciation of
their neighbours has been pleased to disguise them,
how could they at the same time dream of deliberately
naming the inventions and ideas which are really new to
them ?
In the case of scientific terms there is a special interest
in preserving them in the form in which they have first
I appeared. To translate words like telephone, phonograph, -
under pretence of purity, is to hamper a work which has
a value quite equal to that of the homogeneity of a
language ; I mean the facility of relations in the Euro
pean community. Would it be worth while to have
260 SEMANTICS

obtained the unification of time, or the uniformity of

tariffs, if, after having lowered the material barriers, we
raised a wall for the mind ? I have before my eyes a
Latin grammar published in Germany, in which the
author has devoted himself to replacing by German
words all technical terms, such as declension, conjuga
tion, indicative, subjunctive, time-hallowed terms which
have been accepted throughout the entire world for
the last ten or twelve centuries. Thus the indicative
becomes die Wirklichkeitsform , the active voicc die
Thätigkeitsart. Were it even a grammar of the German
language that was in question ! But since it concerns
Latin grammar, why boggle at Latin words ? The
ancient words have even the advantage of having become
purely conventional terms : to translate ablative into
der Woherfall, merely makes it more difficult for the
child to understand the use of the ablative with in, in
which case it is really a Wofall.
Men do not belong only to an ethnic or national
group : they also, according to their studies, their pro
fession, their manner of life and degree of culture, form
part of ideal communities which are at once more general
and more limited. The mathematician lives in a per
petual interchange of ideas with the mathematicians of
other countries. The geologist needs to cominunicate
with his colleagues in Germany or America. The
merchant wishes to know what is happening on the
market of the whole world. It would be unreasonable
to oppose obstacles, in the name of an idea of purity, to
the use of terms which are the common property of men
 WHAT IS PURITY OF LANGUAGE ? 261

devoted to the same interests and to the same pursuits.

In this matter youth teaches us a lesson which has not
yet been fully understood. On the pretext that certain
games which have gone over to France from England
used formerly to be played in France, it has been pro
posed to substitute for the English words the old
names by which thcy were originally known. But this
consideration does not seem to have had much weight
with the lovers of foot-ball or of lawn- tennis. These
thought, not without reason , that, in order to be on the
samc footing as thcir British rivals, to be abreast of the
progress of their sport, to communicate with the masters
of their particular branch, and, in case of need, enter
into competition with them, it was better to know and
be able to handle their language rather than the lan
guage of ancestors, who, though undoubtedly fine fellows,
were never to be met with on the field .
The adoption of foreign words to designate ideas or
objects which have come from outside, and which give
rise to an international exchange of relations, is not
therefore blameable in itself, and can well be justified.
In such a case, it is only to be hoped that the borrowing
will be done intelligently, and that, in the passage from
one nation to another, there will not be substitution of
any kind . This happens more often than one would
think : reft from its proper surroundings, the borrowed
word runs the risk of every kind of deformation and
mistake. It was in this way that the French contredanse
(quadrille) became in English country- dance, and that
renégat became runagate. Probably a vague recollection
262 SEMANTICS

of to run away helped this strange transformation . In

the common speech of Holland a rhetorician is called
rederijker, "rich in speech.”
As happens in the case of all emigrants, the borrowed
words are withdrawn from the stream of thought of the
mother country. They take no part in the changes
which may, in the land of their birth, modify the term
which they represent ; so that when, sooner or later, the
copy is again confronted with the model, there is no
longer any resemblance to be found. The French loyal
and the English loyal no longer express the same
sentiment.
English has at all times readily imported new terms.
It has thereby gained in doubling its vocabulary,
having for many of its ideas two expressions, the one
Saxon, the other Latin or French , It can say at will
either kindred or family ; an event can be called either
lucky or fortunate. One would have to be greatly
infatuated with " purity " to despise this increase of
riches ; for it is impossible but that among these
synonyms there should arise differences which will form
so many new resources for thought. But it is clear that
these mixtures are products of history, not deliberate
and premeditated acquisitions.
When one investigates the repulsion inspired by foreign
words in some even first-rate minds, one finds that it
arises from associations of ideas, from historical recol
lections, from political views, in which, truth to tell, the
Science of Language is interested to but a slight extent.
To German purists, the presence of French words recalls
 WHAT IS PURITY OF LANGUAGE ? 263

a period of imitation which they would willingly efface

from their history. The Hellenic philologists who
banish Turkish words from the vocabulary are continu
ing after their own fashion the War of Independence.
The Czechs, who go so far as to wish to translate the
German proper names, in order not to leave one trace
of a language under which they had so long suffered ,
join to their work of expurgation the hope of a coming
autonomy. “ Purity ” in such a case serves as label for
aspirations or resentments which may be legitimate in
themselves, but which should not delude us as to the
final cause of this linguistic campaign. A nation which
is sympathetically receptive to outside ideas does not
fear to welcome the words by which these ideas are
habitually designated. It is abuse alone that is to be
condemned. It would be abuse to welcome under foreign
names what we already possess. It would also be abuse
to employ the foreign words on every occasion , and
before every audience.
To find the true measure, we must remember that
Language is a work of collaboration, in which the hearer
takes an equal part. A foreign word which would be
quite in keeping were I addressing specialists, would
seem an affectation, or might be a source of obscurity
with an uninitiated public. I am not at all shocked to
find English words in a French article on horse -racing or
on coal-mines ; but a man who reads a novel or hears a
play, expects the language used to be intelligible to
everybody. There can therefore be no uniform solution
for this question as to foreign words. The societies which
264 SEMANTICS

devote themselves to the purification of Language can

only legitimately attend to the language of conversation
and of literature. As soon as they push their pretensions
any farther, they are but performing a useless and
troublesome function .

When it is our moral life that is in question, the

presence of foreign words may produce the impression
of a discord. The more intimate the sentiments to be
expressed, the more contracted docs the linguistic circle
becomc. There is here for the reader or listener an
intellectual pleasure of a most subtle kind. Like the
houscwives of old, who prided themselves on consuming
the milk froin their own farın , the fruits of their own
garden alone, a dainty mind is sensitive to a language
in which all the words are products of the same soil, and
have an air of familiarity and of kinship. This pleasure
may become a very keen one when, in this simple
language, the writer expresses generous sentiments or
weighty thoughts. We seem then to receive the same
impression as when we see a fine deed simply done.
We have at the same time the vague feeling that all this
must have been known to our forefathers, since they
already possessed what was nccdful for saying it, and
that in conscqucncc we must be the children of a very
ancient and noblc racc. In such a case, the use of a
ſorcign word is not only devoid of inotive, it is actually
harmful. This was already understood by the author of
the Pricellence du langage françois, when in speaking
of Italian words, at that time so frequcnt in France,
 WHAT IS PURITY OF LANGUAGE ? 265

he said that they were " non pas françois, mais gdte
françois " (not French , but mar-French ).
It may scem pucrile to wish to limit one's vocabulary
to words admitted into some one or other particular
official collection. Yet I remember hearing a master of
the art of writing say, that the idea of the Diction
naire de l'Académie was a reasonable and a true one,
since it teaches the French what words to use if they
wish to be generally understood. As the limits of thi
vocabulary have not seemed narrow to men of the highes
genius, very weighty reasons should be requisite
dctcrminc any one to scck clscwhere the cxpressic
necessary to thought.
It is not the admixture of foreign words that t
purity of a language has chiefly to fear, it is rath
the unseasonable use of scientific terms. I am speaki
now of that extraordinary prose which disguises un
abstract substantives the most ordinary things of
Take for instance the French phrases : un dynami
modificateur de la personnalité, une individualité au -de.
de toute catégorisation, une jeunesse qui sentimentalis
passionalité.
The impropriety is not always involuntary ;
intended to magnify things by exaggeration of langi
as when the French talk of les imperiosités du dés
célestes attentivili's. We find other studics as we
philosophy ſocding this obscure and pretentious
guage with ncologisms : mcdicinc, music, cxegesi
Middle Ages. Whilc verbs give birth to the
useless substantives (des frappements de grosse cais
 266 SEMANTICS

ferraillements de verrerie, les perlements de la peau, les

serpentements des bras), wc find, on the other hand,
substantivos producing no less cxtraordinary verbs ( il
soleille lourdement, une idée contagionne les esprits, etc.).
We cannot accuse these neologisms of being contrary
to analogy : from the point of view of grammar they
are impregnable ; but their mistake lies in being super
fluous, in substituting an expression which is both
heavy and colourless for what can be said more simply
and more vividly. Voltaire has defined what the French
call le génie de la langue, “ the genius of the language,”
as " an aptitude for saying in the shortest and most
.

harmonious manner what other languages express

less happily ." If we accept this definition , we may
say that the authors of these neologisms sin against
the genius of the French language. This same French
language has sometimes been reproached with not lend
ing itself easily to the formation of new words : with these
examples before my eyes, I am rather inclined to think
that it lends itself too much. English and German have
the resource of compound words : but an unsuccessful
compound, such as is made every day in both these
languages, possesses fewer drawbacks than other neolo
gisms, since the two terms thus momentarily associated
separate again at once, while these French abstract
nouns, welded together by means of suffixes, seem to
be forged for perpetuity .
Everything which is used is exposed to wear and tear :
we must not, therefore, be surprised if the same vocables,
the same metaphors, when employed during a protracted
 WHAT IS PURITY OF LANGUAGE ?

space of time, no longer produce the same impressio

the mind. The invention of new forms may thercran
accounted for. The important point is that the
sumption should not be more rapid than the produ
It is irony, caricature, the use of inverted comma
fierce struggle between the pulpit and the pres
exaggerations of the drama and the " feuilleton ,"
hasten the inevitable changes of Language. Dela
will has much more power to undo and to destro
to create. The origin of words is nearly always
semi-obscurity ; but we can often give the na
those who have discredited or lowered certain
or emptied them of their meaning.
This question of Neologism presents the most
aspects.
To condemn Neologism in principle and al
would be the most annoying and the most u
prohibitions. Each onward step of a langua
work first of an individual, then of a more or
minority. A country in which innovations
bidden would deprive its language of all c
development. By Neologism , we must under
bestowal of a new meaning on an old word a
the introduction of a wholly new vocable .
change which modifies pronunciation is at o
ceptible and constant, to such a degree that
who returns to a country after thirty years
can appreciate the march of time, so also ist
of words being ceaselessly transformed by a
268 SEMANTICS

cvents, of new discoveries, of revolutions in ideas and in

customs. A contemporary of Lamartine would find it
difficult to understand the language of modern French
newspapers. We all work more or less at the vocabulary
of the future, whether we are scholars or unlettered ,
writers or artists, men of society or men of the people.
Children have a part in it which is by no means
small : as they take up the language at the point to
which the preceding generations have brought it, they
generally are ten or twenty years in advance of their
parents.
The limit at which the right of innovation ceases is
not determined by the idea of “ purity ” alone, which
can always be disputed : it is also imposed by the need
which we feel of remaining in contact with the minds of
those who have gone before. The greater the literary
past of a nation, the more does this need impress itself
as a duty, as a condition of dignity and of force. Hence
arises the idea of a classic epoch, offered for the imitation
of subsequent ages ; an idea which is in no way artificial
or chimcrical, if the classical cpoch be not relegated to
too remote a period. In such a case, it is not the
philologists alone who must be consulted, for they might
allow themselves to be guided by more or less pro
fessional motives. The Swedish philologist, Erik Ryd
quist," placed the classical age of the Swedish language
at about the year 1300. An analogous point of view,
though not always openly expressed, is held by many
scholars. If they have to decide between two gram
I Died at Stockholm in 1877.
 WHAT IS PURITY OF LANGUAGE ? 269

matical forms or constructions, it is usually towards the

more ancient one that they incline. Thus, in Germany ,
it is Middle High-German which serves as criterion.
Every nation must judge for itself how far back into
the past it can cast its glances without losing its touch
with the present.

It is inevitable that Neologism, having first tried its

hand on words, should next attack the construction and
the grammar. But it here meets with greater resist
ance. We can barely, up to the present time, enumerate
three or four new turns of phrase which have to any
degree succeeded in gaining adoption. There are good
reasons for this. To change the construction or the
idioms, is to meddle with living things ; it is an attack
on an inheritance which represents centuries of research
and of efforts.
It is only right at this point to make due acknowledg
ments to a succession of obscure and modest workers,
whose names are nowadays but rarely mentioned, but
whose achievements still survive : I mean the series of
French grammarians, from Ménage to Olivet. I wish
to draw attention here to the meed of recognition which
is due to them , for modern philology is but too ready to
deny or even to condemn their influence.
These worthies, whose names were Du Perron, Coef
feteau, Malherbe, La Mothe Le Vayer, Vaugelas, Chape
lain, Bouhours, were not scholars by profession, but for
the most part men of the world who had been led by a
natural taste to occupy themselves with the problems or
270 SEMANTICS

difficulties of the French language. What they had in

view was above all the purity of the language, which
signified on the one hand clearness, and on the other
hand propriety. To eliminate all improper or misbe
gotten expressions, to make war against all double
meanings, to discard all that is obscure, useless, low,
trivial, such is the enterprise to which they devoted
themselves with much abnegation and perseverance.
They sought for the rules, if needs were, they invented
them . They were " admirable rules." Vaugelas declared
that he had found “ a thousand admirable rules " in the
writings of La Mothe Le Vayer. " I hold this rule ,” he
says elsewhere , " from one of my friends, who had it
from M. de Malherbe, to whom must be given the honour
thercof.” And yet again : " This rule is most admirable,
and conducive to the purity and clearness of the
language. ... It is true that in speaking one does
not observe it, but style must be more accurate. . . .
Neither the Greeks nor the Latins have scruples on this
point. But we are more exact in our language and in
our style than the Latins, or than any of the nations
whose writings we read. ” The public felt the same on
this point, and only asked to be directed.
We have some difficulty now - a -days in imagining a
public welcoming interdictions and ready to make pro
hibitions yet more stringent. The philologist has in this
contributed to the education of the public. The modern
philologist rejects nothing ; all that exists has a reason
for existence. But the point of view of these law -givers
was a different one, and, in truth, the languages to
 WHAT IS PURITY OF LANGUAGE ? 271

which a period of strict regulation has been lacking

will under examination still betray, as it were, a lack
of primary education. What must alone be regretted
is that the purification did not come at an earlier
date. The religious wars brought about a delay of
more than half a century . If placed under discipline
sixty years sooner, the French language would certainly
have maintained a greater suppleness, for these worthy
masters were as assiduous in preserving as in pruning,
and, since they took thought " for all the graces of our
language," they would doubtless have saved some of the
earlier freedom.1
They liked and esteemed the task with which they
had voluntarily burdened themselves. They recognised
its importance, because " it needs but a bad word to
bring one of a company into contempt, to discredit a
preacher, a lawyer, or a writer. Finally, an evil word,
since it is easily remarked , is more capable of harm
than a bad argument, which few people perceive.” They
are conscious of the permanency of their work. " I
propound principles which will last no less a time than
" I will take as example the gerundive, the use of which has
been regulated to excess. To explain what I mean, we will take
this sentence : “Mon père m'a fait en partant mille recommenda
tions (“ My father gave me on departing much good advice " ).
Modern grammar insists on restricting "en partant ” exclusively
to the subject. There is here a certain exaggeration, since “en
parlant ” means nothing but “ at the moment of departure," and we
have a right to interpret it as befits the general meaning. Italian
has in this respect reserved for itself a greater freedom . It is just
to add that this rule was not always observed in the seventeenth
century .
272 SEMANTICS

our language and our empire. They are maxims

which will never change, . for even if there be any
change in the usage which I have observed, the differ
ence will still, both in writing and speaking, be in
accordance with these same observations. ” 1
It would be a mistake to regard them as bigoted
logicians. On the contrary , they had arrived at the
conviction that logic was admissible everywhere, save
in the matter of Language. “ The beauty of language
actually consists in this illogical way of speaking, pro
vided always that it is authorised by custom. Here
only is strangeness fitting. . . . It is noteworthy that all
the ways of speaking which custom has established in
contravention of the rules of grammar, should, far from
being regarded as vicious, or as errors to be avoided, be
on the contrary cherished as an adornment of language,
which exists in all beautiful languages whether living or
dead .”
The want of order and of rule is not confined to
words, it extends to idioms and to phrases. “ It is
indubitable that each language has its special phrases,
and that the essence, the wealth , the beauty of all
languages consists principally in using those phrases.
Not that one may not at times make new ones, whereas
it is never allowable to make new words ; but it must be
done with infinite precaution ..." Otherwise, instead
of enriching the language we corrupt it.
These scholars of the seventeenth century are there
fore convinced that in all cases there is one right form ,
1 Vaugelas, Remarques sur la languefrançaise.
 WHAT IS PURITY OF LANGUAGE ? 273

and one only. So without hesitation they proscribe

“ the bad form,” which is often merely the less used or
the earlier form .
The idea of utility is with them dominant over every
other consideration. Since men have received Language
in order to make themselves understood, to sanction two
forms between which a choice were possible, would be
to open the door to misunderstandings and discussions.
It is not therefore the business of the grammarian to
avoid, to Ainch from difficulties. He must look them in
the face, and establish certain rules. We may smile
at this tone of authority, but it is fortunate for the
permanence of the French language that it has in the
past possessed minds of this temper.
But it is not in the name of their own personal
authority that these students of Language pronounce
their judgments. It is in the name of right usage, and
if asked where they find this right usage, they answer
without hesitation that they find it at Court. The
language of the provinces can but contaminate the purity
of the true French language. Fénélon , in this, agrees
with Vaugelas : “ The most polished persons have diffi
culty in correcting themselves of certain mannerisms of
speech which they have acquired in their infancy, in
Gascony, in Normandy, or even at Paris, from inter
course with servants . . . " The Court even is not
exempt from blame : “ It suffers a little," continues
Fénélon , " from the dialect of Paris, where children of
the highest rank are usually brought up."
I have quoted these opinions on purpose to show how
r
274 SEMANTICS

far removed they are from the theories which now gain
credit.
In the eyes of modern philology, all forms, from the
moment that they are used, have a right to existence.
Indeed the greater the alteration, the greater the interest.
The true life of Language is concentrated in dialects ;
literary language, artificially arrested in its develop
ment, has not nearly the same importance. One should
beware of treating the mother-tongue as a subject for
instruction ; such a course merely disturbs the free
expansion in the children of their faculty of Language.?
As we have been shown by the historian Savigny that
the idea of right and of ethics is not applicable to
the historical development of a people, so also is the
idea of good and evil not applicable to the development
of a language.
These doctrines do not seem to carry conviction to
Mr. Noreen. Since Language is our primc method of
communication, there must be some agreement as to the
manner of using it. Who is to be judge in this matter ?
We here ask permission to quote the actual words of
the Swedish writer : “ It will not,” he says, “ be the
historian of Language, who can speak of the past alone ;
it will not be the philologist, whose function is to
describe the laws of Language, but not to dictatc thcm ;
it will not be the statistician, who merely registers the
usage. To whom then is this authority to be attributed ?
It pertains to the inventor, to him who creates the forms
Jacob Grimm , preface to the first edition of his Deutsche
Grammatik.
 WHAT IS PURITY OF LANGUAGE ? 275

used subsequently by the mass of mankind, to the

writer, to the philosopher, to the poet. ... We are the
crowd, who clothe our thought in the garment which
they have created ; we wcar this garment and we wear
it out. By ourselves we can contribute but little to the
development of Language ; it is under the sole direction
of these masters. We must resign ourselves to being
mere pupils, and it is not the business of the pupils to
command . "
If these words came from less far away, we should
doubtless think them less striking. We have frequently
heard something of the same kind both in prose and in
verse. But it is interesting to find at Stockholm , in
a man who has at his command a science of which
Vaugelas and Bouhours did not possess even the first
elements, the confirmation of principles which these
writers followed instinctively in their observations and
criticisms. The idea of a type of correctness and purity,
furnished by refined society and by the best writers,
from having been almost a commonplace during two
centuries, was proclaimed vain and inadequate in the
name of a science which claimed to draw inspiration
from a higher principle : this same idea returns to us
to-day from the north, propounded, with conviction and
' force, by onc of the masters of Scandinavian philology.
THE HISTORY OF WORDS
 THE HISTORY OF WORDS
UNDER this title : La vie des mots étudiés dan
significations (“ The Life of Words as the Sym
Ideas," English translation , published by Kegar
Trench and Co.), M. A. Darmcsteter, a professor
Sorbonne and a distinguished novelist, has lately
a pleasant little book , calculated to add to the pop
of works on philology. We learn in it successive
words are born, how they live together, and ho
die. It has to do with the meaning of words, n
transformations of forms, which belong to another
of science. Of all departments of philology this
tainly the one most likely to interest the general
All display of advanced scholarship would here bu
place. The facts under observation contain n
mystery. The changes which have supervened
meaning of words are as a rule the work of the
! We hero reproduce extracts of what we have written e
on 11% sur 7 , lay Arudne armonteer. In this arti
dates from 18H7 , will he found the first iden of our demand
this person, is well as for the sake of some of the example
we think that this partial reproduction may not be withou
(sco above, p . 3)
279
280 SEMANTICS

and, as always when the popular intelligence is in

question , we must be prepared, not for a great depth of
reflection , but for intuitions, for associations of ideas,
sometimes unexpected and strange, but always easy to
follow . It is therefore a curious and fascinating spectacle
to which we are bidden by this history.
A mind, however, which is not satisfied with mere
appearances may desire to penetrate, behind the varied
and changing aspect that it presents, to the first cause,
which is none other than the human intellect : for to say
that words are born, live together and die, is indisputably
pure metaphor. When we speak of the life of Language,
or when we call languages living organisms, we are using
figures of speech which may help us to make ourselves
more intelligible, but which, if taken literally, will land
us in the realm of dreams. M. Darmesteter has perhaps
not been always sufficiently on his guard against this
kind of setting. As it is easier for men to observe out
ward objects than to read within themselves, we are
usually more ready to discuss the products of the mind,
than the faculty from which these emanate. But while
yielding, for the sake of greater facility of speech, to this
natural inclination , it is advisable from time to time to
rectify the illusion. We must not be afraid of occasion
ally examining the interior of the instrument to which
we owe these projections. Apart from the human mind,
Language has neither life nor reality.
Almost simultaneously with the book of which we are
speaking, there appeared in Germany the second edition
of a somewhat difficult and ponderous work, which ,
 THE HISTORY OF WORDS 281

among other questions, discusses the subject treated by

M. Darmesteter. We mean the Principles of Language
of Mr. Hermann Paul. The author is professor of the
'German language and literature at the University of
Freiburg. In the main these two works complete each
other : they are both volumes of Semantics.

By a noteworthy coincidence, the two authors start by

agreeing on one particular point : this is that each ,
although having doubtless at his command a fairly large
number of languages, has preferred to take for his special
field of study his mother-tongue. This is an indication
which is not without importance. The investigation in
question is one which demands an intimate and direct
knowledge of the subject, and is thus unlike an inquiry in
phonetics or morphology. The modifications which have
come about in the body of the language, such as the
suppression of a letter or syllable, the welding of a new
inflection , the substitution of one termination for another,
strike the eye at once ; but the observations which
engross the student of Semantics are more elusive. It
is chiefly when we have to note the impression made on
the mind by words that the chances of error are multi
plied ; they are almost inevitable in dealing with a foreign
language. A German writer who has touched on these
subjects repeats again and again in his books that the
French word ami is far from possessing either the ring
of sincerity or the depth of the German Freund. An
ingenuous prejudice, but one which is not difficult to
understand ! A few years ago, another scholar detected
282 SEMANTICS

in the French merci a certain wounding and degrading

element : he was thinking of the Latin mercedem . These
kinds of delusions show us the danger. They are a proof,
that the most familiar ground is also the best for this
kind of research . When the general outlines of the
science of Semantics have been traced out, there will be
no difficulty in verifying in other languages the observa
tions taken in the mother -tongue. Once the general
divisions established, we can add to them all facts of a
like order gathered from no matter where.

Let us therefore, without further delay, penetrate into

the domain of Semantics, and observe some of the causes
which govern this world of speech.
We will begin with a point which has a real import
ance for the history of meanings, and which, till these
last few years, we have not taken sufficiently into
account : this is the influence which the words of a
language exercise over each other from a distance. The
signification of certain words becomes more and more
restricted, because they have colleagues which are expand
ing their meaning. In dictionaries in which each term is
studied on its own merits alone, we hardly perceive the
play of this kind of compensation and balance ; it is
only in the most recent and most fully developed voca
bularies, as, for example, in the continuation of Grimm's
dictionary, that the authors have begun to give due place
to this interesting series of parallels. Thus in the old
French language, the verb traire maintained all the uses of
the Latin trahere : traire l'épée, traire l'aiguille, traire les
 THE HISTORY OF WORDS 283

cheveux, were habitually said . Whence comes it that so

common a term has ended by being reduced to the single
signification which it bears to - day, of traire les vaches,
traire le lait ? It is entirely owing to the fact that a rival
of Germanic origin - tirer - has in the course of centuries
invaded and occupied its whole domain. The human
mind is averse to hoarding useless riches : it little by
little discardswhat issuperfluous.However,and thisis y
an observation on which M. Darmesteter is right to insist,
a word may be in jeopardy, may indeed even succumb,
without its compounds and derivatives being at all
affected. As witnesses to the ancient usage, the French
still have the compounds extraire, soustraire, distraire,
the substantives trait, attrait, retraite.
A similar fate befell the French verb muer, which had
to yield all save a small corner of its territory to a
new-comer, changer. Commuer and remuer have sur
vived the ruin of their primitive. The history of sevrer
is the same, séparer having almost entirely ousted it.
This kind of contest, or, to use the Darwinian language,
this struggle for life, is particularly striking, when the
two competitors are, as in the last example, children
of the same stock . This kindred origin makes, moreover,
no difference in essentials.
In the central provinces of France, towards the six
teenth century, the r when placed between two vowels
took the sound of an s or z. This change of pronuncia
tion brought about the change of chaire (cathedra) into
chaise. Commines, in the fifteenth century, could still
say : “ Ladite demoiselle était en sa chaire et le duc de Clèves
284 SEMANTICS

d côté d'elle. " The modern form having prevailed, the

carlier vocable was compelled to beat a retrcat, remain
ing only to designate the pulpit or the professor's chair.
Every new word introduced into a language causes a
disturbance analogous to that resulting from the intro
duction of a new-comer into the physical or social world .
A certain length of time is needed for things to settle
down and subside. At first the mind hesitates between
the two terms : this is the beginning of a period of
fluctuation . When, to denote plurality, it became the
custom in fifteenth -century France to employ the peri
phrase beau coup, the ancient adjective moult did not
incontinently disappear, but it began from that time
forward to age. Then, after all kinds of uncertainties
and contradictions, one of the two competitors steals a
decided advantage over the other, distances its adver
sary , and reduces it to a small number of uses, cven if
not absolutely effacing it. In expounding these facts,
we find ourselves in our turn falling into the figurative
language with which we reproach M. Darmesteter, so
naturally does this form present itself to our minds.
But every one understands that it is a question of mere
acts of the mind. When, for any reason , we have begun
to adopt a new term, we little by little engrave it in
our memory, we render it familiar to our organs, and
we transfer it from the reflective to the spontaneous
regions of our intelligence ; so that it is the same with
this new term as with a gesture, which by dint of
repetition becomes peculiar to one's self, and ends
eventually by forming part of one's personality.
 THE HISTORY OF WORDS 285

Truth to tell, the acquisition of a new word , whether

it has come to us from some foreign idiom, or has been
formed by the association of two words, or has issued
suddenly from some ignored corner of society, is of l
comparatively rare occurrence . What is infinitely more
frequent, is the application of a word which is already in
use to some new idea. Therein lies in reality the secret
of the growth and renewal of our languages. For we
must note that the addition of a new signification in no
way injures the original meaning. They can both exist,
without influencing or harming each other. The more a
nation has advanced in culture, the more numerous and
diversc are the acceptations accumulated by the terms of
which it makes usc. Does this result from poverty of
language ? or from sterility of invention ? He would be
but a superficial observer who could think so . We will
now explain how things really comc about.
In proportion as a civilisation gains in variety and in
richness, the occupations, the acts, the interests of which
the life of the society is composed, are divided among
different groups of men : neither state of mind nor
trend of activity are the same in the cases of priest,
soldier, politician, artist, merchant, farmer. Though they
have inherited the same language, the words become in
each casc coloured with a distinct shade, which pervades
and finally adheres to them. Habit, surroundings, the
whole ambient atmosphere, determines the meaning of
the word and corrects its too general signification. The
most comprehensive words are therefore for that very
reason those which have the greatest aptitude for lending
286 SEMANTICS

themselves to numerous usages. At the word operation,

if pronounced by a surgeon, we have a vision of a patient,
a wound, the instruments for cutting ; if spoken by a
soldier, we think of armies on a campaign ; if by a
financier, we understand that the matter concerns Auctu
ations of capital ; if by an arithmctic master, it is a ques
tion of additions and subtractions. Every science, cvery
art, cvery trade, in composing its terminology, marks with
its impress the words of the common language. Imagine
all these divers acceptations cnumcrated in succession ,
as is done in our dictionaries ; we should be surprised at
the number and variety of the significations. Does this
result from the poverty of the language ? No ; it comes
from the wealth and activity of the nation.
I have before my cycs a German-French dictionary, in
which, to gain room, the author begins by distinguishing
in the French language two hundred and thirty -four
different occupations, sciences, or professions, of which
he gives a list, and to each of which he affixes a number.
The reader is warned that he must always refer to this
table. When the word is followed by a 1 , it is taken as
a term of theology, 7 indicates anatomy, 9 arithmetic, 21
astronomy, 51 the language of carpenters, 118 that of
bookbinders, 233 that of coachmen. One single word,
for example effet, exercice, conversion , in the body of
the dictionary, is followed by five or six different transla
tions, each of which has its number. We see how
mistaken are those who, to estimate the wealth of a
language, think it sufficient to count the vocables.
No name has so far been given to the faculty which
 THE HISTORY OF WORDS 287

words possess of appearing under so many aspects. It

might be called Polysemia. We may say in passing, that
the inventors of new languages (and the number of these
has specially incrcased within the last few years) do not
take sufficient account of this faculty ; they think that
they have done all that is nccdful when they have
provided each word with one representative, never
dreaming that for one single word they ought often
to create six or cight ; or again, if in their language
they succeed in reproducing the French Polysemia, the
English and the Germans reasonably complain that they
are made to speak French in Volapük.
How is it that this multiplicity of meanings produces
neither obscurity nor confusion ? It is because the word
reaches us already prepared by what precedes and by
what surrounds it, interpreted by the time and place,
determined by the characters who are on the stage.
Curiously enough, it has but one meaning for the hearer
as for the speaker, since there is an active way of listen
ing which accompanies and anticipates the speech. We
have only to light unexpectedly upon a conversation
already in full swing, to realise that words by themselves
are but uncertain guides, and need that setting of circum
stances which, like the key in music, determines the
value of the signs. Writers of comedy fully realise this
faculty of Polysemia, which is the source of the cross
questions and crooked answers with which they enliven
their audiences.

The diversity of social surroundings is not the only

288 SEMANTICS

cause which contributes to the growth and renewal of

the vocabulary. Another cause is the need inherent in
our natures of representing and depicting by means of
images that which we think and feel. Words too fre
quently employed cease to make any impression. It
cannot be said that they wear out. If the only function
of Language were to appeal to the intelligence, the most
ordinary words would be the best : thc nomenclature of
algcbra docs not change. But Language is not addresscd
solely to reason : it strives to touch , to persuade, to soſten.
Thus, for things which are as old as the world, we find
new figures of speech arising, sprung from no one knows
where, perhaps from the brain of a great writer, more
often from that of somc unknown person. If the images
are sound and picturesque, they are welcomed and
adopted. Employed originally as figures of speech,
they may in the long run become the actual name of
the thing.
The subject of Metaphor is inexhaustible. There
is no genuine connection or fugitive resemblance which
has not furnished its contingent ; no rhetorical trope is
too bold for Language to use every day as a matter of
course. The examples are so numerous, that the only
difficulty is to choose. The maritime vocabulary seems
at all times to have possessed a special attraction for
the dweller on dry land : hence, for the most ordinary
actions, a continual deposit of nautical terms. Accoster
(to accost) a passer-by, aborder (to broach, lit. board)
a subject, échouer (to fail, lit. to strand ) in an under
taking, all these are metaphors taken from the sea.
 THE HISTORY OF WORDS 289

Words in constant use, such as to arrive, have the

same origin. It must not be thought that this is the
case in modern languages alone. The Latin verb
meaning " to bear," portare, which began at an early
date to strive for mastery with fero, and which Terence
already employed when speaking of news which is
brought, signified " to take into port.” We have re
sumed some part of it in to import, export, deport. It
was a terın belonging to the incrchant service. Grock
in this matter showed less desire for innovations, so
that portare belongs exclusively to the Latin tongue.
As a rule, when, for the expression of a familiar idea, one
of the ancient languages departs from the usage of its
brethren, we may take it for granted that it has adopted
a metaphorical expression. We know that opportune,
importunate, are in a like manner figures of speech
borrowed from the idea of a more or less easy landing.
placc.
The horse and horsemanship have supplied a large
amount of figurative expressions. A whole volume has
been compiled out of them. They may be classed by
epochs, the most ancient terms having already passed into
a condition of colourlessness. The French, for example,
say of a man who has through some shock momen
tarily lost the use of his faculties, that he is désarçonné
(unsettled, lit. unhorsed) or démonté (unhinged, lit.
dismounted) ; of a confused speaker they say that he
s'enchevêtre (gets entangled) in his arguments, comparing
him to a horse whose legs are entangled in the tether of
his halter (chevétre = capistrum ). They continue the same
U
290 SEMANTICS

comparison of an animal at pasture by saying that he

appears to be empêtré ( impastoriatus, entangled ) ; em
barrassé would be more polite, but would bring us back
to the same idea of a bar serving as a restraint. Finally
there are words of whose metaphorical origin we are no
longer conscious. Thus travail (work), which plays so
large a part in economical discussions, and which is
commonly used by a writer or an artist in speaking of
his productions, leads us back to the same image of a
horse shackled and bound. Thanks to the turf, this
factory for mctaphors is not likely to come to a stand
still. At the present day we hear of pupils qu'on entralne
(who are being trained), and of amateurs qui s'embal
lent (who lose their heads, lit, run away).
Again, what a number of expressions, and those of the
most various kinds, Language owes to sport. When in
familiar speech the French say of some one that he
seems déluré (wide-awake), they use a metaphor bor
rowed from falconry, the falcon , which is déluré or
déleurré, being one which does not allow itself to be
taken by the lure. In a quite different style, when
Pauline, speaking of dead Polyeucte, crics :
Son sang, dont ses bourreaux viennent de me couvrir,
M'a dessillé les yeux et me les vient d'ouvrir,".

the heroine of Corneille uses an image of like origin ,

dessiller (which should be written déciller) meaning
merely to unstitch the eyelashes of the falcon , which
had been temporarily blinded for taming purposes.
We see how different a fate may in the course of time
 THE HISTORY OF WORDS 291

befall two terms of identical origin : so great a diverg

ence is explained by the successive stages of the journey,
and by the good or bad treatment which the word has
undergone by the way. Dessiller les yeux has been
employed in religious language : this is what gives it
dignity and nobleness. For it is a great and inestimable
benefit for a nation to have in its literature a sacred book
that all read and know. The language may subse
quently sustain all kinds of injuries, it will always
possess a source of purification. This service was
rendered to the English language by the Bible of
1611 , to German by Luther's translation. The great
preachers of the seventeenth century rendered an analo
gous service to the French language. There are, on
the other hand, dark corners of literature which wither
all that they touch, and which, if they appropriate an
expression, return it tarnished and dishonoured.
Like shells which strew the seashore, relics of animals
which existed, some but yesterday, some long centuries
ago, languages are filled with the spoil of modern or
ancient ideas, some still living, others long since forgotten.
Every civilisation, every custom, cvery conquest, and
every dream of mankind, has left a trace, which may,
with a little attention , be brought to light.
Consistency of style, coherency of metaphor, however
right to recommend, is absolutely to seek in Language ;
or rather, it is possible and necessary for the topmost
stratum only. Otherwise, we should deprive ourselves
of the simplest expressions, and speech would become
as difficult as is the daily intercourse of life among those
292 SEMANTICS

Asiatic religions in which all that has ever had life is

regarded as impure. The ancient languages are subject,
in this respect, to the same conditions as the modern
ones, being ancient in relation to us alone, and having
themselves inherited the legacies of centuries. When
Sallust makes Catilina say : “ Cum vos considero, milites,
et cuin facta vostra æstumo, .
. " he thinks no more than
we do ourselves of the origin of expressions which seem
to him quite simple. Yet considero is a metaphor
borrowed from astrology and astumo from banking.
Were we to believe the lists of roots drawn up quite
arbitrarily by llindoo and Arab grammarians, we might
bc dcluded into thinking that languages began with the
most general ideas. We perpetually find among them
roots which signify " to go, resound , shine, speak, think ,
feel.” But it is our ignorance of a previous age which is
sole cause of this delusion .
Collections of rhetoric contain no catachresis , litotes or
hyperbole of which popular language does not perpetually
furnish abundant examples. A grammarian of the
eighteenth century, Dumarsais, wrote a treatise on tropes,
one edition of which had the unexpected honour of being
dedicated to Mme. de Pompadour. But what are these
examples, culled from the surface of the soil, compared
to those which are exposed to view by deeper excava
tions ? Were we to say that there exists an idiom
in which the same word which designates the lizard
signifies also a muscular arm, because the movement
of the muscles under the skin has been compared to
that of a lizard, our explanation would be received with
 THE HISTORY OF WORDS 293

doubt, or at all events we should be supposed to be

referring to the imaginations of some savage people.
Yet the word in question is lacertus, which means lizard ,
and which has often been used by poets and prose writers
to designate the arm of a hero or athlete. At other times
the mouse has been substituted for the lizard, whence we
get musculus, a word which mcans, as our readers are
aware, sometimes mouse and sometimes muscle. This
singular image scems to have been a favourite one at all
timncs. Littré draws attention to the fact that in French
the leg muscle of a leg of mutton is called souris. In
inodern Greck the rat is called mys pontikus (water rat),
or more shortly, pontikos. Now the adjective has also
replaced the substantive in the case of the other meaning,
and pontikos designates the muscle.
Our author has tried by means of pictures, or, as we
now call them, schemes, to demonstrate the radiation or
connection of the different meanings of a word. Some
times he explains it as a star, sometimes as a broken
line. But it must be remembered that these complicated
figures of speech are only valuable to the philologist :
the inventor of the new meaning forgets at the moment
all previous meanings save one, so that associations of
ideas are always in pairs. It is in no way the business
of the people to peer into the past ; the popular mind
knows but the signification of its own day. In this
connection an ingenious comparison has been drawn
with those bold climbers who remove from under the
right foot the cramp iron which supported it, as soon
as they have placed the left foot on the following one.
294 SEMANTICS

The philologist is alone in his search for the trace of

these moveable steps.

Any one who, in compiling the history of the variation

of meanings, took the words only into consideration ,
would either run the risk of losing a portion of the facts,
or be in danger of explaining them wrongly. A language
is not formed solely of words ; it is composed of groups
of words and of phrases.
We have all, when looking out some uncommon word,
often read in our dictionaries, " Now used only in the
expression .
Then there usually follows some
proverbial idiom or some technical term , or a phrase
more or less devoted to some special use. If we reflect
on the cause of this phenomenon we shall be led to
new
consider the elements of Language under a
aspect. The philologist attributes to the word a
personal cxistence which is continuous through all the
associations and combinations in which it takes part.
But in reality, as soon as the word has entered into a
formula which has become at all common, we perceive
only the formula. Vocables which have long since
ceased to be employed on their own merits, and which
are difficult to recognise when withdrawn from the one
position left to them, are still preserved in certain special
associations. What, for example, is the French word
conteste ? It has dropped out of use for so long, that
we should find it difficult to say even to what class it
belongs. But it still appears in the idiom, Sans con
teste. What, as the name of a colour, is bis ? It used
 THE HISTORY OF WORDS 295

to designate brown or black. À tort ou à droit, à bis ou

d blanc... l'un veut du blanc, l'autre du bis, were
common expressions. It was the Italian bigio. But it
is now used only when speaking of bread . Demeure, in
the sense of delay, has almost disappeared ; but every
body understands the expression, Il y a péril en la
demeure ( there is danger in delay ).
It is not the word which forms a distinct unity for our
mind, it is the idea. If the idea be simple, it matters
little that the expression of it should be complex : our
mind will be conscious of the totality only. One might
even go farther, and wonder whether the majority of
men have a clear and distinct conception of the word.
Every one knows that illiterate people are guilty in
writing of the strangest separations, as also of the most
curious couplings. This does not preclude some of
them from handling thought with accuracy, speech with
correctness. Their intelligence, while embracing the
wholes, has never had leisure to descend to details.
Missionaries who first fix in writing the language of
savage peoples, know how difficult it is to discover
where the words begin and end. If Etruscan has so
far resisted all attempts at deciphering, it is owing in
part to the defectiveness of the divisions.
Accustomed as we are to the service which writing
renders us, we are tempted to show ingratitude towards
it. The new school of fonetists does not perhaps pay
enough attention to this, at least in the case of the
more advanced portion of the school - for I do not
wholly disapprove of the undertaking. In our modern
296 SEMANTICS

languages, in which so many vocables of different origin

and signification have become identical for the ear, the
words are not engraved on the mind by their sound
alone, but also by their aspect. In the absence of ortho
graphy, we should be driven to have recourse to an
explanatory commentary , as do the Chinese, and as we
do ourselves when we say : the site of a building, the
sight of a rifle.
Once it has been framed in an idiom, the word loses
its individuality and has no further concern with outside
events. It is therefore not accurate to speak, even in
metaphor, of the life and death of words. A word may
be dead for the mind, and yet continue to figure in some
special context, in which it is perceived not as a word,
but as the integral portion of a whole. Imprisoned in this
retreat, it escapes all changes of Language, all revolutions
of usage and of ideas. The French say res -de-chaussée,
although rex ( rasus) has dropped out of ordinary use.
Faire un pied de nes survives in spite of the metric
system. They still speak of rhumes de cerveau , although
in the eyes of modern medicine the brain has little to
do with a cold in the head.
As soon as a word has passed into an idiom, its
proper and individual meaning is obliterated for us.
These kinds of incoherences usually strike foreigners
more than ourselves, especially if they have learnt the
language by scientific methods, rather than by use.
Thence the purism so readily affected by foreigners who
speak and write the English or French which they learnt
at school.
 THE HISTORY OF WORDS 297

One may deduce from this class of facts some con

clusions on the manner in which languages are modified
and disintegrated. If we depended on the teachings of
Phonetics alone, words would be transformed singly and
independently, according to the number of syllables and
position of the accent, and in conformity with invariable
rules. Moreover, inflections which were doomed to
perish would be simultancously cxtinguished in all
words of the same class. The construction would be
modified in a uniform manner in all phrases composed
of the same logical elements. But we find nothing of
the kind. Regularity does not exist, and the reason for
this is, that a language is by no means a collection of
words only, but includes groups alrcady collected and
so to speak articulated. In the Christian inscriptions of
the first centuries, we find that in the middle of a Latin
which is extremely incorrect and already half Roman,
there survive whole formulas of a very tolerable Latinity :
these are formulas which had been saved from oblivion
: by daily usage, and in the case of which analysis and
comprehension of the elements had been rendered un
necessary by a previous familiarity. A nation which
forgets its language somewhat resembles the school-boy
who recites a half- remembered lesson : though there
may be some parts in which the words present them
selves disjointedly and imperfectly to his memory, there
are also others in which they come back in a rush,
whole phrases at a time. Again, we observe something
of the same kind when two languages touch and mingle,
for example on the frontiers of two countries. It is
298 SEMANTICS

not words alone but entire phrases which pass from

one people to another. M. Schuchardt's work on the
mingling of languages furnishes examples of this fact
as strange as they are various.
We are taught, and rightly, that the cases of the
Latin declension no longer exist in French : yet leur
and Chandeleur are genitive plurals. It is doubtless
through no special gift of longevity that they have
survived their kindred : it is owing to the idioms in
which they were, so to speak, embalmed.
Fèvre, in Old French, meant " workman " ( faber );
orfèvre preserves the Latin construction. When the
French say la grand'rue, la grandmère, they speak
the language of the thirteenth century . They are
genuine fragments of Latin or Old French, borne
along by the language of to-day, regardless of the
changes in grammar and in construction.

We each of us possess an assortment of abbreviated

expressions, intelligible only to our intimate friends.
Supposing that these abbreviations are adopted all
round us, that they become of current usage among a
whole class of people, and that they are disseminated
in the press, they may one day take their place in the
language. Such is the origin of general. It is evident
that this is an insufficient expression for the designation
of a military grade. But if we go back to the six
teenth century, we find that the phrase is completed in
captain general. There are, in the animal kingdom,
certain crustaceans which, when seized by one of their
 THE HISTORY OF WORDS 299

claws, fall hastily to the ground, leaving the enemy in

possession of the claw, and making use of the nine
others to flee with utmost speed. Our idioms undergo
an amputation of the same kind with this difference,
that the claw which remains takes for us the place of
the whole animal. What is the meaning of the name
école centrale ? Absolutely nothing. The words des
arts et manufactures must be added. I have often
taken part in interminable discussions on l'enseignement
spécial, and on the meaning which the founder had
probably attributed to that adjective. No one, not even
the founder himself, had thought of referring to the
charter of foundation , in which there is mention of un
enseignement spécial pour l'agriculture, le commerce et
l'industrie. The finest epoch of the French language was
familiar with this jargon. There was canal, when the
king and his court amused themselves on the canal at
Versailles. There was caveau , when the court of Mon
seigneur played in the little room so named. Those
very names of monseigneur, monsieur, madame, are
ellipses concealing a more complete and resounding
title.
The philologist notes that in all languages the ad
jective has a tendency to replace the substantive. This
law, which seems to belong solely to grammar, pre
supposes another which belongs to psychology and to
history. A few examples will help to make my mean
ing clear. French has lost the ancient word which
served to designate the liver ( jecur), and has replaced it
by an adjective signifying "fed on figs" ( ficatum , "le
300 SEMANTICS

foie " ). But what are we to conclude from this change ?

That we here have a word belonging to the language of
cookery. Those who listen in restaurants to the calls
between the dining-room and the basement may detect
many an ellipse of the same kind. There is mention
in French law -books of a certain class of loan which is
called le prêt à la grosse (bottomry -bond ): this adjective
might leave us long in doubt, did we not discover from
elsewhere that it concerned a loan à la grosse aventure,
a sort of contract applying to risks by sca. The more
one is familiar with some profession or class of life, or
for that matter the more one wishes to appear so, the
more use will one make of this shorthand language . A
private passes from active service (l'active) to the reserve
(la territoriale ). In Paris a man of fashion attends all
the premières. Besides a saving of time , there is in these
sous -entendus something flattering to one's vanity , like
the charm of an initiation . All progress, all modern
inventions increase their number. We wait for the
express, or we take a special in the railway stations .
At the time of the Paris Exhibition of 1878 , visits were
made to le captif des Tuileries. Slang makes use of the
same process. Cache ta menteuse," says one of Zola's
characters to his chattering daughter. These examples
are drawn from close at home, from the language of
the present day ; but we might quite equally take them
from other modern languages or from the languages
of antiquity. Brother is in Spanish hermano, which
represents the Latin germanus, already employed in
the same sense ; but in itself it is an adjective which
 THE HISTORY OF WORDS 301

signifies "real, natural." Cicero , when saying in one of

his familiar letters that on a certain occasion he had
behaved like a real ass, uses this word : Me asinum
germanum fuisse.
Our examples have so far consisted almost wholly of
substantives ; but something similar exists in the case of
verbs. By force of habit complements can make them
selves understood by implication, so that, from being
transitive, the verb becomes neuter. This is the counter
part of what we have seen in the case of the adjective
turned substantive. Are you exhibiting ? is a question
that is perfectly clear to a painter. A woman who
entertains is universally allowed. Purchasers under
stand the meaning of a shop which is selling off, or,
in France, of une maison qui liquide. Our spoken
language is full of these expressions : so much so in
deed, that it has been said that an abundance of neuter
verbs is a sign of civilisation. Sometimes the idiom is
curtailed from the middle : of all methods of abbrevi
ation, this is undoubtedly the least desirable. Yet
geologists hold forth on the subject of tertiary man .
A French doctor would talk of paralytiques,progressifs.
I have heard a member of the French Academy, speak
ing of Mr. Max Müller, call him philologue comparé.
Among candidates at the Sorbonne, everybody knows
what is meant by un bachelier scindé. Again , to quote
from what are, we grant, terrible barbarisms, when, in
religion, we talk of Protestants and Catholics, we are
making use of an ellipse which, though more ancient,
is nevertheless similar in kind.
 302
SEMANTICS
Weconclude that in a quest
ion of Language there is
one rule which holds sway over all others . Once a sign
has been found and adopted for some particular object,
it becomes adequate to that object. You may mutilate
and materially reduce it, it still maintains its value . On
one condition that is to say : that the usage which
attaches the sign to the object signified , remains unin
terrupted. To reconstru a language with the sole help
of etymology is a risky ucntdertaking , which may succeed
up to a certain point in the case of the generality of
words , but which comes into collision with the particular
kind of obstacle resulting from idioms . One realises this
when deciphering the text of a language which has not
reached us through a living tradition . The origin of the
word is often clear , the grammatica form leaves room for
l
no doubt , yet the inmost meaning escapes us . They are
like faces, whose features we can discern , but whose
thought remains impenetra . The only ancient lan
ble
g w w r k a t w
uages hich e eally now re hose hich ave ome h c
down to us with an accompani of lexicons and of
commenta
m
--Latin , Greek , Heebnrtew , Sanscrit , Arabic ,
Chinese .ries

Littré, in a charming work cntitled Pathologie du

Langage, has collected a certain number of facts of the
same kind. We strongly recommend this fragment to the
notice of our readers : it is an extract from his great dic
tionary, and a collection of interesting and curious facts.
X 1 Littré, Études et Glanures. (This portion has been published
in the Bibliothèque Pédagogique. Delagrave.)
 THE HISTORY OF WORDS 303

But what the great French scholar calls pathology is

the normal development of Language, and an every
day event. It is through this only that languages can
adapt themselves to the expression of new ideas. It
implies no disease. When by a circuit they have attained
to the creation of a new term , they obliterate the path
along which they had travelled. Thus for the most part
etymology has but an historical interest. In ordinary
life, in the discussion of philosophical or political ideas,
the investigation of the origin of a word may serve as
starting point ; but it would be the mark of an inferior
mind to dwell too insistently upon this, or to base upon
it too long or important a chain of inference.
It has been rightly said that words are like glasses,
which must be polished and rubbed for a long time, lest,
instead of showing, they but conceal things. Too vivid
a recollection of etymology often spoils the expression of
thought, which it is in danger of disturbing by all kinds
of false reflections. The work of centuries and the
benefit bestowed by a long succession of thinkers has
been to enfranchise and emancipate words, without, how
ever, wholly estranging them from their kindred or from
their place of origin.
The only case in which we may legitimately speak of
pathology is the case in which one word is erroneously
employed for another, whether on account of some
resemblance in sound, or as the result of some other
accident. Such is the confusion which has arisen
between the French words habit (clothes) and habillé
(clothed) : this last, which should be written abillé, is a
304 SEMANTICS

metaphorical expression meaning " prepared, arranged .'

It was first used in connection with wood. Du bois en
bille ( wood in logs) is still said. The ancient meaning
survives in some idioms, such as, habiller un poulet
( to truss a chicken ), le voilà bien habillé (he is well
equipped ). Here, again, we note the constancy of
idioms, which continue their existence without paying
any attention to the general stream of progress.

A language is not composed solely of words and of

expressions : there must also be the machinery for con
taining and maintaining these materials.
Wilhelm von Humboldt tells us that we bear in our
minds a sort of grammar which, sooner or later, ends by
leaving its mark on Language. It is what he calls Die
innere Sprachform (the inner form of Language). There
is no reason why we should not accept this expression
provided we thoroughly understand it. It is quite clear
that the inner linguistic form is not a gift of nature,
since it varies in different languages, and since in one
single language it becomes modified in the course of
ages. The inner form of Language is nothing but re
membrance of the mother-tongue. But, in its turn, this
remembrance is impressed on the still fluctuating parts
of the language, and forces them into the established
framework.
This is, moreover, not the only problein of the kind.
Here is another no less curious.
1 We borrow this etymology from a verbal communication made
by M. Gaston Paris to the Société de linguistique.
 THE HISTORY OF WORDS 305

The material death of an inflection does not put a stop

to its usage. Long after the inflection has disappeared,
Language can still appeal to it, and demand the same
services as though it still existed. It is a remarkable
fact, that the absent infection continues to render these
services. Yet more, we find the grammatical function,
of which it was formerly the exponent, still propagating
itself, though deprived of all means of expression ; so
that the most important portion of its history is some
times that in which it has lost its outward and tangible
representative .
This survival of inflections may be observed in all
languages. We find a striking example in French in
.
expressions like la Rue Monsieur le Prince, l'Hospice
Cochin, l'Institut Pasteur. Although French has for
centuries lost the exponent of the genitive, we here use
genuine genitives. It is necessary, indeed, in order that
a fact of this kind may be produced, that the language
should have preserved a certain number of models. Ex
pressions like l'Hôtel-Dieu, l'Église Notre- Dame, la Place
Dauphine, have been the type which the language has
continued to imitate. We have but to peruse a list of
the streets and places of Paris to realise that the genitive
has never been more used than since it has been shorn of
all sign. We must add indeed that, as this usage is for
the most part confined to proper names, the popular con
sciousness has somewhat changed with regard to it, and
at the present day discerns in these names rather a kind
of baptism than a case denoting possession,
While on this subject I should say that one must
х
 306 SEMANTICS

beware of confusing languages which have had an

inflection and lost it, with those which have never pos
sessed one. English, with a facility to be envied by
other nations, transforms its substantives into verbs.
It takes, for example, the substantive grace, and says :
It would gract our life. What we are here conscious of
is an absolute infinitive : though in no way expressed,
the idca of the infinitive presents itself uncquivocally
to our mind. The phrasc takes its place in an ancient
mould which was formed at the period of the inflcction,
and which has survived it.
The various languages differ notably on this point.
Clearness of speech depends on the greater or less use
i that is made of these survivals. A language derives its
character from what it implies as much as from what it
expresses. A just proportion in this matter constitutes
the merit of a language, like the proportion of solids
and voids in architecture.
German has preserved the turns of phrase of a syn
thetic language, though many inflections have disap
peared or ceased to be recognisable. When Goethe, in
his Iphigenie, says : Denkt Kinder und Enkel, “ think of
your children and descendants," it is a genitive that he
is professedly using. But there is nothing external to
indicate it. The difficulty of the German language arises
1
partly from these keys, so to speak, of expression,
which sound for the inward ear alone.
This is not the place to multiply examples. But the
inner linguistic form to which Humboldt alludes does
not limit its action ; it is, to speak metaphorically,
 THE HISTORY OF WORDS 307

present at all development of Language, skilful in

retrieving losses and in rescuing, by means of useful
additions, inflections which are in danger, ready to
profit by accidents, prompt to extend acquisitions. It
is this which has given to English its triple possessive
pronoun , his, her, its, of which the Romance languages 0

possess no cquivalent. It is this which has cnriched

thc French conjugation with tcnscs unknown to Latin. i
It causes phenomena of utterly different origin to com
pete for one and the same goal. It infuses a signification
into syllables originally empty or indifferent.

We arrive in this manner at an extremely important

and delicate question : how large a share has intention
in the facts of Language ? Modern philologists in general
unhesitatingly reject the idea of intention. The utmost
that they will admit is, that accidents which have come
about inevitably and quite unexpectedly have been
utilised in a spontaneous and unconscious manner.
There is no doubt that in former days philologists
greatly abused the intentions ascribed to Language, and
attributed to it in detail all kinds of distinctions and
hidden purposes of which it is innocent. But the con
trary doctrine is no less far from the truth. Modern
philology seems to confound intelligence with reflection .
Though not premeditated, the facts of Language are none
the less inspired and directed by an intelligent will. Be
tween the popular act which on the spur of the moment
creates a name for some new idea, and the act of the
scholar who invents a designation for a recently dis
 308 SEMANTICS

covered scientific phenomenon , there is a difference as to

the promptitude of result and as to intensity of effort, but
none at all as to nature. On cach side the faculty
brought into play is the samc. It would be a grcat cx
aggeration to supposc thc agent to be, on the one hand,
Intelligent and free, on the other hand blind and un
conncloun.
Even that other, inoro inatcrial, department of phillo
logy which treats of sounds, L'honctics, for which , to
gether with the unconsciousness of physiological pheno
mena, the precision of mathematical laws is now clained,
is not of an absolutely different order, for it is the brain
as much as the larynx which is the cause of changes.
At least a distinction should be made between pheno
mcna which clcrivc from the structurc of organs and froin
an iinperious nccessity of pronunciation, and those which
come from the instinct of imitation and from mcre pre
ferences. Without dwelling longer on these consider
ations, we will merely say that they are passing exag
gerations of a genuine and excellent principle, to wit, the
regularity of the phenomena of speech. But we have no
doubt that philology, abandoning its paradoxes and its
prejudices, will become less unjust towards the prime
motor of languages, that is to say, towards ourselves,
towards the human intelligence. The mysterious trans
formation which has caused French to issue from
Latin, as Persian from Zend, and as English from Anglo
Saxon, and which exhibits on all essential points a strik
ing collection of similarities and identities, is not the
A mere product of the decadence of sounds and of the wear
 THE HISTORY OF WORDS 309

and tear of inflections. Beneath these phenomena, where

everything speaks to us of ruin, we feel the action of a
thought which stands out from and strives to modify the
form to which it is bound, and which often derives profit
from what at first sccm but loss and destruction. Mans
agitat molem .
 APPENDIX

THE SCIENCE OF MEANING '

By the selection of the Science of Meaning as the

subject of my address to-day, I find myself in a some
what singular position. I am like one who revisits, after a
lapse of almost twenty years, the land of his recollection
But the scene before me is not the one so commonly
seen on such occasions. I cannot point to a country
civilized almost beyond recognition, where flat marsh or
rolling prairie has disappeared before fields and orchards,
hamlets or homesteads, or even towns and manufactories.
The region on which I ask you to-day to interest your
selves with me is almost a waste. It is true that here and
there a spot has been cleared. Special contributions of
value or importance have been made in the works of a
number of inquirers, in those, for example, of Brugmann,
Bechtel, Heerdegen, Paul, and Sweet. In the particular
province or, if you like, outlying dependency called
Grammar, now, we may hope, finally rescued from the
thraldom of Logic, there has been great and fruitful
1 An Inaugural Address delivered on October 6 , 1896, at the
opening of the Session 1896-7, at University College, London,
published in the Fortnightly Review of September 1897, and
reprinted by the kind permission of Messrs. Chapman and Hall.
 312 APPENDIX

activity. Lastly, and this is most important, the

materials and appliances for research in this our subject
are greatly augmented and improved. But after all,
when we consider the actual total of results, we must
own, I fear, that the words of an inquirer who has aided
the study both with pen and purse do but too faithfully
rcficct the facts.
" I have found nothing," wrote Lady Welby, in a
letter to me at the end of August last, “ which con
tributed except indirectly, implicitly, or casually to the
study of ' Sense, Meaning, or Interpretation .'”
I should be wasting the precious moments of your
attention were I to enter into a disquisition on the causes
of this state of things. One question, however, rises
before the mind with such pertinence and such urgency
that it must be at once resolved. It may - nay, under
s
the circumstance it must be asked : Are there not good
reasons, as we call them, for the neglect which we
observe ? If the study has so signally failed to attract
inquirers, perhaps the explanation is that it is destitute
of importance , or that it is too difficult or too uncertain
to pursue.
In considering this question I shall, as I am in private
duty bound, begin with its aspects towards the subject
which I have the honour to represent in this College.
The investigation of meaning is, as I shall hope to
show, of considerable importance to other branches of
+ knowledge ; but to Comparative Philology it is vital.
Language is but articulated sound used as a medium
for the interchange of thought between man and man ;
 APPENDIX 313

and what philologer would care to determine whether t

is pronounced with the teeth or the palate, or how many
vibrations per second there are in the sound -waves of e,
if that must be the stopping -place ? No : to him these
researches into the vehicle of language are only auxiliary
to that of what it is intended to convey ; and to decline
the systematic investigation of meaning is to confess
that the so -called Science of Language is of its essence
unscientific . And here I will pause to answer a not
unnatural objection which may possibly have formed in
the minds of some. What,” it may be asked, “ is the
need of this new science ? Etymology was planted and
has thriven without it. Why should we not do as here
tofore, and content ourselves with the tracing of meaning
as it appears in each particular manifestation ? ” The
answer is easy. That in the first instance we must so
proceed, that further, by groping from one particular to
another particular, we may develop a sort of instinct
which will keep us from error in the main, may be
readily admitted. But what of that ? Is this to be an
argument against a more systematic procedure ? A
savage can find his way across a trackless brake ; long
before cartography was a science or the potency of the
needle known, mankind had traversed land and sea .
But this is no argument against roads and maps, against
the chart and the compass.
I will now ask you to turn with me to observe how
far these weighty considerations are recognized in the
practice of professed and competent philologers at the
present day. In doing so we will confine ourselves, as
314 APPENDIX

elsewhere in this introductory address, to the sufficiently

vast region of Indo -European philology ; that is, the
comparative study of that great group of languages to
which our own speech and those of most European and
several Asiatic nations belong. The vocal sounds of
these nations have been minutely examined and described.
The changes of these sounds one into another have been
classified and tabulated. Laws have thence been deduced
which are formulated and applied with almost mathe
matical precision. No exceptions are permitted. If one
appear, and there are still apparent exceptions, it is
instantly explained away. So much care is shown for
the form and clothing of language ; but how much for
its substance and soul ? The difference is seen in the
very terms employed when an unsatisfactory derivation
is censured. If it offend against the canons of meaning,
or, to speak more exactly, against the critic's conception
of these canons, it is called harsh, unsuitable, or im
probable. No reflective critic would employ a stronger
term. But let it offend against the generalizations of
sound -change, the mildest censure will dub it baseless
and arbitrary ; most commonly will it be branded as a
flagrant breach of phonetic law. And thc reason for the
difference is plain. In sound-change all the ground has
been examined ; the facts are known or believed , at
least, to be so ; criteria of truth and falsehood have been
determined. But in the case of meaning not one of
these conditions is present. Hence it is that human
fancy, that worst rebel against the sciences, driven from
the domain of sound -lore, runs riot in the region of
 APPENDIX 315

meaning, where it does its best to undo the work of

sober judgment and research in the territory from which
it has been excluded.
The importance of the science of meaning to Com
parative Philology I trust I have shown : of its importance
to other branches of science I shall have a word or two
to say anon. Its difficulty need not appal us. Since
the days of Prometheus nil mortalibus ardui est has
been the scientific watchword of mankind ; and I think
I shall be able to show that this, though difficult, is not
one of the most difficult branches of human inquiry.
There remains, however, another question, and that
the most momentous of all -- the question of the certainty
of its results.
We have seen that to impute uncertainty to the results
of the investigation of meaning is implicitly to impute it
to the results of the science of language as a whole.
And, as a matter of fact, the objection usually takes this
form . We shall best appreciate the nature and pertinence
of this allegation by considering a few examples. You
will forgive me if these are too familiar. Strange
contrasts and even contradictions are perceived in the
past or present meanings of words. The German

schlecht once meant " simple ” or “ right,” now it means

“ bad .” The original sense of the English silly was
happy.” The Greek oủ ndu ought to mean “ not
quite ” ; it means not at all. Nor are such discrepancies
confined to contrasts between the present and the past.
At this very day the English rather expresses a high or
a low degree according to circumstances, as "rather
316 APPENDIX

fine ," but in answers Rather ! And the same Latin

writer, Plautus, apparently uses procliuis (whence comes
the English proclivity ) now for "easy" and now for
“difficult." Again, it has been urged , as by Mr. Herbert
Spencer, that many words are derived from proper
names, and that the etymologist will go sadly wrong if :
he attempts to find cognates for these in other words, as
he will infallibly do if he is not acquainted with the
special circumstances of their birth . A favourite example
is the verb to “ burke," which comes from the name of a
murderer executed in 1829, but is now used meta
phorically and without any sense of its origin. It is not
surprising that, with such examples before them, the
critics of Comparative Philology have drawn the con
clusion that its conclusions were as a whole uncertain ;
it would have shown the rarest self-restraint if they had
refrained from so doing. But it is most unreasonable
to make these “ sports ” or eccentricities of language a
basis of argument unless we know what proportion they
bear to the whole. We have seen, or at least have
heard, of quadrupeds with five or more legs, of men
with six fingers on each hand. But how absurd would
it be to use these monstrosities to attack the propositions
that a quadruped is an animal with four legs, and that
the complement of human fingers is ten !
Professors of my unfortunate subject have occasionally
strange questions asked them. A gentleman, who wrote
from Ireland, but whose name was obviously German,
requested ine not so long ago to inform him how many
words there were in some four different languages, of
 APPENDIX 317

which English and German were two. I was unable,

for several reasons, to answer this modest inquiry : one
was that I did not ':now what he understood by a
“ word .” In “boxing the compass” and “boxing a
boy's ears” we have, according to the great English
Dictionary, two different words "to box ." In “ booking
to Edinburgh" (of a passenger) and " booking an account "
of a tradesman is there only one ? Certainly not - for
the science of meaning. For in use the words are as far
apart as if they had come from the most different roots.
What does this mean to our science ? It means that its
materials are at once multiplied we cannot tell how
many fold . For in each distinct sense of a word we
have a fresh fact for the science of meaning. And
" fact " is the right name to employ, for the derivation
of the sense is, in nearly every case, certain and clear.
But furthermore this does not represent, by any means,
the whole of the materials available for the study of
meaning in any language. A new word or, if you like,
a new differentiation of a word is not the growth of a
day ; time and custom, and the consenting usage of a
number of persons, are required before it becomes
“established" as we call it. Every human speaker and
writer, from time to time, uses words in another than
their proper sense, employing metaphor or some other
of the figures of speech. Of these figurative uses some
pass into common currency ; most do not. But all may
furnish materials for our student. With such stores to
our hands, how perverse it would be to begin with the
speech of remote ages, which has been imperfectly and
 X
318 APPENDI

often , we must fear, erroneously recorded, or to busy

ourselves with the conjectured senses of hypothetical
forms ! No ; the study of meaning should, like charity,
begin at home, and should, at least at the outset, confine
itself to the study of actual living language and well
attested literary documents. Let me fortify what I
have said by a reference to Mr. Henry Sweet's paper
on “ Words, Logic, and Grammar. ”
“ One of the most striking features of the history of linguistic
science as compared with zoology, botany, and the other so-called
natural sciences, is its one-sidedly historical character. Philologists
have hitherto chiefly confined their attention to the most ancient
dead languages, valuing modern languages only in as far as they
retain remnants of older linguistic formations - much as if zoology
were to identify itself with palæontology, and refuse to trouble itself
with the investigation of living species, except when it promised to
throw light on the structure of extinct ones.
“ Philologists forget, however, that the history of language is not
.
one of decay only, but also of reconstruction and regeneration .
These processes are of equal, often more, importance than those
by which the older languages were formed , and, besides, often
throw light on them. They have further the great advantage of
being perfectly accessible to the observer. Thus the growth of a
language like English can be observed in a series of literary docu
ments extending from the ninth century to the present day, affording
examples of almost every linguistic formation .”

It is clear, then, that the science of meaning has a

more than ample field for its investigations without
trenching on doubtful or disputed ground. Those very
paradoxes of use, to some of which I have referred,
which amuse the curious and alarm the timid inquirer,
have each of them its explanation. Schlecht did not
pass at once from “good” to “bad.” Of words, as of
 APPENDIX 319

men , it is true nemo repentefuit turpissimus. The change

took time ; it was a journey of stages, all of which can
be traced. Rather, " somewhat, " is not the same word
as rather / " very ." Though they are spelt alike, their .

sound, as well as their sense, is different. These and

other similar examples are not proofs demonstrative of
the uncertainty of our science ; rather are they helps to
the attainment of certainty, as they warn us of the limits
outside of which it may not be found.
Having now completed our vindication of the science
of meaning as a part of the science of language, we may
turn to consider it as a pursuit - what it can teach us
and how it may best be pursued .
Throughout the whole history of the human race there
have been no questions which have caused more heart
searchings, tumults, and devastation than questions of
the correspondence of words to facts. The mere mention
of such words as " religion ," " patriotism ," and " property "
is sufficient to demonstrate this truth. Now, it is the
investigation of the nature of the correspondence between
word and fact, to use these terms in the widest sense,
which is the proper and the highest problem of the
science of meaning. That every living word is rooted
in facts of our mental consciousness and history it would
be impossible to gainsay ; but it is a very different
matter to determine what these facts may be. The
primitive conception is undoubtedly that the name is :
indicative, or descriptive, of the thing. From which it
would follow at once that from the presence of the name
you could argue to the existence of the thing. This is
320 APPENDIX

the simple conception of the savage, who cannot under

stand a name which is not significant. There is the
well-known story of the Indian who refused to believe
that the name of Colonel Short could be applied to a
person of that officer's unusual stature . This view is at
the bottom of those superstitions, as we now call them,
which an und so frequently in ancient and, for that
matter, even in modern times. Thus because penthos
means " mourning ” in Greek, Bacchus in the play of
Euripides tells Pentheus that his name is the right one
for misfortune. Such fancies may be sport for the
humourist, as in the adventures of A. Mishap which the
Sea Pie made so diverting ; but they have exercised ,
nay, with many they still exercise, a most potent
influence.
This primitive view of a real connection between the
name and the thing has by no means entirely disappeared
even from the select circles of scientific inquiry. There
it takes the form of an undue exaggeration of the direct
testimony of language. The information which Language
will give us directly is often either superfluous or mis
leading ; it is that which may be indirectly extracted
from her which, in most cases, is alone of value. We
may in fact compare her to what the lawyers call a
hostile witness who will only give up the truth upon
cross-examination. Let me, as before, illustrate by the
simplest of cases : it will be in the present instance a
proper name. Suppose a person called James Johnson :
what facts could a student of language infer ? In this
case he could, as it happens, draw from the first name a
 APPENDIX . 321

direct inference as to the person who bore the name;

he could correctly deduce the fact that this person was
a male. But what about the second name ? Could he
deduce from it that James Johnson was related by kin
ship, nearly or remotely, to other Johnsons ? Certainly
not. James might have been adopted ; he might have
assumed the name for reasons of his own. Could he
conclude that James was an Englishman or even a
member of the Anglo-Saxon race ? Again we must
answer, No. Johnson is a common negro name. So he
could not even tell whether James was a black or a
white. But what could he deduce indirectly ? Не
could deduce the existence of a progenitor called John,
from whom the family of Johnsons derived their descent.
And since John, like James, is the name of a male,
he could infer that in the community to which these
Johnsons belonged descent was traced through the
paternal, and not, as in some others, through the maternal
line. These deductions are of no value so far as our
knowledge of the individual James Johnson is concerned ;
but they might be of very high value for our knowledge
of the family or the community with which he is by
name associated.
There is a fascinating branch of philological specula
tion which has been dignificd with the titlc of Linguistic
Palæontology. In Indo-European philology its origin
ator was Adolphe Pictet, an ingenious but capricious
Frenchman ; its leading exponent now is O. Schrader,
a German of far greater learning, judgment, and sobriety.
The besetting sin of this form of philological activity,
Y
322 APPENDIX

the value of which within its proper limits I should be

one of the first to maintain, is its proneness to false
estimates of the direct evidence of language. Thus it
has been argued from the vagueness, the insufficiency,
and the apparent inconsistency of the expressions for
colour in Homer, that the poet or even the whole of his
contemporaries were colour- blind . The inference is un
warranted. The fact that the Homcric Grecks have no
cxpression for green does not prove that they could not
scc the colour, but that they did not want the word.
Thcrc was no name for magenta until the battle was
fought and thc dye discovered ; but who would venture
to assert that it was not till 1859 that the human cyc
becamc capable of distinguishing the colour ? If these
inferences are hazardous when, as in the case of Homer,
we have a real language for our foundation, what must
they bc when the language is itself a creature of infer
enccs ? To construct the Indo -European civilization,
except in thc merest outline, from the Indo -European
language, is to build with cards upon the sand.
May I quote some words of Professor Schrader in his
work The Prehistoric Antiquities of the Aryan Peoples,
with which I thoroughly agree ?
“ Comparative Philology of itself is not in a position to recon
struct the primitive civilization of the Indo -Europeans, and if we
are to secure our advance step by step over this difficult ground,
we can only do so on the condition that the three sisters, Linguistic
Research, Prehistoric Research, and History unite in the common
work." (English translation , p. 149.)
Before I leave this portion of my subject, I will
endeavour to indicate by an example how the method
 APPENDIX 323

of indirect reference may be employed to throw light

upon obscure difficulties of language as well as upon the
circumstances of times of which no direct knowledge is
possible. Of all the apparent contradictions in language
that I have met with, there is none at first sight more
inexplicable than the usage of procliuis to which I have
already adverted . In Plautus, the oldest Latin author
of whom we possess considerable remains, procliuis
(lit. “ sloping ”) is clearly used in two opposite senses :
"easy " in the phrase in procliui est, “ difficult ” in the
expression “facere rem planam ex procliui," " to make a
matter level,” that is " easy," instead of " sloping," that
is “ difficult ." The apparent contradiction is a grave
onc. If, under the same circumstances, in the same
period of a language, the same word may be used for
two such opposite ideas, we might well renounce the
investigation of meaning as a serious study. Yet every
avenue of explanation seems shut. There .can be no
question of irony or euphemism , through whose work
ings the apparent meaning of an expression is so often
inverted. The phrase is simple and straightforward.
Nor, again, can we take refuge in a hypothetical sense
for procliuis, from which we might derive the meanings
of “ difficulty " and " ease ” by different routes. We
cannot say that the word once meant “ steep," and that
this meaning was retained in certain phrases where
difficulty was to be expressed , while in others it was
weakened into that of " sloping," " gently sloping,"
" easy ." The original sense of procliuis is incontest
able ; it means " sloping forward " or "down "; cliuus,
324 APPENDIX

like its English derivative " declivity," which still bears

the imprint of the original meaning, means a slope, an
ascent up which a horse can be driven . Ovid, referring
to a common Latin proverb, says, “ my horse pants at
the beginning of the slope." There is no avoiding, then,
the conclusion that the word in both these phrases
meant originally " downhill.” That “ downhill" should
come to mean " casy ” nccd surprise no one ; the meta
phor is a common and a very natural onc. But its usc
for “difficult " is certainly astonishing. On cxamining
this use we remark two things : first, that, unlike the
other use, it is not found after Plautus, from which we
conclude that it belongs to the older stages of the
language ; secondly, we observe that procliuis, “ down
hill ” in the sense of " difficult,” is always opposed to
(and contrasted with ) planus, that is, to " flat" or " level,"
from which we infer that the proper application of the
use was to vehicles , and that it dates from a period
which cannot be very much anterior to Plautus , when
Roman waggons and carriages had no efficient brakes.
It will have been gathered from what has been said
alrcady that we may expect to learn much more from
language about persons than about things ; that a word,
when interrogated, will not, in general, tell us much
about that of which it is used, but a good deal about
the people who used it. And this leads us directly to
the chief of the external uses of the science of meaning
-the supplying of materials to the student of human
thought. That language is a mirror of man has long
been recognized by Icading authorities in mental philo
 APPENDIX 325

sophy and psychology. In these sciences direct observ

ation is impossible. The workings of our thoughts and
feelings cannot be examined, they can only be inferred ;
and language, the expression of those thoughts and
feelings, furnishes one of the most important modes of
inference. Now, if we find an individual applying the
name of anything whatever to something else, it is an
inevitable conclusion that there is some similarity or
connection between the notions whose expression is
thus associated . And if wc find this to be the case in
a number of cases, that is, in the general usage of a
word or a group of words by a community, we may
indeed we must - further conclude that this association
is no peculiarity of the individual, but that it is the
product of something in the mental configuration of
that community, or even of the human race. The merit
of this evidence is that it is entirely unconscious, and
therefore, so far as it goes, absolutely reliable. The
mind may have been misled by an unreal or fallacious
resemblance in things : that very error is a fact for the
savant. It may have been moved by a desire to gloss
over and disguise an unpalatable idea : that repugnance,
that evasion, are duly noted and recorded by the psycho
logist. After the picture which I drew at the beginning
of this address of the present condition of the science
of meaning, I should involve myself in flagrant incon
sistency if I referred to the finds that might be expected
from this almost virgin soil otherwise than quite tenta
tively and with the greatest reserve. But the importance
of the subject is such that I cannot bring myself to say
326 APPENDIX

absolutely nothing upon it. I will therefore mention

two or three remarkable tendencies of language in
general upon which we may hope to have more light
hereafter

It is not strange to find “ concrete " expressions used

for " abstract," e.g. it seems quite natural to say “ this is
the mark of a good man " instead of " a mark of good
ness.” But what can be the reason for the converse
substitution ? Why should Latin , a language which is
unusually prone to the concrete, use seruitium and
seruitus, “slavery,” for “ slaves ” ? This usage, which
is by no means confined to any onc language, sccms
to indicate some relation bctween concrete and abstract
ideas, the nature of which has not been fathomed.
Again, a number of the languages possessing the definite
article place it when used with certain adjectives in a
different position to the one which it usually occupies.
One of these adjectives is all. Why do we say "all the
land ," but “the vast land," "the whole land " ? This
might be thought an accident of English expression but
for the fact that in Greck and in the Romance languages
descended from Latin (which, as you know, did not
possess an article) the same position has been developed
independently . Lastly, in the evolution of the powers
of the mind a very important role has been assigned to
the sensations of the muscles, by which are meant the
sensations of effort and strain which accompany move
ments of the body, the support of weights, and the likc.
Now, it has interested me to observe that a very large
number of the words which language uses for tninking
 APPENDIX 325

sophy and psychology. In these sciences direct obsery

ation is impossible. The workings of our thoughts and
feelings cannot be examined, they can only be inferred;
and language, the expression of those thoughts and
feelings, furnishes one of the most important modes of
inference. Now , if we find an individual applying the
name of anything whatever to something else, it is an
inevitable conclusion that there is some similarity or
connection between thc notions whose expression is
thus associated. And if we find this to be the case in
a number of cases, that is, in the general usage of a
word or a group of words by a community, we may
indeed we must - further conclude that this association
is no peculiarity of the individual, but that it is the
product of something in the mental configuration of
that community, or even of the human race. The merit
of this evidence is that it is entirely unconscious, ang
therefore, so far as it goes, absolutely reliable. Thu
mind may have been misled by an unreal or fallaciou
resemblance in things : that very error is a fact for th
savant. It may have been moved by a desire to glo.
over and disguise an unpalatable idea : that repugnan
that evasion, are duly noted and recorded by the psych .
logist. After the picture which I drew at the beginni.
of this address of the present condition of the scier
of meaning, I should involve myself in flagrant inc
sistency if I referred to the finds that might be expec
from this almost virgin soil otherwise than quite tei
tively and with the greatest rescrve. But the importa
of the subject is such that I cannot bring myself to
 ot
is

re

or
an ,
not
en
, be
nifi.
ther
сcial
It
.

bject.

Prm is
deals
 APPENDIX 327

are derived from words such as " ponder," " to weigh ,”

which express or imply the presence of muscular play.
There is yet another use to which the methods of the
science of meaning may be applied. It is one replete
with interest and not devoid of profit, though it cannot
claim the importance of those which have just been
mentioned. I mean the deduction of individual and
national character from an examination of their language.
How far it would be worth while to investigate character
by means of its reflcction in specch we are not in a
position to judge, as this matter has never been properly
investigated . But the observation that national character
is rcfccted in national speech is one which cssay writers
have taken for their own, and I fear that now it runs the
risk of being considered too hackneyed to be important.
I will therefore confine myself to reading a passage from
Marsh's Lectures, which will admirably illustrate the
method employed. You must not suppose that I
necessarily subscribe to all the conclusions of Landor.
' I think the language of Italy is a case in point. Landor argues
the profound and hopeless depravity of the Italians from the abject
character of their complimentary and social dialect, and the phrase
ology expressive of their relations with their rulers or other
superiors, as well as from the pompous style by which they magnify
the importance of things in themselves insignificant, and their
constant use of superlatives and intensives with reference to trifling
objects and occasions. Were it true that the Lombards, the Pied
montese, the Tuscans, and the Romans of the present day had not
inherited, but freely adopted, the dialect of which Landor gives
a sort of anthology, it would argue much in favour of his theory.
A bold, manly, and generous, and truthful people certainly would not
choose to say " umiliare una supplica ,” to humiliate a supplication ,
for, to present a memorial ; to style the strength which awes, and
328 APPENDIX

the finesse which deceives, alike “ onestà ,” honesty or respecta

bility ; to speak of taking human life by poison, not as a crime, but
simply as a mode of facilitating death, " ajutare la mortc" ; to
cmploy " pellegrino," foreign, for admirable ; to apply to a small
garden and a cottage the title of " un podere," a power ; to call
every house with a large door " un palazzo," a palace; a brass
carring, una gioja,” a joy ; a present of a bodkin, " un regalo," a
royal munificence ; an alteration in a picture, " un pentimento,” a
repentance ; a man ofhonour, “ un uomo di garbo," a well-dressed
man ; a lamb's fry, “ una cosa stupenda," a stupendous thing 66; or a
message sent by a footman to his tailor, through a scullion, una
Ambasciata ," an embassy.' (Marsh, Letters on the English
Language, Lecture X., from Landor's Imaginary Conversations,
Sccond Scrics, No. V.)
I now come to the last division of my subjcct - thc
prosent and urgent nccds of the Scicncc of Meaning.
In the first place it lacks a terminology. An attempt
was made to provide it with one in some posthumous
articles which appeared in the earlier volumes of the
Journal of Philology 1 from the pen of the late Professor
Grote, the able brother of the inore famous historian,
which I mention here because apparently they are not
as well known as they deserve to be. Grote's nomen
clature is, however, too cumbrous and repellent to be
generally adopted. I lay no stress on the fact, signifi
cant though it be, that, unlike philology and its other
branch phonology, it is itself at present without a special
designation and must be indicated by a periphrasis. It
is a matter so much more scrious that its very subject.
matter is unnamed.
I have hitherto called this words. But the term is
unsuitable. The objects with which our science deals
* IV. pp. 53 foll., 157 foll., V. pp. 153 foll.
 APPENDIX 329

are the expressions of ideas or notions, with their con

nections and qualifications, in speech. These may be
and commonly are words, no doubt ; but they need not
be. They may be groups of words or phrases ; they
may be tones, stresses, or pauses of the voice. In
English, for example, as no doubt you know , an asser
tion may be changed to a command or to a question by
a change in the modulation of the voice. New titles
then arc rcquired, and I propose to take them from the
Greek , that storchouse of scientific terms, and to call
the expression of a single idca or notion a rheme,
from prua, " a thing said ," and to distinguish the expres.
sions of qualifications and connections of such rhemes
by calling them epirrhemes, though, as a general term,
Theme may serve for both. If these terms be approved
of, I should propose to call our science Rhematology, or
the study of rhemes. Thus, in “ hands off the table ,”
we may say that there are four rhemes (more strictly
three rhemes and an epirrheme), hands, the-table, off,
expressing the notion of removal and the tone of the
voice which conveys the speaker's will. But in " to lay
hands on the plate" there are only two, because “ to
lay hands on ” stands for a single notion. My little
daughter said the other day, " I know three new words
-scandalous, Matthew's man, and pretty creature.”
She said " words," but she meant "rhemes " ; for to her
the noun with the genitive in the one case and the noun
with the adjective in the other are single notions, though
for us they are double. I take the sense of singleness
or unity in a conception as the test of singleness in a
330 APPENDIX

rheme. By this I mean that it must present itself to

the mind as such ; that the mind must feel it as single.
I know of no other criterion. But it is by no means
easy to say in a given case whether this sense of unity
or singleness is present. It will be one of the most
delicate tasks of rhematological inquiry to follow and
ascertain the origin and growth of this unifying sense .
The contention that in " laying hands on ," whether
applied to a burglar or a bishop, there is only one
rheme, will probably be a shock to some ; but all will
admit it in the case of the Greek επιχειρείν, « to put hand
to " or " undertake," a word of similar origin. And the
rcason is only this, that in the Greek word we are
further from the point of fusion, and that the complete
ness of this is indicated in the external form ; for
&TIXelpeîv is a single word, and “ lay hands on " three.
I must now say a word or two upon the contents of
rhemes. I shall speak here only of rhemes which are
expressed by the same sounds, or, as we say, by the
same words. These often differ from one another as
much as the very different persons who bear the same
name.

“ Strikes the clock ? The hour is gone.

Strikes the man ? The day is won ."
In these two lines strikes is used first for as simple a
rheme as can be found, " the impact of a sound -wave
on the ear," and in the second for one of very great
complexity, which perhaps you will excuse me from
analyzing.
In the composition of rhemes we may distinguish the
 APPENDIX 331

principal or central portions from the outlying and acci

dental portions. Any name will illustrate this distinction.
Ask a dairy -maid and a New York belle what they
understand by a cow . It will be impossible to miss the
constancy of the central portion of the rheme. But say,
“You will meet a cow round that corner," and observe
the effect. You will as clcarly appreciate the variation
in its accessory portions. Here, as elsewhere, you will
observe that the central portion of a rheme enables us to
identify it, and the accessory portions to reason falla
ciously about it.
If the differences of conception are so great, and, by
consequence, the field of misapprehension so large, in
rhemes whose centre is strongly rooted in a physical
sense, what must this be in rhemes which have no
such attachment ? If " cow ” excite mental feelings so
different, what may we expect for “virtue " ? Let the
history of moral philosophy give the answer to this
question. Thus much is clear, that as we rise higher
into the reasons of abstraction, there is a tendency for
the central, the constant part, the core we may call it, of
a rheme to shrink and diminish ; while the accidental
extraneous and fluctuating portions spread and increase,
so that it reminds us of nothing so much as a comet,
whose tail is out of all proportion to its head. Some
say (the matter is far beyond our present scope) that some
times the nucleus disappears altogether, and all that is
left is the mere sounds which constitute the name and
a shifting, shapeless nebula of associations, as unsuitable
to be the objects of knowledge as is a will-o'-the-wisp
 332 APPENDIX

itself. Such rhemes are those meant when it is said that

men have often given their lives for a name. I have
dwelt with some emphasis on this topic because it has a
special bearing upon our own language - this modern
English, of which we have many reasons to be proud.
The looseness and ambiguity of English expression is
well known both to men of letters and statesmen. That
great and statesmanlike writer, the late Sir John Sceley,
once observed to me in conversation, that this was so
great as to make it a most difficult matter to draft a
treaty in English. For my own part, I must confess
1
that as a vehicle of clear expression I prefer Latin, in
spite of its inherent inferiority to my native tongue, and
I shall be only too happy if to-day I have not given
another illustration of the striking words of R. L. Steven
son : " Do you understand me ? God knows, I should
think it highly improbable."
We sec now how great to our science is the importance
of lexicography. The lexicons and dictionarics in most
languages are far from perfect ; but we have a few
models to point to. Such are Littré's great dictionary
of French, and that of English by Dr. Murray and Mr.
Bradley now in progress, to which it is the duty of every
Englishman who can afford it to subscribe. A model
lexicon will provide us with a complete biography of
every word. It will. give all its senses, illustrated by
well-chosen and significant examples. It will distinguish
carefully between the usages of poetry, set prose, familiar
conversation, slang, and vulgar speech. It will exhibit
them arranged in order, and, as far as possible, in the
 APPENDIX 333

natural order of development. (I may observe in pass

ing, that in this respect the Latin lexicon of Lewis and
Short is superior to the Greek one of Liddell and Scott.)
Above all, it must never lose sight of chronology. The
date of the first appearance and, if possible, of the
last appearance of a usage must be recorded. Here the
New English Dictionary has set an example which the
lexicography of other modern and ancient tongues will
do well to follow . A special department of lexico
graphical study which may be separately pursued is the
discrimination of synonyms. Much good work has been
done here in many languages. In Greek I may mention
the researches of Dr. J. H. Schmidt.
A cognate subject is the one to which I have already
made reference -- the investigation of metaphors and
figurative language. This also may be pursued apart.
There is no stage of language which the student of
the science of meaning may ncglcct. He must have no
prejudiccs, no prcdilcctions. But he will find that for
certain purposes some stages and forms of languages are
more instructive than others. Foremost among these
may be placed the languages of savages. The great
propagandist societies and their devoted emissaries have
earned the imperishable gratitude of humanity by all
that they have done to rescue from oblivion these perish
ing forms of speech. But it is not enough to know what
is the word (that is, the sounds) which a savage language
uses for a certain idea : we must know how he uses it,
and why. In most of the dictionaries of savage lan
guages that I have seen there is little help given towards
 334 APPENDIX

determining the original sense of a word. What is done

here must be done quickly : savage habits of thought
can only be properly interpreted by actual intercourse
with savages ; and these races and their languages are
rapidly disappearing.
The rustic must excuse me if I mention his speech
next in order to the savage's. There is ample material
for the study of bucolic cxpression in the numerous and
copious dialects of English which are now about to
rcccive a fitting home in the dialect dictionary of Dr.
Wright, the publication of which has already begun.
The dialects which have esulted when one race or

nation has adopted the speech of another have a special

importance. Such lingos as Negro -English, Negro
French , Pigeon-English, have a value for us out of all
proportion to their literary merits. We can gather from
them how far the mental stock of the two races coincides
by observing which portions of the total stock of rhemes
have been retained , which rhemes have been rejected as
too strange or dropped as superfluous.
The language of children , for specimens of which I
may refer you to the writings of my distinguished col
league and friend, Professor Sully, must also be pressed
• into the service. Its almost savage primitiveness and
simplicity throws light upon much that might otherwise
be obscure.
Nor, lastly, must the diseases and imperfections of
language be neglected. The study of the phenomena of
Aphasia, or the destruction of the power of rational
speech in all its forms and stages, has long been recog.
 APPENDIX 335

nized as an adjunct to psychology. To us even its

slightest manifestations, what are called mere slips of
the tongue, are worthy of serious attention .
And now I come to the last but the greatest and most
urgent need of our science. When we survey it, we sec a
subject of almost limitless extent, but of workers not a
score. “ Apparent rari nantes in gurgite uasto." " A
gulf immcnsc, a swimmer here and there." I hope, I
trust that this state of things will not last much longer.
I trust that nothing I have said to-day will contribute
towards its continuance. I feel sure that I have not
underrated the difficulties of the subject. I fear only
that I may have given you an exaggerated impression of
their magnitude. It behoves me then to say, that neither
in itself is this the most difficult of studies, nor is it
now in its most difficult stage. It does not demand the
rare and intense thinking of the higher mathematics, or
the facility of invention and delicacy of manipulation
which are essential to success in some of the physical
sciences. It is as yet in its beginnings ; its prime need
is the collection of facts. The collecting and arranging
of facts does not require the highest energies of the
human intellect. Industry, care, and circumspection it
does require; and these no subject can dispense with.
This is a study in which, if analogy is of any value as an
argument, Englishmen may expect to excel. It is also,
as I have said, one that has a special value for English
men. There is no need to travel to remote climates or
distant ages. It is a small disadvantage to be unable to
do so. The student of this subject must begin with
 X
336 APPENDI

what he knows. In your own speech, in that of your

friends, in the monuments of your native language - here
you are on familiar ground. Your facts can be checked ,
your conclusions verified. The experience so gained
will grow into an aptitude which will guide you safely
through regions where the way is darker, the access
more precipitous, the foothold less secure. The labour
will be its own reward. I do not know if in anything I
have said to -day there is sufficient hint or indication of
the great and varicd attractiveness of these studies ; but
this I can honestly asseveratc—that I have not con
sciously turned out of my path to paint an alluring
picture ; and I therefore feel myself free to conclude,
with expressing my deep and sincere conviction that
those who venture upon the exploration of these regions
will find themselves well repaid .
J. P. PostGATE.
 INDEX

ABLATIVE absolute, survival of the, | Avis, Latin words containing, 137

53
Ablaut, 58 Bourgeois, 105
“ Abstract and concrete ” Briller, 118
themes, xlviii Brugmann, K., on gender, xiv
nouns, 134 , 245 Buhle, 101
Abtreten , 193 Busse, 113
Accabler, 119 But, 202
Accusative, various significations of
the, 221 Calvados, 175
Adjectives become substantives , Cases, deterioration of the signifi.
298 cative value of, 196
Adolere , 153 ; why Latin cases have not
Adulterium , 111 passed into Romance languages, 16
Adverb as part of speech, 181 Cedere, 192
Adverbs, Latin, in c, 87 Cerberus, heads of, xvii sqq.
Æger , 94 Chaire, 283
Aflicted , 102 Children, language of, 334
Aigre, 130 Cicero, pro Milone, $ 54, xxix
All and whole, article with, 326 Civitas , 135
Amant, 101 Classis, 136
Analogy, xlv, 60 Cliens, 95 , 104
Antiquity of languages, what is to Common and proper names in one ,
be understood by, 88 xxxvii
Anxious , 103 Comparative, the Latin , in Romance
Aorist, extended by analogy to all languages , 13
moods, 75 Compelle, 103
Aphasia, lvi, 334 Compounds, 155
Ardea , ardea , xxxviii Concretion ofmeaning ,134
Aristophanes, Birds, 1313 899., X Conjugation , French , 66
Clouds, 659 sqq ., xi Germanic, 24
Armare, 121 Greek, 61 , 90
Article, origin of the, 208 Conjunction as part of speech , 185
Articulated groups, 166 Consules, in
•Alre, I'rench words in, 42 Contagion, 200
“ Attraction,” xxxii
Audire, 95 Dame, 92
Augment, modifying adverb or pro. Danger, 142
noun , 74 Declension, Greck, 63
Avis, French derivatives of, 95 Latin, survival of the, 50
ጊz
338 INDEX

Declension in the modern languages Gagner, gain , 116

of India, 217 Gas, 174
Defendere, 153 Gemma, 125
Defunctus, 152 Gender and sex , ix sq .
Delbrilck's classification of roots, Genitive, with verbs, 197
xlix English , 22
Delovi, 73 Germanus, 300
Dé!url, 230 Gerundive, Latin, 45
Dissiller , 290 French , usage of, 271
Deterioration o : meaning, 102 Getreide, 110
Differentiation , law of, 27 Giants, iheir snaky character, xxi
Duntaxat, 169 ! Goshtha, 120
Grammarians, French , 269
-ei, the German suffix, 68 Groups, articulated , 166
.eln , German verbs in, 43
false inflection of the plural in | Habiller, 304
English, 56 Hart, hind, liv
Enallage, xxxi llendiadys in sentences, xxxiv
Entenire, 153 Herr , 104
Epices, 109 llesiod , Works and Days, 242, xliv
Epirrhemes,xxviii,and Pref. passim , Homu, 114
329 Horace, Carm . II. 13, 34 , xvii sqq.
•cr , false inflection of the plural in Carm . III, 11, 15 sqq.,xvii sqq.
German , 57 Hydra, xx
Er and sie,German pronouns, origin . | Ilyperbaton, xxxvi
ally formulas of politeness, 104 Hysteron proteron, xxxiv
Erudire, 125
E.rainen , 144 Nilioms, 294
Expansion of meaning, 115 jcrin , German verbs in, 82
Exstinguere, 125 Inıperative, 235
Extinction , 101 Industrius, 131
Infinitive as new acquisition , 79
Facio, 110 Inflections, survival of, 50
Faclio, 110 false, 56
Facullas, 135 Inversions, 54
Fast (German ), 231 Inverted Sid KOIVOû construction ,
Fatigo, 102 xxxvii
Fec ( English ), 117 Invitare, 103
Felis, 108 Irracliation, 39
Felix, 73 ish, English verbs in, 68
tienum , 109 Italian langunge, Landor on , 328
Ferme, 232
l'igurative Inngungo, 333 " from
Johnson ," what may be inferred
l'igures of speech, *** nninc, 331
Fingere, 174
Flare, derivatives of, 96 Lacerlas, 293
Frau , 92 Langungens rcſccting individual
Frazer, J. G. , on gender, xv n. character , 327
Frequentative verbs replacing simple Latin use of abstract nouns, xlix
verbs, 95 Law , what is to be understood by, 11
Fructus, 135 Legio, 136
 INDEX 339

Lexicography, 332 Order of words, 212

Liberi, 149 Ordo, ordiri, 126
Linguistic palæontology, 323 Osthoff, H. , on anomalous forma.
Lisi (German ), 100 tions, liii
Livy, 26. 19. 8, XXX Ovid, Ars. Am. 1. 145, 2. 254.
Logic of language, 219 xxxii
Lucan, 7. 9,xl Trist. 3. 7. 47, xxix
8. 316 sqq., Xxxiii
Lucre, 96 Pune, 232
Luna, xii Participle, middle participle in -mini
Lustrare, 119 extended to all tenses, 74
Luxus, 124 past participle in French, 203
Parts of speech, 181
Mactare, 153 Passive, as new acquisition, 84
Magis , 14 Paunia , 116
Maitresse, 101 Pejorative, the so - called tendency ,
Manifeslus, 157 100
Mansio, 136 Pentheus , 320
Martial, 7. 5. 5 sq ., xxix Periculum , 100
Malurus, 145 Personifications, sex in, xvi
Neditor, 196 Pitere, 191
Meliorative tendency , 102 Plato , Theutetus, p . 151 c, xxxü
Memini, mens, 31 Plautus, Pscudolus, 35, xxxviii
Mentiri, 100 Pliny, N. H., 32, § 144, xxxiv n .
Metaphor, 122, 288 Plonger, 118
Middie Futures from Active Presents Pluril " ungrammatical " after with ,
in Greek , 1 cum , etc., xxvii n .
Migrare, 152 Mus, 14
Ninne, 101 Polysemia, 139, 287
Mortuus, 69 Porcus, porca , xii n.
Muer, 283 Portio, 136
Musculus, 293 Post, 69
Muth, 112 Pradium , 115
Mythological concepts not different | Prastare, 73
from other concepts, xix Prætor, 111
Mythology and science of meaning, Pregnant words, 151
xvi, xxiv Preposition as part of speech , 183
Prepositions, why they have replaced
Names, proper, 176 cases , 16
proper in Greck ,' 158 l'resent tcnsc, periphrastic and
Nognilon, ncgntivo words in French, simple in Mod. English, llii
200 l'roclivis, 325
Neuter nominative and nccusative, I'ronoun ns part of speech, 187
xxviii reintivo, 205
survival of the Latin, sa l'ropertius, peculiar use of synonyms
Noctu , 69 in, xxxix
Novellee, 150 Prostratus, 73
Prude, 101
Obliviscor, 71 Purity of language, 255
Omnes , 73 Putere, 123
Oppido, 231 Putzen , 124
340 INDEX

Quamvis, 168 Sweet- heart, 45

Que, for " or,” XXV
Õueen and qucan, different words, Tegmen, 106
xliii
Templum , words containing, 127
Tempus, 117
Rather, 315 Thucydides, 2. 54. 3; xliv
Reduplication , replaced by the aug . Tibullus, i. 3. 71 sq., xxiii n .
ment, 75 Totus, 92
Regina, 92 Tourmenter, 102
kegio, 136 Traire, 282
Religens, religiosus, 33 Tranquillitas, 125
Restriction of meaning, 106 Transitive force, i89
Rhcmatology, lvii Tribreries, III
Rhemes, xxviii, and Prcf. passim , Triumvir, 157
329 Truncus, 144
classification of, xlix Turda , xv
Rità , 138 Twain, Mark , on gender, ix
Rivalis, 127
Runagate, 261
Rutherford, W. G., 1 Ulmus marita, xiii
Umlaut, 58
-sal, .selig, German suffixes, 45 Unilerstand , 193
Savages, language of, 334 lingeführ, 23 !
Schlecht, 315 Ungrammatical constructions,
xxvii
Science of meaning, methods of, Urbs, 113 .
Ivi
• sco, Latin verbs in, 39 -urio, Latin verbs in , 40
Scrupulum, 117 Vadium , 115
Scylla, xxi
Sehr, 103 Valerius Flaccus, 3. 227 sq ., xxiii n.
Senex , 94 Valetudo, 100
Sentio , 33 Veneror, 176
Scquor, 196 Venus, 176
Si, origin of conditional iden, 203 Verbs, active and neuter, 189, 301
Silly, 100 auxiliary, 24, 210
Slips of the tongue , 335 Verslchen, 193
Smart, 102 l'estis, 135
Solidus, 117 l'occ081, 102
Sophocles, Ajax, 841 1940, Tvill n . l'indomia, 119
Clipas Colonciis, 1321.2, Vocatives, irrational, xxx1
xxxiii ; 1685 899., xxix Volupe, volumas, 32
Sound and mcaning, inter - action of,
xlv s99. IVetten , 115
Spatium , 118 Wheeler, B. I. , on gender-endings,
Specialisation, law of, 11 xiv
Species, 109 Witz, 112
Spes, spes, xxxvii, 33 " Word, a ," meaning of, 317
Splendere, 119
Subjective elements, 229 Zeugma, xxxi
Sweet, II. , liji
 INDEX M1

INDEX OF GREEK WORDS

άλογο », του κτήματα, 109

άνειρα , 92
ανήρ, άνθρωπος , 34 nówv, leo, for lioness, xii n .
αρχαί, 30 λιμός and λοιμός, xliv
• & tos, the suffix , 65
μανία, 31
βαίνω, βίβημι, βάσκω, 38 μέμνημαι, 31
βουκολίω, 120 μίνος, 31
βυσσοδομεύω , 131 μηχανάω, 131
μιμνήσκω, 31
γέρων, 94
πειρατής , 109
εκατομβή, 120 πλείον , 14
«ίκειν , 192 ποιητής, 152
ελπίς, ελπομαι , 32 πονηρός, 33
εμποδών , 70 πόρος , 127
επίσταμαι , 193
επιχειρείν, “ lay hands on ," 33ο reis or docis, not ducis, in modern
Greek, xliv
ηδονή, 32 σοφίζω, 128
στοιχεία, 30
θα , 83, 167 συμφωνέω, 133
-1dw, verbs in, 41 τάλας, 32
έκνέομαι , 191 oratos, thesuffix, 65
τείνω, τιταίνω, τανύω , 38.
•κα , perfect in, 47 τλήμων, 31
κάμνω , 103 τολμάω, 31
κατά, 20
κεφαλή, caput, apparently for | φεύγω, φεγγάνω, 37
plural, xxiii n. φράτωρ, 15ο

Richand Clay & Sons, Limited, London & Bungay.

 STANFORD UNIVERSITY LIBRARIES
CECIL H. GREEN LIBRARY
STANFORD , CALIFORNIA 94305-6004
( 415 ) 723-1493

All books may be recalled after 7 days

DATE DUE

2 1937

SEP 24 2001
20301
AUG
1
 P 741 .B813 1900 C.1
Semantics :
StanfordUniversityLibraries

3 6105 039 190 017

DATE DUE 17 1998

JUL 15 1985 JUN 1990

JUNE 1992

STANFORD UNIVERSITY LIBRARIES

STANFORD , CALIFORNIA 94305
564,92K
0473
PAS067973
9429 "
13 1900 C.1
:
ord University Libraries

05 039 190 017

DATE DUEL 17 1988

51985 JUN 1990

32

D UNIVERSITY LIBRARIES
DRD, CALIFORNIA 94305
 STANFORD UNIVERSITY LIBRARIES
CECIL H. GREEN LIBRARY
STANFORD , CALIFORNIA 94305-6004
(415 ) 723-1493
All books may be recalled after 7 days
DATE DUE

4E2 1337

SEP 02,4 2001

231 001
AUG