Digitized by VjOOQIC



282 The Indo-European Organism,

forms, and similar facts may be reckoned by thousands
in all the Indo-European idioms. But, as I have said,
each language has made its choice, each has combined
and amalgamated in its own way the suffixes taken
from the common stock ; and herein it is distinguished
from its kindred, hereby it displays its own genius
and vitality.

A good many ready-made themes may, however, be
found throughout the entire family, more especially
the simpler ones, those which have only a letter or a
syllable, called formative, attached to the root. The
origin of these suffixes, so slight and meaningless yet
so persistent, is the subject of many hypotheses. Is
the extra vowel a very early appeal to the attention of
the auditor ? Is it a survival from a time when our
idioms, like many Asiatic and African idioms, avoided -
final consonants ? In that case the dissyllables ihara,
vaka, (feri), voca, would be earlier than the contracted
roots hhaVy vak, fer, voc. It is possible, but more
often, perhaps, the superfluous vowels are merely a
rest for the voice, an easy liaison^ what have since
been called euphonic letters, whose use commonly
escapes us.

It is almost always in company with this vowel that
the root forms part of a compound word : Theophilus,
Philotheos, lover of God ; Patrocles, Cleopatra, pride of
his (or her) father; and a study of this great class
of words often shows us the true theme of a noun or
verb, preserved from attrition by the second term in
the compound.

Composition, properly so called — ^that is, the juxta-
position of two or more themes, of which the second
only is declinable — belongs to the inflected languages
only. Sanscrit, Greek, and German have made great



Digitized by VjOOQIC



Indo-European Roots. 283

use of this power, indeed even to abuse ; the Romance
languages, on the other hand, have almost lost it.

In a more general sense, the word may be applied
to the prefixing of adverbs, prepositions, conjunctions,
which do so much in the way of differentiating and
altering the meaning of the root. Maleficits is a com-
pound, just as are magnificus^ 'padficfos, camifex ; nor
should I exclude contradiction intercipere, pro- or post-
panere^pro/undtts, from the category of compound worda

The formation of Indo-European words is, then, the
result of two processes : derivation, by means of suflSxes,
and composition, by the joining of two or more themes,
and by the addition of prefixes. We have only now
to consider the terminations which show the function
and place of words in the sentence.



Digitized by VjOOQIC



CHAPTER 11.

PARTS OF SPEECH-THE NOUN.

Original identity of substantive and adjective — ^Parallel formation of
noun and verb — Declension — Case-endings are postpositions of
demonstrative suffixes — The attrition of terminations a proof of
their great antiquity — Nominative, accusative, subject, and
object in old French — Grenitive, dative, vocative, and ablative —
Insufficiency of cases — Gradual substitution of the preposition for
the postposition — Degrees of comparison — The disappearance of
the suffix before the adverb.

Before we touch on the mechanism of the declension,
we have to solve one or two preliminary problems,
which offer a far greater interest than the nomenclature
of the forms special to the various families of languages.
What is a noun ? What is a verb ? Is the one class
anterior to the other ? What are case and verbal
endings ? what their na*^ure and their office?

The noun is the individual or generic designation of
an object, in the most general sense of that word, of
a sensible or intellectual object, concrete or abstract.
So at least it now appears to us when we say lion,
Hger, sheep^ horse, house, or even pain, pleasure,
feeling, idea. But although we apply to these words
the term noun-substantive, although all hold the same
place in the sentence, and are brought into relation
with one another in the same manner, and with other
parts of the proposition by the same methods, yet
even a cursory examination will reveal to us shades of

difference among them. Most of them are derived

284



Digitized by VjOOQIC



Parts of Speech. — The Noun, 285

from derivatives ; and it is through forgotten meta-
phors and transpositions, through changes innumerable
in form and meaning, that they have come to express
for us the, so to speak, real and objective image of
the being, thing, or quality with which we identify
them. They represent no substance. But, by a useful
illusion, eliminating more or less voluntarily all acces-
sory or approximative meanings, we imagine that we
conceive clearly the object designated. Are there now,
were there ever in the past, true substantives ? Ye^
and no. That is to say, that in the earliest times,
when man was yet unable to analyse his sensations,
the meeting with an animal, the passage of a meteor,
a blow received, pain or joy experienced, may have
provoked cries, phonetic gestures, which answered ade-
quately to the impression received ; and these sounds,
these primitive names, may, by the merest chance,
have come down to us. I am speaking here of the
Indo-European languages only. For it cannot be
doubted that a great number of Chinese monosyllables
were signs, attached either to objects or to vaiious
aspects of objects, aspects considered as new objects.
M. Michel Br^al is disposed to believe that naked
roots like srp^ like at;, 05, va^ wp^ or aqo^ gvau, which
are found in serpens, in ovis, in as^i, in vata, in aqua^
in hos^ were essentially what I shall call raw substan-
tives, and signified before anything else serpent^ sheepy
breath, wind, watery ox, and by analogy only came to
be applied to other beings, or the actions of other
bein<rs, and took other meanings, such as to glides
undidating, to breatTie, life, to run, agile, to wander, to
walk, earth. This opinion is plausible, yet the haste
with which speech seizes upon those raw substantives
to express qualities or actions seems to prove that



Digitized by VjOOQIC



286 The Indo-European Organism.

tfaey themselves are the resalt of an unconscious
analysis, and that they describe the most salient
peculiarity of the object heard, seen, or touched. Our
sensibility, in fact, having five ways of perceiving
external things, is itself an instrument of abstraction.
The senses co-operate and supplement each other, but
it is always the one which is most directly affected
which determines the impression on the brain and its
expression by the voice ; and the vocal symbol neces-
larily differs according as it corresponds to an indication
of sight, hearing, touch, smell, or taste. Hence the
number of synonyms ultimately rejected, or reserved
for approximate shades of meaning, or qualities per-
ceived in the same object by the eye, the ear, or the
hand. So that the noun, even when the most in-
voluntary expression of the primitive impression, can
only be the expression of a quality.

There is, then, no original difference between the
substantive and the adjective. Both are names which
express a quality, a manner of existence, either gene-
ralised and applicable to all the objects which possess
it, or specialised and identified with the whole of the
object, of which it really designates but one property.
'* All substantives," says M. Breal, " were, to begin
with, adjectives taken substantively."

How did the adjective come to be distinguished in
the long-run not only in meaning but in form ? In
the first place, the adjective, habitually used to signify
an object, lost its qualifying value, and came to be
solely the name of the object. For instance, defva,
which means the shining, and which has still in
Sanscrit the three degrees of comparison, ended by
meaning the god. Sourya^ the brilliant, became the
name of the sun. Akva, the runner, became the



Digitized by VjOOQIC



Parts of Speech. — The Noun. 287

narae of the horse. Manu^ the intelligent, signified
man. The epithet was forgotten in the thing desig-
nated by it. Other words, on the contrary, laghu
(6r. eKayy(i^ Lat. le{g)rAs, light), tami, hrghu {tenuis^
hre(g)ms), nava, new, not being specially attached to
any object, retained, with their qualifying power, the
faculty of taking the three genders, which the sub-
stantives lost, and the comparative and superlative
forms. Phonetic change, by obscuring the meaning
of the roots, contributed also to separate the two
classes of words. The Hindu, whose language is
less modified, can perceive the relationship which
exists between akva and a^, rapid. But what Greek
would have guessed the affinity between wKvg and
tinrog ? Then a choice was made between the suffixes,
and that as early as the Indo-European period. If
in Latin the suffix ti forms alike nouns and adjectives,
pestis, vestis, ficstis, mentis, fortis, mitis, tristis, there
are others, such as mxin and men (agman, the sky,
nomen, documeny foramen, eocam^en, agmen), such as tra,
tro (irXrJKTpov, rastruvi, cultrum, m/mstrum), which
were used for nouns alone. We sometimes find this
sorting process taking place in a single language. In
the Vedic dialect, for instance, the suffix as still forms
adjectives, tar-as, penetrating, ap-aSy action; but in
Sanscrit it rarely forms anything but nouns, man-as,
gan-as, as in Greek and Latin, juLcvog, gen-us, op-us.

The verb, which we shall consider separately, is
closely related with the nonn and the adjective. At
first it expresses, like them, a state, a manner of
being, an action ; it even borrows their form for its
supines, participles, gerundives, and infinitives. Ee-
duced to its simplest elements, it is composed, like
them, of an attributive root or theme, and of a



Digitized by VjOOQIC



288 The Indo-European Organism.

demonstrative root Bhara-s, the bearer, the burden,
is the same as b?tara-ti, he bears ; for the two suffixes
s and ti represent the same pronoun, of which the
sound varies between sa and ta, and which replaces,
announces, or recalls the subject expressed or under-
stood. For the rest, the fundamental identity of noun
and verb is proved to us by the monosyllabic lan-
guages ; the same word may be noun, adjective, or
verb, according to its place in the sentence. But in
the agglutinative idioms we see the beginning of the
differences, which become more marked in inflected
languages. Affixes placed before or after a root,
which may turn into a verb, are pronominal, and not
simply indicative ; already some rudimentary artifices
add to the action the idea of present or of past time.
The complicated edifice of the Indo-European conjuga-
tion is raised on the same foundation, and by the aid
of similar materials. Only the joints, which a ruder
construction allowed us to see, have here disappeared,
hidden by the fusion in which the theme, itself in-
flected, contracted, lengthened, reduplicated, and the
auxiliaries and terminations are welded together.

In his ingenious essay on the chronology of language,
Curtius rejects the hypothesis of a parallel develop-
ment of the noun and of the verb. He believes that
while the former remains as a naked root, or at least
as an indeclinable theme, the latter acquired the six
personal terminations, '' invariable characteristic of all
the Indo-European languages." The declension is.
even, in his view, of later date than the insertion
between the root and the termination of the suffixes
a, ja, nu, na, pa, ta, ^a, which give an intensive, causa-
tive, frequentative, desiderative meaning to the verb,
and later also than the period of auxiliary verbs and



Digitized by VjOOQIC



Parts of speech, — The Noun. 289

compound tenses. Tor my part, I can see neither the
reason nor the probability of these successive forma-
tions. If the method of the conjugation has been
longer preserved, it is because it corresponded longer
to the needs of speech ; perhaps also because it was
only just completed, and present to all memories at
the time when the different tribes which composed
the family began their slow exodus. In any case, we
find the declension far more worn by time, an indica-
tion of an antiquity at least as great ; even in Sanscrit,
Zend, and Lithuanian, where we find it most complete,
it can hardly account for some of its forms. The
cases whose endings are worn and altered become
confounded and mislead the auditor, so that recourse
is had to prepositions. The Latins, Greeks, and Ger-
mans lose two or three terminations as their language
progresses, though some remnants still embarrass our
modern idioms,

Though it be impossible to recover the primitive
form of the case-endings of nouns, it is none the less
certain that they were equivalent to the prepositions
which have reinforced or replaced them ; when they
were joined. to attributive roots, the mother-tongue
admitted only of postpositions or enclitics ; these
latter, perforce abridged in pronunciation, became con-
tinually effaced, and numerous independent particles,
already provided with a meaning and a form of de-
clension, were used to supplement their failing and
decrepid predecessors. As for Schlegel's mystical
theory of the vegetation of words, each word throwing
out, like certain insects or molluscs, filaments, buds,
ephemeral or durable excrescences, it has long been
forgotten ; even if the testimony of the three first
persons of the verbs in m% si, ti^ did not prove the



Digitized by VjOOQIC



290 The Indo-European Organism.

pronominal origin of these terminations, it would be
impossible to separate them from the suffixes. Now
no suffix proceeds from the root ; all are added to it.
This is equally evident for case-endings and personal
terminations.

The history of the noun and that of the verb can-
not be separated ; they arose together, bound together
by a common radical, and gradually distinguished by
suffixes hardly different from each other, but increas-
ingly used for special purposes. One of the functions
of the verb being to put two nouns into relations, to
mark the action of one upon another, the case-endings
seemed appropriate to lend the phrase force and clear-
ness. When the prepositions had acquired definite
meaning, the verb and the noun both made use of them,
the one to enable it to govern a given case, the other
to enhance some weakened or doubtful termination.
And in the meantime, the natural process of derivation
was always producing verbs from nouns, and from the
verbs verbal nouns and adjectives, which themselves
produced new forms.

The earliest known form of Indo-European declen-
sion has eight cases : nominative, vocative, accusative,
genitive, dative, ablative, locative, and instrumental; but
it might admit, it has admitted, of many others; for there
are other shades in the relations of words to each other,
and even these cases express imperfectly the meanings
of with, for, through, on, under ^ &c. The Latin me-cum,
vdbiS'Cum, is a combination which is apparently a form
of case ; and the Umbrian dialect offers examples of
true cases which it has created for itself : anglu-tOj
from the angles ; anglom-e, near the angle ; totam-e,
dsam-e, near the city, near the altar ; tdta-per, for the
city ; asa-co, with the altar ; asam^ar (for ad), towards



Digitized by VjOOQIC



Parts of Speech. — The Noun, 291

the altar. These examples are precious because they
are still lucid. The Umbrians formed these cases iu
imitation of those they already possessed ; they did
not innovate. Their method is that which gave rise
to the declension ; and the suflSxes they used are not
different in kind from earlier terminations.

Since each language has used the cases in its own
way, whether it suppresses them, whether it com-
pensates for their loss by lengthening or accentuating
the syllable which should bear, their sign, or whether
it modifies their meaning and function, it is impossible
to present a general table of the Indo-European declen-
sions, unless it be in Sanscrit and Zend, in which all
the forms are moulded on the same type. Yet even
here examples of these two idioms, embracing the
comparison in detail of the various masculine, feminine,
and neuter themes, in the singular, dual, and plural,
would overwhelm the reader with. their innumerable
variations, and would only leave a confused and
fatiguing impression. This mechanism, too delicate, too
minute to be solid, should be studied separately and at
leisure in each language.

Omitting the instrumental, which is peculiar to the
Oriental and Letto-Sclavonic groups, we will give a
few succinct notions, incomplete but clear, about the
other cases, showing in especial certain agreements or
curious divergences.

If the declension, instead of being the unconscious
work of time, had been invented by some judicious
grammarian, it would seem that the nominative might
have dispensed with any outward sign. And, in fact,
many neuter and some masculine nouns in m\ as, i,
and u, have no termination in the nominative. But
in general, the naked theme has been reserved for the



Digitized by VjOOQIC



292 The Indo-European Organism.

vocative, with abbreviation of tbe final vowel : akva^
rrx€, domine, amice, O horse, master, O friend.
In the great majority of noans and adjectives, mas-
caline and even feminine, whether the theme ends in
a vowel or a consonant, the nominative termination is
an 8, the remnant of the demonstrative sa. The s is
too familiar to those who have any knowledge of Latin
and Greek to need examples ; it persists in tbe Scla-
vonic languages and in Lithuanian ; retained in Gothic,
it has become r in Scandinavian ; German has almost
entirely lost it. But it is clear that at the time of the
separation of the idioms this case-ending was already
decaying. Before certain words, and in virtue of
euphonic laws which are often strange, Sanscrit re-
placed it by soft breathing. Zend compensated its
loss by a change in the final vowel : mazddo, for
mazdaSy just as the Provencal says chivao for chivals
(whence the French chevau-Uger), In Latin the s was
faintly heard, and was elided before a consonant
{omnihu, Jovi) from the time of Ennius ; and while the
classic Latin re-establish)3d and maintained it scrupu-
lously, the popular speech cared little for it. There is
no trace of it left in Italian, and very little in the
other Romance languages. French retained it longest,
using it, however, with very little sense of its meaning.
As early as the fifth century, M. Brackel tells us,
long before the appearance of written French, popular
Latin reduced the number of cases to two, subject and
object, and, in order to distinguish them, chose the
two terminations which recurred the most frequently,
bonus, bomt-m, muru-s, muru-m, French grammar,
which is a continuation of Latin grammar, inherited
this system in part ; it could not revive the final m
of the accusative, nor the i of the plural mu7H, which



Digitized by VjOOQIC



Parts of speech. — The Noun, 293

had dropped; but it retained the s wherever tho
Latin of the people pronounced it, and constituted a
declension which was very useless, but simple and
acceptable : in the singular li murs and le mur, plural
li muT (illi muri)y les murs (illos muros). In the
thirteenth century, taking as type the second Latin
declension, the 8 of the nominative singular was extended
to those forms which had never possessed it ; and men
wrote li pastreSy the shepherd. This artificial con-
struction ruined the declension which it was intended
to confirm. Rejected by the people from the thirteenth
century, and often neglected by the lettered class, the
French declension died out in the fourteenth century.
Thenceforward only one case was used for each number,
and as the objective case was longer in Latin, and
therefore better able to resist the strong tendency of
modem French to contraction, this case prevailed
and was chosen as the type. Murum became mwr,
muros, murs. Thus all our nominatives represent
Latin accusatives. Some remains of the ancient
nominative singular may be found, however, in tl e
nine following words : fils^fonds, lacs, legs, lis {lilius), lez
(lotus), puits (pitteus), rets (retis), qu£ux {coquus) ; which
were in the accusative: fil (whence Jiliation), fond
(fundum), leg, li, U (still used in the phrase %m U
d'itoffe) ; latum, puit (puteum), ret (retem, rMaire),
queu or coq (ship's cook). Such is the end, in no sense
to be regretted, of the earliest and most inconvenient
of case-endings, of that which engendered perpetual
confusion between all the cases and both numbers.

The general sign of the accusative singular, m, has
everywhere, as we have seen, marked the direct object,
with the primitive sense, doubtless, of movement to-
wards an object, a place, a being : eo Romdvi, eo
20



Digitized by VjOOQIC



294 '^^^ Indo-European Organism.

Lugdunum^ I go towards Ronie, towards Lyons. When
the verb, at first intransitive, took the active sense,
implying in itself movement influence on the object
" governed," the original sense of the termination m
became fainter, but the sign remained no less useful
to distinguish the object, and mark its subordination
to the subject of the sentence. Universally adopted,
it perished nevertheless, and its disappearance began
early in our Western languages. Greek, rejecting the
sound of m at the end of a word, replaced it everywhere
by a n, an equivalent nasal, which became labialised
before a p or a &. They pronounced rofji iroXejuLOP, war,
€19 rafjL TToXiv (whence Istamboul, towards the town).
But in a number of forms this n dropped, preserving
only the vowel a of the primitive theme, which would
otherwise have been weakened into e ; avSpa for avepav,
Kvva for Kupavi, dog. In Latin, in Umbrian, in the
earliest inscriptions, the final m is no longer written, and
even in the classical Latin of the Augustan age it was so
little pronounced that its elision was the rule in poetry.
It became completely obliterated from popular Latin
and from the Romance languages ; so much so, that in
order to pronounce it in reading Latin, an Italian is
obliged to double it and add a mute e, sanctumme, like
GerusaleniTne, Retained in Gothic, tunthum, the tooth,
it becomes lost in the German n, or rather it remains
only as the sign of the dative (derriy gutem), in the use-
less pronominal declension of literary German. One
of the reasons for the disappearance of this interesting
m is again to be found in its double use. Not only
was it attached, we know not why, to the neuter
nominative (dananiy donum\ but it also terminated the
genitive plural, devasam, deorum, rosarum, omnium.
Thua it lost its distinctive value and its vitality. It



Digitized by VjOOQIC



Parts of Speech. — The Noun. 295

existed also, but followed and soon absorbed by an s, in
the accusative plural : yueyciXoi/?, dominos, are contrac-
tions of iuLeydXov9, dominumSy a regular phenomenon
of assimilation in Greek and Latin.

The genitive or possessive case is peculiar ; it is, so
to speak, retroactive ; it seems governed by the nomina-
tive, but it contains the true subject. Liber Petri, it
is Peter who is the possessor of the book. It would
not therefore be surprising if the sign of the genitive
were derived from an enclitic relative suffix, "The
book, Peter, which," " the book which Peter " (under-
stand possesses, or wrote) ; hence the frequent con-
fusions with the ablative of or hy, " the object which
is of, or made by, Peter." This case, the genitive, was
represented by several terminations, especially sya for
vowel themes, and as, os, is, for consonantal themes.
Greek used both, but the intermediate 8 has dropped
as commonly happens between two vowels, and the
Homeric 010 (Oeoio) for ocrio has been contracted into
01; — Xoyov. The other case-ending has remained with-
out alteration, (pXayog^ of the flame (as also the plural,
«i/), apSpog, of the man, j9a<r/X€[f]a)?, Ai[F]69. Latin
has completely rejected sya, and replaced it by an i in
three declensions : rosai, diei, domini. It has kept as
under the archaic forms 08 {senatuos, magistrcduos,
manuos) and U8 ( Venerics, Cererus) ; and finally under
the classical form is {fratris,8ororis),wh\ch unfortunately
created a confusion with the numerous nominative sin-
gulars in is and with the ancient plurals frugiferenteis, ,
parenteis. The Germanic languages still use the geni-
tive in 8, and it is one of the few traces of grammar
retained in English,

The comparative table of the dative shows the inde-