John Beames.

A Comparative Grammar of the Modern Aryan Languages of India: The noun and ... online

. (page 17 of 27)
Online LibraryJohn BeamesA Comparative Grammar of the Modern Aryan Languages of India: The noun and ... → online text (page 17 of 27)
Font size
QR-code for this ebook

melted down into one form distinguished by the ending dhi, di,
or e. We thus account for the fact that in Hindi and Panjabi

^ '^Gujar^tlbh^bdno itihas," p. 44, quoted from a poem called Munjar&sa, the
dato of whick » not glTen.

Digitized by VjOOQ IC


the oxytone noun in d forms its base in ^^ as ghofd, ghare. The
same takes place with Sindhi nouns in o, which correspond to
Hindi oxytones in (i, as wft " head," ^|%. Hindi and Panjabi
have only this one method of forming the oblique, and, as I have
stated before, only employ it in oxytone nouns ending in 4*
All other nouns yirtually end in a consonant, and are not there-
fore open to any change ; they have rejected all inflection ab*
solutely. In mediaeval Hindi, words of this class use the form
tf , with no connecting vowel, as i^T^ff " to EAm," and when,
in process of time, this affix was dropped, there remained nothing
but the bare stem, incapable of inflection. Even in those nouns
which, strictly speaking, end in other vowels than d, the same
rule is followed, because these languages take no heed of final
vowels, and in speaking, at least, reject them always ; and ev^i
in writing they are of little value.

It is in the languages of the western group— as might be
expected — that the greatest diversity exists, and to them we
must now turn. Sindhi takes for the general type of its
singular oblique the vowel a, for which Trumpp hints at a
derivation from the genitive; but we have pointed out that,
before the period of the rise of the modem languages as such,
the Apabhransa Prakrit had already nearly obliterated all dis-
tinction between the genitive and other oblique cases, bringing
them all down to the common form dhi. If this be the case in
the written Apabhransa, — which, though wandering far from
the central type of Prakrit, must still, as a written language,
be supposed to have retained greater regularity than the spoken
language, — we are justified in supposing that, in the spoken, a
still more complete fusion of all the case-endings must have
taken place ; and it is not likely that a rude pastoral race would
carefully observe such minute distinctions as that between dhi,
dhe, dho, and dhin. Moreover, we notice that even in the
written language, in one case at least, the final short vowel had
been rejected, so that the ablative ends in dh or d. Sindhi is

Digitized by VjOOQ IC


very prone to the shortening of vowels. Thus it would be quite
in harmony with the gen^*al practice of the language, to pre-
serve out of all these endings nothing more than short a. In
masculine nouns ending in u, we have really the consonantal
Ending of barytones of the a-stem, and it is therefore only what
we should expect to find that the oblique of this class should
reject the w, which is hardly audible in pronunciation, and con-
clude with a. Thus ^TH "slave," obi. ;^TO, may be traced
back to a Prakrit form nom. ^\H\ ^T^j obi. ^I^lf^, ^TOT^,
^rer^, ^lifllffi, all of which in the spoken language would
fuse into ^TOf^, as in Hindi, and thence into ^TOf, and finally
^W. The reason why nouns in o form their oblique in e after
the Hindi fashion, appears to confirm the view I have taken
of the origin of these nouns. When they owe their long final
vowel to the fact of their being derived from Skr. oxytones, the
presence of the accent on the final syllable prevents the termi-
nations dhi, dhCj etc., from shortening their d into a ; all that
takes place therefore is the rejection of ¥, and the termination
thus becomes d'i, which by a natural process becomes e. Sindhi
nouns in short a and t do not differ in their oblique from the
nom. This is a further confirmation of the view expressed above.
The Prakrit oblique of such nouns would end in ihi, ahi; but the
i and a belong to the stem, not to the termination ; and when
the hi is rejected, there remains nothing, so that the oblique
cannot imdergo any change. Kouns in long i and A add an a
to the stem, which is again a relic of the common form ahi
deprived of its final hi. Long before the epoch of the formation
of these modem cases, the Prakrit had disencumbered itself of
the habit of making an euphonic combination between the final
vowel of the stem and the initial vowel of the termination; thus
the long i and A hold their place, imchanged by any commotions
which might vex the termination.

In Gujarati the only change that occurs is in the oxytone
nouns in o, which make an obi. ia d. I have often before

Digitized by VjOOQ IC


mentioned that Ghijarati strikes the student as an archaic
dialect of Hindi, whose development has been arrested by its
isolation ; and it would be consonant with this view to regard
the oblique form in d as derived from the full Prakrit dhi, by-
simple rejection of the final hi — a process which we have shown
to be followed by the other languages. Beyond this there is
no further change for the oblique in Oujarati ; the termination
in e belongs to the instrumental, which will be explained here-

Marathi remains to be discussed ; lumbering along as usual
with its old-world Prakrit baggage of terminations, it ofiers
many troublesome problems to the inquirer. In nouns with a
consonantal ending it lengthens the mute a into d : as bdp^ obi.
bdpd; ghar, obi. ghard. In Old-Marathi the nominative of
nouns of this class, like the corresponding class in Sindhi still,
and Hindi and the rest in former times, ended in w, the Prakrit
barytone o. Thus we have ^J]^, ^^, if«^4J, and the like. The
Marathi consequently did not consider words of this class as
ending in a consonant, nor does it now ; technically these words
still end in short a. Not having, like the Hindi, rejected this
final vowel, and with it all power of modifying the termination,
it has been able to retain the oblique form in d from dhi, merely
rejecting the hi, as in the others. This it has been able to do
in barytones, whereas the others only preserve this ending,
softened to ^, in oxytones. Similarly in a certain class ending
in long d, it exhibits the oblique in d ; but in this case it may be
equally correct to regard the d as merely a retention of the form
of the nom. ; and this is rendered more probable by the parallel
case of nouns in I and ^, which also, as shown in the table, have no
separate form for the oblique. The noims in long vowels would
be unable to form a separate oblique, because the obKque of the
Prakrit woidd merely differ from the nom. by the addition of
hi, as dhi, Ihi, Mi; so that when the hi came to be rejected, there
would remain nothing. Marathi differs from Sindhi and all

Digitized by VjOOQ IC


the other languages in being still bothered with notions of
aandhi, and still hardens the final vowel of a stem before a
termination beginning with a vowel. Thus in a class of nouns
ending in d, i and H, it hardens the vowel before the imiversal dhi
of the oblique, making from bhdu, bhdvd and the like. In mascu-
lines and neuters of this class, the Prakrit masc. form dhi has been
used, leaving d: thus they say, hhdvdld "to a brother,'* which is
hhdv'd{hi)'ld\ but in the fern, the Prakrit fem. is used. In
these words so early as the Mahar&shtri Prakrit, the fusion of
cases had taken place. Thus in the feminines mdld^ devi, bahu,
we find only the following narrow range of endings :

ifTWr'' garland," ace. TTRfy abl. 4i|<j||ff , Instr. Gen. Loc. ?nwni-
^^ "goddess," „ ^pl „ ^^VfV f» » » 4^ll»
^If "wife," „ Hjf „ inrff » " »» ^^^8fII-

The abl. differs very slightly in sound from the other oblique
case, because e is short in Prakrit, and consequently to the
vulgar ear the general type for the fem. oblique would be e.
Thence it would result that in the words ^rW» ^jft* after the
final vowel had recrudesced into its semivowel, the form of the
oblique to be added would not be d, as in the masc, but e, and
we therefore find adsaveld, stripeld, which are sdsav-e-ld, striy-
-e-ld. In this case the Marathi is more s^isitive than the
Prakrit, for it does not permit the hiatus where the other does.
The principle of changing the final vowel into its semivowel
having been once introduced, has been iguorantly extended —
through the influence of that blind groping after analogies
which has been so fertile a cause of change in many languages —
to nouns ending in d ; and as these have no semivowel of their
own, the most frequently used of the two semivowels, % has
been applied to them, so that we get an oblique sdsaryd from
a nom. adsard. It is precisely on the same principle that the
weak declension of nouns and the weak conjugation of verbs
have gained so largely, and are still gaining, both in English and

Digitized by VjOOQ IC

DECLENfilOIf. 217

Germany on the old strong forms. The sixth declension in
Marathi, namely that which we are now considering, is one of
the weak declensions in fact, and as such has gained ground on
the strong declensions. Another weak declension is that which
comprises masculines and neuters in ^ and ^, which lose their
final vowel and form their oblique as though their nom. had
ended in mute a, or in former times in u. Thus rd/sarii makes
its oblique neither vdtsaru nor vdtmravd, but vdtsard, where the
final vowel of the nom. has been ignored, and the masc. oblique
sign d has been added to a stem tdtsar. On analyzing the
words which fall imder this class, it becomes apparent that in
most cases the final A or An is a modem invention, and not
organic. They are, first, words compounded with the Skr. ^S^
and 9^, in which the final u is short, and might thus easily be
confounded in Old-Marathi with barytones of the a-stem like
^j;^, so that they formed their oblique in d, and the
lengthening of the final vowel of the nom. is only another
instance of the fondness of Marathi for final long syllables :
secondly, they are neuters ending in the diminutive syUables
vi and ^ which, as I have shown in § 24, are in the other
languages TT or ^, and "^ or '^, respectively, and thus come
imder the head of oxytones of the (Z-stem, and the oblique
would regularly be d. There is, however, very great irregularity
and confusion on this subject, the language not having made up
its mind as to which of the three forms available it will use.

Nouns ending in short a, corrupted from feminines in Skr. in
d, of which the type is ftQET, M. ^sft^, form their oblique by
adding e, as ^i^. This is the same rule as that followed in
feminines in long i and A, and the e is the regular Prakrit
oblique. Thus the Pr. obi. would be TTPirnC; but in this case, as
the Marathi has lost the long d, it merely adds the e to the final
consonant. This it does also in Tatsamas which retain the long
d ; thus Yrr?n makes ^nif • Here I smell the Pandits. I suspect
that the nom. had become* ifm? ^ in Hindi, in which case the

Digitized by VjOOQ IC


form J(ift for TrnfTT would be regular, just as ^^ for ^in[;
but the Pandits have subsequently restored the long d. It is
no answer to this to say that 4i|dl is found in the earliest
writings in the language, because, in the first place, the writings
are no guide to what the speech of those days really was ; no
Indian writer could ever resist the temptation to use grander and
more Sanskritic words than occurred in the spoken language ;
the attractions of the so-called sddhu bhdshd have always been
irresistible; and, secondly, the formation of the oblique form
took place long before the earliest writings that we have; and
it is therefore quite possible that when the oblique in e was
formed, the nominative in current use was ^TRT.

There is, as before stated, also another method of forming the
oblique in use among the nouns of this class, namely that in I,
which arises from the fact of their being derived from noims
which in Skr. ended in t, or I. In masculines of this class the
oblique Prakrit is in ihi, in feminines it is in ie, both of which
have left in Marathi only L Examples of this class are the


Skr. lrf%r "fire," "MU^, HnHtf. ^fPTt ^IPft-

Skr.^ "belly," ^Bf^, f^frtf, ^, f^.

Skr. ^ffe "fist," Tfff, 4J^lff, Jj^, ^^.

Skr. ^Jt^ "assembly," ^f^» 'ftf^* ^^t^» 'ft^-

This last word is almost a Tatsama; it would be completely so
had it not lost the final i in the nominative ; it is used in the sense
of " talk, gossip, conversation,*' also of "an affiiir, case, business."

§ 47. We now come to the oblique forms of the plural, which
are in all respects simpler and more imiform than those of the
singular. Hindi has but one form for all classes of nouns,
namely ijf , which must, I think, be distinctly referred to the
genitive of the older languages. Sanskrit forms the genitive
of the a-stem in ^IPlt for all three genders; the nom. and acq.

Digitized by VjOOQ IC


plural of the neuter end also in dni, which I have shown to be
the origin of the nom. pi. in Hindi. The similarity of the
genitive ending to that of the nom. has perhaps led to its
preservation for the oblique cases, as in the period when the old
inflectional case-endings had died out, and the use of the modem
particles had hardly become fixed, there would be no distinction
between the different cases of the plural; and Chand accordingly,
as pointed out previously, uses such forms as ^T^l, ^«*IR, both
for nom. and obi. plural. In his Gslth& passages, where he
employs archaic constructions, we find a genitive in ^Ipt, as in

"The sweet sound (made) by the anklets ofwamen.^^ — ^Pr. R. i. 17.

(^=W^> ^=Ip^, ^5T^ = ^3T^ instrumental of a
fem. form ^TJ>)

Prakrit, in the principal dialect, makes its genitive in inJ^,
and extends this form to all classes of nouns, totally rejecting
the Skr. genitive in dm used in so many bases. In fact the
terminations of the a-stem have, as a rule, completely over-
ridden and supplanted all the others. Hindi has rejected the
final anuswslra of the Pr. and turned the n into anusw&ra, and
this rejection and softening are the probable causes of the
present form in ^, the long vowel o having its origin in an
effort to compensate for the loss of the n. Panjabi, which is not
so sensitive, retains simply ^ for the oblique plural. There
seems to be no room for doubt that the Pr. genitive is the origin
of these forms, because the other cases have a different type
altogether. Thus the Maharashtri has instr. in ehi or ehin, abl.
in aunto or hinto, loc. in esu, emn ; and though the Aprabhansa
has a different range of endings, yet they do not, on the one
hand, approach the Skr. genitive, nor afford, on the other hand,
materials for the construction of the Hindi oblique pL, the long
of which is in my opinion to be accoimted for by a still
further lengthening of long d, a letter which occurs only in the

Digitized by VjOOQ IC


genitive. That tlie Hindi form is comparatively modem is
shown by the fact of its not occurring in any of the middle-age
poets, in whose writings the form in ^rfif , ^1, or ipf, is used
for the oblique as well as for the nom. This is why I referred
at the beginning of this section to the similarity between the
nom. neuter in dni and the gen. of all three genders in dndm.
I believe that this similarity is the cause why no separate form
for the oblique was struck out for so many centuries. It is a
further confirmation of this view, that Gujarati, with its arrested
development, has no oblique form for the plural, nor have
BengaU and Oriya, both of which languages must have sepa-
•ated themselves from the central Hindi type certainly earlier
than A.D. 1400, as we find Bidy&pati in a.d. 1433 in full
possession of a distinct set of forms. The Bhojpuri dialect of
Hindi also does not possess the form in ^, but makes its
oblique plural still in ^if , as ?ftTT ^> etc. ; so also the Marwari
dialect, which uses only irt, which, like the Panjabi form, is the
legitimate descendant of Chand's plural in ^if .

By the aid of this view the terminations in use in Sindhi are
also explainable. The Apabhransa dialect, which is more
especially connected with Sindhi, has fused all its plural endings
into a small range of forms, as instr. Mm, dhin, ihin, (Lhin, abL
ahuUf gen. ahan, nom. and ace. du. Only the loc. retains a
distinct shape dsu; and even in this, when we remember the
facility with which Sindhi changes ^ into 1, it becomes
probable that a form dhu would not be long in making its
appearance. Later Apabhransa genitive forms in iheUy uhen, are
also found ; so that we really get as materials for the Sindhi
oblique little more than one form with trifling variations.
The oldest and fullest form of the oblique in this language ends
in fif preceded in each class of nouns by the final vowel of the
stem. Here we have the if or ^ of the Skr. forms 'SRTf'f and
^Hliftt Pr. mftr and W^, fused together. The other forms nt>
% are readily deduced from the Apabhransa forms ^ and*f>

Digitized by LjOOQ IC


the vowel preceding whicli is merely a yariant derived from the
final vowel of the stem.

Precisely similar are the Marathi forms, which are strictly
analogous to the singular oblique forms of each class, only
differing by the insertion of anusw&ra, which evidently points
back to the ^ of the Fr. gen., and preceded in each instance by
the phonetic peculiarities which mark the singular.

On the whole, then, we conclude that in both singular and
plural the terminations of the oblique descend from a general
form produced by the fusion of all the oblique cases of Sanskrit ;
but there is this difference between the two numbers, that
whereas in the singular no one case has retained its individual
existence, or impressed its own special type upon the modern
nouns, in the plural the genitive, by virtue of a special strength
of type, and by its similarity to the neuter nom., which had
usurped the place of the other nominatives, has preserved its
individuality, and in a great majority of classes absorbed into
itself the other cases. It might also, however, be said that to a
certain extent, even in the singular, the genitive has had the
preponderance, as the form in f^, although its earliest appear-
ance in Prakrit is in the capacity of an ablative, is yet more
easily derivable from the Skr. gen. in '^TS than from any other
classical form. Thus, although the imiversal written Prakrit
termination is ^TW, yet it is phonetically more natural that a
form ^irfH should have arisen, which — by the operation of the
tendency to change ^ into ^, a tendency which certainly exists
in all the languages, though more extensively in the western
members of the group — would become "^rff . It must be noted,
also, that the change took place at a time when these western mem-
bers were most powerful — ^Eastern Hindi, Bengali, and Oriya,
not having then arisen. The period of the origin of these
forms cannot be put later than the seventh century, when the
decay of Buddhism brought about those great linguistic changes
which laid the foundation of the modern languages; and at that

Digitized by LjOOQ IC


epoch the eastern parts of India were, as far as we know, com-
paratively sparsely peopled by men of Aryan race. A phonetic
change, therefore, of the character we are now discussing, would
naturally be in accordance with the tendencies and peculiarities
of the western tribes who then constituted the immense majority
of the Aryans.

§ 48. The forms of the oblique are not, however, the only
traces which still survive of the old Sanskrit inflections.
Simplest of all of these, the locative, which ended in H, has held
its own down to the present day in many languages. In the
Oriya poems this locative exists, as ij^ " in the village," iflf^
" in Gop," though it has now been superseded by the analytical
locative formed by ^, and modem Oriya uses lf(^ \ or iffq yf^ ;
in the latter of these '3T= WPf •

In Bengali it still survives, as in Ift^ " in anger," ^if^ " in
fear," "^ " in a chariot." Here also, as in Oriya, the tendency
to an analytical construction led at an early date to the addition
of the particle %, so that in Kasi Das's Mahabharat forms
^i% It, T^ ^9 occur, although pleonastic, and often more with
the sense of an ablative. After nouns ending in long d, this
ending takes phonetically the form of J(, as ^tTRI "in a
horse;" but after nouns in other vowels, the modem termination
% is more usual.

Hindi does not know this locative form : having adopted e as
the oblique ending for the only class of nouns in which it
admits a separate oblique form, there was no room in its system
for the special locative. Ghijarati regularly retains it in all
cases, with complete disregard of phonetic combinations, so that
it is added to nouns ending in a vowel quite as freely as to those
which end in a consonant. Thus we have ^itl( " in a custom,"
?f1^ " in a tent ;" but in nouns of the masc. o-class, in which
the oblique differs from the nom., the change of termination is
possible, and they consequently write ^|f |^ " in a day." So

Digitized by VjOOQ IC


also in the plural, which universally terminates in o, the e of
the locative is added, giving oe^ no account being taken of the
fact that the locative of the plural in Sanskrit ends in Tjg.
This is the way with the modem languages. Having got into
their heads the idea that a certain termination is typiced of a
certain case, they stick it on to their nouns all round, over-
riding the more intricate distinctions of the older languages,
and thus gaining in simplicity and regularity.

Marathi has a locative in ^, universally employed in the
older poets, but now going by degrees out of use. The form
is the same for both singular and plural, and appears to have
arisen from the Pr. locative in ^lrf^[, which in Bhagavati
appears as %f^y or as Weber reads it ^ftf. The later form
was probably ^f^, which, by rejection of the a, is, strictly
speaking, a portion of the stem, and throwing forward the
anusw&ra, becomes t^. Marathi has by degrees got rid of the
^, as in the similarly constructed forms of the oblique, and the
lengthening of the final vowel is the usual Marathi custom.
It agrees in practice with Gujarati, in using the singular form
for the plural also. This may be pointed out as another instance
of the preference of the Prakrits and modem languages for the
older or pronominal declension, as this termination comes ulti-
mately from forms like fl^f^^-^

Panjabi resembles Marathi in having a locative in \, which,
however, is not restricted to the plural, and is not of very
general use ; thus, ^T^ " in houses,*' If^ " in hands." In the
singular, a locative in ^ is occasionally found, as lEf^ ; but
this is more strictly an ablative, and I suspect we have here,
not a relic of a synthetical case, but an abraded particle, as will
be explained in another place.

* The corrupt Konkani of Goa uses a locative in X> as ?rf% " on the bank," from
Tf^, Skr. IfJ, where classical M. would have ?f^T. (Bumell*s Specimens of
S. Indian Dialects, Mangalore, 1872.) This is probably only a shortening of the
Skr. locative in 1(.

Digitized by VjOOQ IC


In SindU there is a locative, but only in nouns of tlie u-class
(=mute a). It ends in t, shortened probably from the Skr. e,
as in ^^, and not, as Trumpp writes, identical with the locative
termination t, because this latter is not used in the declension of
nouns of the a-stem, from which the Sindhi u-stems descend.
Moreover, the declensional forms of the a-stem have to so great
an extent swallowed up those of all the other stems, that we are
hardly justified in looking to any forms but those of the a-stem,
unless it be the old pronominal forms of words like ^.

Besides the locative, several of the languages have also a relic
of the old synthetical instrumental. This case in Skr. ended in
the a-stem in ena ; and Marathi retains this form shortened into
Tt> as ^*^ " by a weapon,'' ^^ " by a house/' Inasmuch, how-
ever, as this termination is identical with that of the nom. pi.
of neuter nouns, it has become customary to use a pleonastic
construction by the addition of one of the modern particles
41^^ " by means of," so that they would now write ^Ji§f ^d^^
" by means of a weapon." Here, in consequence of the back-
wardness of Marathi, we are enabled to see in force a process
which has occurred in the other languages also at a former
time, namely, the gradual wearing away of the synthetictd

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 19 20 21 22 23 24 25 26 27

Online LibraryJohn BeamesA Comparative Grammar of the Modern Aryan Languages of India: The noun and ... → online text (page 17 of 27)