Skip to main content

Full text of "A Sanskrit grammar; including both the classical language, and the older dialects, of Veda and Brahmana"

See other formats

?*?.; v .'. 

r-v^Hr * 



presented to 

Xibrarp , 

of tb& 

; * * % "'- 

J^>' .TUnivcrsttp of Toronto 

fcl%^%< 'V' u I',* 

pt^:<Hb'U^t : j|-. "V. "^ -r - 

, ^^pfessor W.o.^, 
























Entered according to Act of Congress, in the year 1879, by W. D. Whitney in the ofrice 
of the Librarian of Congress at Washington D. C. 

(The Right of Translation and Reproduction is reserved .) 

Printers : Breitkopf & Hartel, Leipzig. 


It was in June, 1875. as I chanced to be for a day or 
two in Leipzig, that I was unexpectedly invited to prepare 
the Sanskrit grammar for the Indo-European series projected 
by Messrs. Breitkopf and Hartel. After some consideration, 
and consultation with friends, I accepted the task, and have 
since devoted to it what time could be spared from regular 
duties, after the satisfaction of engagements earlier formed. 
If the delay seems a long one, it was nevertheless unavoid- 
able ; and I would gladly, in the interest of the work itself 
have made it still longer. In every such case, it is necess- 
ary to make a compromise between measurably satisfying a 
present pressing need, and doing the subject fuller justice 
at the cost of more time; and it seemed as if the call for 
a Sanskrit grammar on a somewhat different plan from those 
already in use excellent as some of these in many respects 
are - - was urgent enough to recommend a speedy com- 
pletion of the work begun. 

The objects had especially in view in the preparation 
of this grammar have been the following: 

To make a presentation of the facts of the language 
primarily as they show themselves in use in the literature, 
and only secondarily as they are laid down by the native 
grammarians. The earliest European grammars were by the 
necessity of the case chiefly founded on their native prede- 
cessors ; and a traditional method was thus established which 
has been perhaps somewhat too closely adhered to, at the 
expense of clearness and of proportion, as well as of scien- 
tific truth. Accordingly, my attention has not been directed 
toward a profounder study of the grammatical science of the 
Hindu schools : their teachings I have been contented to take 


as already reported to Western learners in the existing 
Western grammars. 

To include also in the presentation the forms and con- 
structions of the older language, as exhibited in the Veda 
and the Brahmana. Grassmann's excellent Index- Vocabulary 
to the Rig- Veda, and my own manuscript one to the Atharva- 
Veda (which I hope soon to be able to make public), gave 
me in full detail the great mass of Vedic material ; and this, 
with some assistance from pupils and friends, I have song] it 
to complete, as far as the circumstances permitted, from the 
other Vedic texts and from the various works of the Brah- 
mana period, both printed and manuscript. 

To treat the language throughout as an accented one, 
omitting nothing of what is known respecting the nature of 
the Sanskrit accent, its changes in combination and inflection, 
and the tone of individual words - - being, in all this, ne- 
cessarily dependent especially upon the material presented 
by the older accentuated texts. 

To cast all statements; classifications, and so on, into a 
form consistent with the teachiogs of linguistic science. In 
doing this, it has been necessary to discard a few of the 
long-used and familiar divisions and terms of Sanskrit gram- 
mar -- for example, the classification and nomenclature of 
"special tenses" and "general tenses" (which is so indefen- 
sible that one can only wonder at its having maintained itself 
so long), the order and terminology of the conjugation-classes, 
the separation in treatment of the facts of internal and ex- 
ternal euphonic combination, and the like. But care has been 
taken to facilitate the transition from the old to the new; 
and the changes, it is believed, will commend themselves 
to unqualified acceptance. It has been sought also to help 
an appreciation of the character of the language by putting 
its facts as far as possible into a statistical form. In this 
respect the native grammar is especially deficient and mis- 

Regard has been constantly had to the practical needs 
of the learner of the language, and it has been attempted, 
by due arrangement and by the use of different sizes of 


type, to make the work as usable by one whose object 
it is to acquire a knowledge of the classical Sanskrit alone 
as those are in which the earlier forms are not included. 
The custom of transliterating all Sanskrit words into Euro- 
pean characters, which has become usual in European Sans- 
krit grammars, is, as a matter of course, retained through- 
out; and. because of the difficulty of setting even a small 
Sanskrit type with anything but a large European, it is 
practiced alone in the smaller sizes. 

While the treatment of the facts of the language has 
thus been made a historical one, within the limits of the 
language itself, I have not ventured to make it comparative, 
by bringing in the analogous forms and processes of other 
related languages. To do this, in addition to all that was 
attempted beside, would have extended the work, both in 
content and in time of preparation, far beyond the limits 
assigned to it. And, having decided to leave out this ele- 
ment, I have done so consistently throughout. Explanations 
of the origin of forms have also been avoided, for the same 
reason and for others, which hardly call for statement. 

A grammar is necessarily in great part founded on its 
predecessors, and it would be in vain to attempt an acknowl- 
edgment in detail of all the aid received from other schol- 
ars. I have had at hand always especially the very schol- 
arly and reliable brief summary of Kielhorn, the full and 
excellent work of Monier Williams, the smaller grammar of 
Bopp (a wonder of learning and method for the time when 
it Avas prepared , and the volumes of Benfey and Mtiller. 
As regards the material of the language, no other aid, of 
course, has been at all comparable with the great Peters- 
burg lexicon of Bohtlingk and Roth, the existence of which 
gives by itself a new character to all investigations of the 
Sanskrit language. What I have not found there or in the 
special collections made by myself or by others for me, I 
have called below "not quotable*' - a provisional designa- 
tion^ necessarily liable to correction in detail by the results 
of further researches. For what concerns the verb, its forms 
and their classification and uses, I have had, as every one 

viii PREFACE . 

must have, by far the most aid from Delbruck. in his Alt- 
indisches Verb urn and his various syntactical contribu- 
tions. Former pupils of my own. Prof. Avery and Dr. 
Edgren. have also helped me. in connection with this sub- 
ject and with others, in a way and measure that calls for 
public acknowledgment. In respect to the important matter 
of the declension in the earliest language. I have made great 
use of the elaborate paper in the Journ. Am. Or. Soc. (print- 
ing contemporaneously with this work, and used by me 
almost, but not quite, to the end of the subject) by my 
former pupil Prof. Lanman; my treatment of it is founded 
on his. My manifold obligations to my own teacher. Prof. 
Weber of Berlin, also require to be mentioned : among other 
things, I owe to him the use of his copies of certain un- 
published texts of the Brahmana period, not otherwise access- 
ible to me; and he was kind enough to look through with 
me my work in its inchoate condition, favoring me with 
valuable suggestions. For this last favor I have likewise to 
thank Prof. Delbruck who, moreover, has taken the trouble 
to glance over for a like purpose the greater part of the 
proof-sheets of the grammar, as they came from the press. 
To Dr. L. Schroder is due whatever use I have been able 
to make (unfortunately a very imperfect one) of the import- 
ant Matriayani-Sanhita. 

Of the deficiencies of my work I am. I think, not less 
fully aware than any critic of it. even the severest, is likely 
to be. Should it be found to answer its intended purpose 
well enough to come to another edition, my endeavor will 
be to improve and complete it; and I shall be grateful for 
any corrections or suggestions which may aid me in mak- 
ing it a more efficient help to the study of the Sanskrit 
language and literature. 

GOTH A, July 1879. 

W. D. W. 



It seems desirable to give here such a sketch of the 
history of Indian literature as shall show the relation to 
one another of the different periods and forms of the lan- 
guage treated in the following grammar, and the position 
of the works there quoted. 

The name "Sanskrit" (samskrta, 1087 d, 'adorned, elab- 
orated, perfected'), which is popularly applied to the whole 
ancient and sacred language of India, belongs more properly 
only to that dialect which, regulated and established by the 
labors of the native grammarians, has led for the last two 
thousand years or more an artificial life, like that of the 
Latin during most of the same period in Europe, as the 
written and spoken means of communication of the learned 
and priestly caste ; and which even at the present day fills 
that office. It is thus distinguished, on the one hand, from 
the later and derived dialects as the Prakrit, forms of 
language which have datable monuments from as early as 
the third century before Christ, and which are represented 
by inscriptions and coins, by the speech of the uneducated 
characters in the Sanskrit dramas (see below), and by a 
limited literature ; the Pali, a Prakritic dialect which became 
the sacred language of Buddhism in Farther India, and is 


still in service there as such ; and yet later and more altered 
tongues forming the transition to the languages of Modern 
India. And, on the other hand, it is distinguished, but 
very much less sharply and widely, from the older dialects 
or forms of speech presented in the canonical literature, 
the Veda and Brahmana. 

This fact, of the fixation by learned treatment of an 
authorized mode of expression, which should thenceforth be 
used according to rule in the intercourse of the educated, 
is the cardinal one in Indian linguistic history; and as the 
native grammatical literature has determined the form of 
the language, so it has also to a large extent determined 
the grammatical treatment of the language by European 

Much in the history of the learned movement is still 
obscure, and opinions are at variance even as to points of 
prime consequence. Only the concluding works in the devel- 
opment of the grammatical science have been preserved to 
us; and though they are evidently the perfected fruits of a 
long series of learned labors, the records of the latter are 
lost beyond recovery. The time and the place of the cre- 
ation of Sanskrit are unknown ; and as to its occasion, we 
have only our inferences and conjectures to rely upon. It 
seems, however, altogether likely that the grammatical sense 
of the ancient Hindus was awakened in great measure by 
their study of the traditional sacred texts, and by their com- 
parison of its different language with that of contemporary 
use. It is certain that the grammatical study of those texts 
(gakhas, lit'ly 'branches'), phonetic and other, was zealously 
and effectively followed in the Brahmanic schools ; this is 
attested by our possession of a number of phonetico-gram- 
matical treatises, firatigaJchyas (prati $ahham, 'belonging to 
each several text 1 ), one having for subject each principal 
Vedic text, and noting all its peculiarities of form; these, 
both by the depth and exactness of their own researches 
and by the number of authorities which they quote, speak 
plainly of a lively scientific activity continued during a long 
time. What part, on the other hand, the notice of differ- 


ences between the correct speech of the learned and the 
altered dialects of the vulgar may have home in the same 
movement is not easy to determine; hut it is not customary 
that a language has its proper usages fixed by rule until 
the danger is distinctly felt of its undergoing corruption. 

The labors of the general school of Sanskrit grammar 
reached a climax in the grammarian Panini, whose text-book, 
containing the facts of the language cast into the highly 
artful ancl difficult form of about four thousand algebraic- 
formula-like rules (in the statement and arrangement of 
which brevity alone is had in view, at the cost of distinct- 
ness and unambiguousness , became for all after time the 
authoritative, almost sacred, norm of correct speech. Re- 
specting his period, nothing really definite and trustworthy 
is known ; but he is with much probability held to have 
lived some time (two to four centuries) before the Christian 
era. He has had commentators in abundance, and has under- 
gone at their hands some measure of amendment and com- 
pletion; but he has not been overthrown or superseded. 
The chief and most authoritative commentary on his work 
is that called the Mahabhashya, -great comment', in which 
Katyayana's strictures on. his rules are examined arid dis- 
cussed by Patanjali. 

A language, even if not a vernacular one, which is in 
tolerably wide and constant use for writing and speaking, 
is, of course, kept in life principally by direct tradition, by 
communication from teacher to scholar and the study and 
imitation of existing texts, and not by the learning of gram- 
matical rules; yet the existence of grammatical authority, 
and especially of a single one, deemed infallible and of pre- 
scriptive value, could not fail to exert a very strong regu- 
lative influence, leading to the avoidance more and more of 
what was, even if lingering in use, inconsistent with his 
teachings, and also, in the constant reproduction of texts, 
to the gradual effacement of whatever they might contain 
that was unapproved. Thus the whole more modern litera- 
ture of India has been Paninized, so to speak, pressed into 
the mould prepared by him and his school. What are the 


limits of the artificiality of this process is not yet known. 
The attention of special students of the Hindu grammar 
and the subject is so intricate and difficult that the number 
is exceedingly small of those who have mastered it suffi- 
ciently to have a competent opinion on such general matters) 
has been hitherto mainly directed toward determining what 
the Sanskrit according to Panini really is, toward explaining 
the language from the grammar. Arid, naturally enough, 
in India, or wherever else the leading object is to learn to 
speak and write the language correctly that is, as author- 
ized by the grammarians that is the proper course to 
pursue. This, however, is not the way really to understand 
the language. The time must soon come, or it has come 
already, when the endeavor shall be instead to explain the 
grammar from the language; to test in all details, so far 
as shall be found possible, the reason of Panini' s rules 
(which contain not a little that seems problematical, or even 
sometimes perverse) ; to determine what and how much 
genuine usage he had everywhere as foundation, and what 
traces may be left in the literature of usages possessing an 
inherently authorized character, though unratified by him. 

By the term "classical'' or "later" language, then, as 
constantly used below in the grammar, is meant the lan- 
guage of those literary monuments which are written in con- 
formity with the rules of the native grammar : virtually, the 
whole proper Sanskrit literature. For although parts of this 
are doubtless earlier than Panini, it is impossible to tell 
just what parts, or how far they have escaped in their style 
the levelling influence of the grammar. The whole, too, 
may be called so far an artificial literature as it is written 
in a phonetic form (see grammar, 103) which never can 
have been a truly vernacular and living one. Nearly all of 
it is metrical : not poetic works only, but narratives, histories 
(so far as anything deserving that name can be said to exist), 
and scientific treatises of every variety, are done into verse ; 
a prose and a prose literature (except in the commentaries) 
hardly has an existence. Of linguistic history there is next 
to nothing in it all; but only a history of style, and this 


for the most part showing a gradual depravation, an increase 
of artificiality and an intensification of certain more unde- 
sirable features of the language such as the use of pas- 
sive constructions and of participles instead of verbs, and 
the substitution of compounds for sentences. 

This being the condition of the later literature, it is of 
so much the higher consequence that there is an earlier 
literature, to which the suspicion of artificiality does not 
attach, or attaches at least only in a minimal degree, which 
has a truly vernacular character, and abounds in prose as 
well as verse. 

The results of the very earliest literary productiveness 
of the Indian people are the hymns with which, when they 
had only crossed the threshold of the country, and when 
their geographical horizon was still limited to the river- 
basin of the Indus with its tributaries, they praised their 
gods, the deified powers of nature, and accompanied the 
rites of their comparatively simple worship. At what period 
these were made and sung cannot be determined with any 
approach to accuracy: it may have been as early as 2000 
B.C. They were long handed down by oral tradition, pre- 
served by the care, and increased by the additions and 
imitations, of succeeding generations; the mass was ever 
growing, and, with the change of habits and beliefs and 
religious practices, was becoming variously applied sung 
in chosen extracts, mixed with other material into liturgies, 
adapted with more or less of distortion to help the needs 
of a ceremonial which was coming to be of immense elab- 
oration and intricacy. And, at some time in the course 
of this history, there was made for preservation a great col- 
lection of the hymn-material, mainly its oldest and most 
genuine part, to the extent of over a thousand hymns and 
ten thousand verses, arranged according to traditional author- 
ship and to subject and length of hymn : this collection is 
the Rig-Veda, -Veda of verses (re) or hymns'. Other col- 
lections were made also out of the same general mass of 
traditional material : doubtless later, although the inter- 
relations of this period are as yet too unclear to allow of 


our speaking with entire confidence as to anything concern- 
ing them. Thus, the Sama- Veda. 'Veda of chants (saman}\ 
containing only about a sixth as much, its verses nearly all 
found in the Rig-Veda also, but appearing here with nume- 
rous differences of reading; these were passages put together 
for chanting at the soma-sacrifices. Again, collections called 
by the comprehensive name of Yajur-Veda, 'Veda of sac- 
rificial formulas (yajusV: these contained not verses alone, 
but also numerous prose utterances, mingled with the former, 
in the order in which they were practically employed in 
the ceremonies; they were strictly liturgical collections. Of 
these, there are in existence several texts, which have their 
mutual differences: the Vajasaneyi-Samhita (in two slightly 
discordant versions, Madhyandina and Kanvd , sometimes 
also called the White Yajur-Veda ; and the various and 
considerably differing texts of the Black Yajur-Veda. namely 
the Taittirlya-Samhita, the Maitrayam-Samhita, and the 
Kathaka (the two last not yet published). Finally, another 
historical collection, like the Rig-Veda, but made up mainly 
of later and less accepted material, and called (among other 
less current names) the Atharva-Veda, 'Veda of the Ath- 
arvans (a legendary priestly family)'; it is somewhat more 
than half as bulky as the Rig- Veda, and contains a certain 
amount of material corresponding to that of the latter, and 
also a number of brief prose passages. To this last col- 
lection is very generally refused in the orthodox literature 
the name of Veda; but for us it is the most interesting of 
all. after the Rig -Veda, because it contains the largest 
amount of hymn-material (or mantra, as it is called, in 
distinction from the prose brahmana], and in a language 
which, though distinctly less antique than that of the other, 
is nevertheless truly Vedic. Two versions of it are extant, 
one of them only in a single known manuscript. 

A not insignificant body of like material, and of various 
period (although doubtless in the main belonging to the 
latest time of Vedic productiveness, and in part perhaps 
the imitative work of a yet more modern time), is scattered 
through the texts to be later described, the Brahmanas and 


the Sutras. To assemble and sift and compare it is now 
one of the pressing needs of Vedic study. 

The fundamental divisions of the Vedic literature here 
mentioned all have had their various schools of sectaries, 
each of these with a text of its own. showing some differ- 
ences from those of the other schools : but those mentioned 
above are all that are now known to be in existence; and 
the chance of the discovery of others grows every year 

The labor of the schools in the conservation of their 
sacred texts was extraordinary, and has been crowned with 
such success that the text of each school, whatever may 
be its differences from those of other schools, is virtually 
without various readings, preserved with all its peculiarities 
of dialect, and its smallest and most exceptional traits of 
phonetic form, pure and unobscured. It is not the place 
here to describe the means by which, in addition to the 
religious care of the sectaries, this accuracy was secured: 
forms of text, lists of peculiarities and treatises upon them, 
and so on. When this kind of care began in the case of 
each text, and what of original character may have been 
effaced before it, or lost in spite of it, cannot be told. But 
it is certain that the Vedic records furnish, on the whole, 
a wonderfully accurate and trustworthy picture of a form of 
ancient Indian language (as well as ancient Indian beliefs 
and institutions) which was a natural and undistorted one, 
and which goes back a good way behind the classical San- 
skrit. Its differences from the latter the following treatise 
endeavors to show in detail. 

Along with the verses and sacrificial formulas and 
phrases in the texts of the Black Yajur-Veda are given 
long prose sections, in which the ceremonies are described, 
their meaning and the reason of the details and the accom- 
panying utterances are discussed and explained, illustrative 
legends are reported or fabricated, and various speculations, 
etymological and other, are indulged in. Such matter comes 
to be called brahmana (apparently 'relating to the brahman 
or worship';. In the White Yajur-Veda. it is separated into 


a work by itself, beside the samhitci or text of verses and 
formulas, and is called the Catapatha-Brahmana, 'Brahmana 
of a hundred ways'. Other similar collections are found, be- 
longing to various other schools of Vedic study, and they 
bear the common name of Brahmana, with the name of the 
school, or some other distinctive title, prefixed. Thus, the 
Aitar ey a and Kamhitaki- Brahmanas, belonging to the 
schools of the Rig- Veda, the Pancavinqa and Shadvin$a- 
Brahmanas and other minor works, to the Sama-Veda; the 
Gopatha-Brahmana, to the Atharva-Veda ; and a Jaimini- 
Brahmana, to the Sama-Veda, has just (Burnell) been dis- 
covered in India; the Taittirlya-Brahmana is a collection 
of mingled mantra and brahmana, like the samhita of the 
same name, but supplementary and later. These works are 
likewise regarded as canonical by the schools, and are learn- 
ed by their sectaries with the same extreme care which is 
devoted to the samhitas, and their condition of textual 
preservation is of a kindred excellence. To a certain 
extent, there is among them the possession of common 
material: a fact the bearings of which are not yet fully 

Notwithstanding the inanity of no small part of their 
contents, the Brahmanas are of a high order of interest in 
their bearings on the history of Indian institutions; and 
philologically they are not less important, since they re- 
present a form of language in most respects intermediate 
between the classical and that of the Vedas, and offer spe- 
cimens on a large scale of a prose style, and of one which 
is in the main a natural and freely developed one the 
oldest and most primitive Indo-European prose. 

Beside the Brahmanas are sometimes found later ap- 
pendices, of a similar character, called Aranyakas ('forest- 
sections'): as the Aitareya-Aranyaka, Taittirtya-Aranyaka, 
Brhad-Aranyaka, and so on. And from some of these, or 
even from the Brahmanas, are extracted the earliest Upa- 
nishads ('sittings, lectures on sacred subjects') which, 
how r ever, are continued and added to down to a compara- 
tively modern time. The Upanishads are one of the lines 


by which the Brahmana literature passes over into the later 
theological literature. 

Another line of transition is shown in the Sutras ( -lines, 
rules'). The works thus named are analogous with the 
Brahmanas in that they belong to the schools of Vedic 
study and are named from them, and that they deal with 
the religious ceremonies: treating them, however, in the 
way of prescription, not of dogmatic explanation. They, 
too, contain some mantra or hymn-material, not found to 
occur elsewhere. In part ($rauta or kalpa-sutras\ they take 
up the great sacrificial ceremonies, with which the Brah- 
manas have to do ; in part (grhya-sutras), they teach the 
minor duties of a pious householder; in some cases (sa- 
mayacarika-sutras) they lay down the general obligations of 
one whose life is in accordance with prescribed duty. And 
out of the last two, or especially the last, come by natural 
development the law-books (dharma-$astras), which make 
a conspicuous figure in the later literature : the oldest and 
most noted of them being that called by the name of 
Manu (an outgrowth, it is believed, of the Manava Vedic 
school); to which are added that of Yajnavalkya. and many 

Respecting the chronology of this development, or the 
date of any class of writings, still more of any individual 
work, the less that is said the better. All dates given in 
Indian literary history are pins set up to be bowled down 
again. Every important work has undergone so many more 
or less transforming changes before reaching the form in 
which it comes to us, that the question of original con- 
struction is complicated with that of final redaction. It is 
so with the law-book of Manu, just mentioned, which has 
well-founded claims to being regarded as one of the very 
oldest works of the proper Sanskrit literature, if not the 
oldest (it is variously assigned, to periods from six centuries 
before Christ to soon after Christ). It is so, again, in a 
still more striking degree, with the great legendary epic of 
the Mahdbharata. The ground-work of this is doubtless of 
very early date; but it has served as a text into which 



materials of various character and period have been inwoven, 
until it has become a heterogeneous mass, a kind of cyclo- 
pedia for the warrior-caste, hard to separate into its con- 
stituent parts. The story of Nala, and the philosophical 
poem Bhagavad-Glta, are two of the most noted of its 
episodes. The Ramayana, the other most famous epic, 
is a work of another kind: though also worked over and 
more or less altered in its transmission to our time, it is 
the production, in the main, of a single author (Valmiki); 
and it is generally believed to be in part allegorical, re- 
presenting the introduction of Aryan culture and dominion 
into Southern India. By its side stand a number of minor 
epics, of various authorship and period, as the Raghuvah$a 
(ascribed to the dramatist Kalidasa), the Maghakavya, the 
Bhattikavya (the last, written chiefly with the grammatical 
intent of illustrating by use as many as possible of the 
numerous formations which, through taught by the gram- 
marians, find no place in the literature). 

The Purdnas. a large class of works mostly of immense 
extent, are best mentioned in connection with the epics. 
They are pseudo-historical and prophetic in character, of 
modern date, and of very small value. Real history finds 
no place in Sanskrit literature, nor is there any conscious 
historical element in any of the works composing it. 

Lyric poetry is represented by many works, some of 
which, as the Meghaduta and Gitagovinda, are of no mean 
order of merit. 

The drama is a still more noteworthy and important 
branch. The first indications of dramatical inclination and 
capacity on the part of the Hindus are seen in certain 
hymns of the Veda, where a mythological or legendary 
situation is conceived dramatically, and set forth in the 
form of a dialogue well-known examples are the dialogue 
of Sarama and the Panis, that of Yama and his sister Yami, 
that of Vasishtha and the rivers, that of Agni and the other 
gods but there are no extant intermediaries between these 
and the standard drama. The beginnings of the latter date 
from a period when in actual life the higher and educated 


characters used Sanskrit, and the lower and uneducated used 
the popular dialects derived from it, the Prakrits; and their 
dialogue reflects this condition of things. Then, however, 
learning (not to call it pedantry) intervened, and stereotyped 
the new element; a Prakrit grammar grew up beside the 
Sanskrit grammar, according to the rules of which Prakrit 
could he made indefinitely on a substrate of Sanskrit; and 
none of the existing dramas need to date from the time of 
vernacular use of Prakrit, while most or all of them are 
undoubtedly much later. Among the dramatic authors, 
Kalidasa is incomparably the chief, and his Cakuntala as 
distinctly his masterpiece. His date has been a matter of 
much inquiry and controversy; it is doubtless some cen- 
turies later than our era. The only other work deserving 
to be mentioned along with Kalidasa' s is the Mrchakafl of 
Cudraka, also of questionable period, but believed to be 
the oldest of the extant dramas. 

A partly dramatic character belongs also to the fable, 
in which animals are represented as acting and speaking. 
The most noted works in this department are the Panca- 
tantra, which through Persian and Semitic versions has made 
its way all over the world, and contributes a considerable 
quota to the fable-literature of every European language, 
and, partly founded on it, the comparatively recent and 
popular Hitopade$a ('salutary instruction'). 

Two of the leading departments of Sanskrit scientific 
literature, the legal and the grammatical, have been already 
sufficiently noticed; of those remaining, the most important 
by far is the philosophical. The beginnings of philosophic- 
al speculation are seen already in some of the later hymns 
of the Veda, more abundantly in the Brahmanas and Aran- 
yakas, and then especially in the Upanishads. The evo- 
lution and historic' relation of the systems of philosophy, 
and the age of their text-books, are matters on which much 
obscurity still rests. There are six systems of primary rank, 
and reckoned as orthodox, although really standing in no 
accordance with approved religious doctrines. All of them 
seek the same end, the emancipation of the soul from the 



necessity of contiiiuing its existence in a succession of 
bodies, and its unification with the All -soul; but they 
differ in regard to the means by which they seek to attain 
this end. 

The astronomical science of the Hindus is a reflection 
of that of Greece, and its literature is of recent date; but 
as mathematicians, in arithmetic and geometry , they have 
shown more independence. Their medical science, although 
its beginnings go back even to the Veda, in the use of 
medicinal plants with accompanying incantations, is of little 
account, and its proper literature by no means ancient. 







Vowels, 8; Consonants, 11; Quantity, 26; Accent, 27. 


Introductory, 33 ; Principles, 36 ; Rules of Vowel Com- 
bination, 41; Permitted Finals, 46; Deaspiration, 50; 
Surd and Sonant Assimilation, 51 ; Combinations of 
Final s and r, 53 ; Conversion of s to s, 57 ; Con- 
version of n to n, 60 ; Conversion of Dental Mutes to 
Linguals and Palatals, 62; Combinations of Final n, 
63 ; Combinations of Final m, 65 ; the Palatal Mutes 
and Sibilant, and h, 66 ; the Lingual Sibilant, 71 ; 
Extension and Abbreviation, 72; Strengthening and 
Weakening Processes, 73; Guna and Vrddhi, 74; 
Vowel-lengthening, 76 ; Vowel-lightening, 77 ; Nasal 
Increment, 78; Reduplication, 79. 


Gender, Number, Case, 80: Uses of the Cases, 81; 
Endings of Declension, 92; Variation of Stem, 95; 
Accent in Declension, 97. 


Classification etc., 99; Declension I., Stems in a, 100; 
Declension II., Stems in i and u, 104; Declension 
III., Stems in Long Vowels (a, I, u): A. Root-words 
etc., Ill; Stems in Diphthongs, 116; B. Derivative 
Stems etc. 117; Declension IV., Stems in r or ar, 
123; Declension V., Stems in Consonants, 127; 
A. Root-stems etc., 129; B. Derivative Stems in as, 
is, us, 138; C. Derivative Stems in an, 140; D., 
in in, 145; E., in ant or at, 146; F. Perfect Par- 
ticiples in vans, 152; G. Comparatives in yas, 155; 
Comparison, 156. 








xx ii CONTENTS. 


VI. NUMERALS 160167 

Cardinals, 160; Ordinals etc., 166. 

VIE. PRONOUNS 168 181 

Personal, 168 ; Demonstrative, 171 ; Interrogative, 
176; Relative, 177; Emphatic, 179; Nouns used pro- 
nominally, 179; Pronominal Derivatives, Possessives 
179; Adjectives declined pronominally, 181. 


Voice, Tense, Mode, Number, Person, 182; Verbal 
Adjectives and Nouns, 185; Secondary Conjugation, 
185; Personal Endings, 186 ; Subjunctive Mode, 191 ; 
Optative, 193 ; Imperative, 195 ; Uses of the Modes, 
196 ; Participles, 201 ; Augment, 201 ; Reduplication, 
202; Accent of the Verb, 203. 

IX. THE PRESENT-SYSTEM . ;. [ ,,, .T.~ ^ J-rt^ > 207 255 
General, 207; Conjugations and Conjugation Classes, 

208; I. Root-class (second or ad-class), 211; II. Re- 
duplicating Class (third or ftu-class), 221 ; III. Nasal 
Class (seventh or rwd/i-class), 229; IV. Nu and w-Classes 
(fifth and eighth, or su and tan-classes), 232; V. Na- 
Class (ninth or fcrz-class), 238; VI. a-Class (first or 
6ftu-classj, 241 ; VII. Accented d-Class (sixth or tud- 
class), 245; VIII. Fa-Class (fourth or diu-elass), 248; 
IX. Accented yd-Class or Passive Conjugation, 252; 
Uses of the Present and Imperfect, 254. 


Perfect Tense, 255; Perfect Participle, 266; Modes 
of the Perfect, 267; Pluperfect, 269; Uses of the 
Perfect, 270. 


Classification, 271; I. Simple Aorist: 1. Root-aorist, 
273 4 ; Passive Aorist 3d sing., 277; 2. the a-Aorist, 
278; II. 3. Reduplicated Aorist, 281; III. Sibilant 
Aorist, 285; 4. the s-aorist, 286; 5. the is- Aorist, 
290; 6. the >-aorist, 293; 7. the sa-Aorist, 294; 
Precative, 296; Uses of the Aorist, 298. 


I. The s-future, 300; Modes of the 5-future, 302; 
Participles of the s-future, 302; Preterit of the s- 
future: Conditional, 303; II. The Periphrastic Future, 
303 ; Uses of the Futures and Conditional, 305. 

CONTENTS. xxiii 

Chap. Page. 


Passive Participle in ta or nd, 307 ; Past Active Par- 
ticiple in tavant, 310; Future Passive Participles: 
Gerundives, 310; Infinitives, 313; Uses of the Infini- 
tives, 315; Gerunds, 319; Adverbial Gerund in am, 


I. Passive, 322; II. Intensive, 323; Present-System, 
325; Perfect. Aorist, Future, etc., 329; III. Desider- 
ative, 331; Present-System, 334; Perfect, Aorist, 
Future, etc., 335; IV. Causative, 337; Present-System, 
339; Perfect, Aorist, Future, etc., 340; V. Denomi- 
native, 343. 

The Periphrastic Perfect, 347; Participial Periphras- 
tic Phrases, 349 ; Composition with Prepositional 
Prefixes, 350; Other Verbal Compounds, 355. 


Adverbs, 358 ; Prepositions, 366 ; Conjunctions, 369 ; 
Interjections, 369. 

A. Primary Derivatives, 373; B. Secondary Deriva- 
tives, 403. 


Classification, 425; I. Copulative Compounds, 428; 

II. Determinative Compounds, 431 ; A. Dependent 
Compounds, 432; B. Descriptive Compounds, 437; 

III. Secondary Adjective Compounds, 443; A. Pos- 
sessive Compounds, 443; B. Compounds with Governed 
Final Member, 452; Adjective Compounds as Nouns 
and as Adverbs, 453; Anomalous Compounds, 455; 
Stem-finals altered in Composition, 455 ; Irregular 
Construction with Compounds, 456. 

APPENDIX 457 460 

A. Examples of Varines Sanskrit Type 457 ; B. Exam- 
ple of Accentuated Text, 459. 






AB. Aitareya-Brahmana. 

APr. Atharva-Prati^akhya. 

AV. Atharva-Veda. 

BB. Bohtlingk and Roth (Petersburg 


9 or ak. Qakuntala. 
$B. $atapatha-Brahmana. 
QGS. Qankhayana-Grihya-Sutra. 
GB. Gopatha-Brahmana. 
H. Hitopade$a. 
K. Kathaka. 

KB. Kaushitaki-Brahmana. 
KSS. Katha-Sarit-Sagara. 
M. Mann. 
MBh. Mahabharata. 
Megh. Meghadiita. 

MS. Maitrayani-Sauhita. 

PB. Pancavin^a-Brahmana. 

R. Ramayana. 

Ragh. Raghuvan^a. 

RPr. Rigveda-Prati^akhya. 

RV. Rig- Veda. 

SB. Shadvin^a-Brahmaua. 

SV. Sama-Veda. 

TA. Taittiriya-Aranyaka. 

TB. TaittirTya-Brahmana. 

TPr. Taittiriya-Prati^akbya. 

Tribh. Tribhashyaratna. 

TS. TaittirTya-Sanhita. 

V. Veda. 

VPr. Vajasaneyi-Prati^akhyj 

VS. Vajasaneyi-Sanhita. 



1. THE natives of India write their ancient and sacred 
language in a variety of alphabets generally, in each 
part of the country, in the same alphabet which they use 
for their own vernacular. The mode of writing, however, 
which is employed throughout the heart of Aryan India, or 
in Hindustan proper, is alone adopted by European scholars : 
it is called the devanagari. 

This name is of doubtful origin and value. A more comprehensive name 
is nagarl (perhaps, 'of the city'); and deva-nagarl is 'nagarl of the gods,' 
or 'of the Brahmans.' 

2. Much that relates to the history of the Indian alphabets is still 
obscure. The earliest written monuments of known date in the country are 
the inscriptions containing the edicts of A^oka or Piyadasi, of about the 
middle of the third century B. C. They are in two different systems of 
characters, of which one shows distinct signs of derivation from a Semitic 
source, while the other is also probably, though much less evidently, of the 
same origin (Burnell). From the latter, the Lath, or Southern Acoka cha- 
racter (of Girnar), come the later Indian alphabets, both those of the northern 
Aryan languages, and those of the southern Dravidian languages. The 
nagari, devanagari, Bengali, Guzerati, and others, are varieties of its northern 
derivatives; and with them are related some of the alphabets of peoples 
outside of India as in Tibet and Farther India who have adopted Hindu 
culture or religion. 

There is reason to believe that writing was first employed in India for 
practical purposes for correspondence and business and the like and 
only by degrees came to be applied also to literary use. The literature, to 
a great extent, and the more fully in proportion to its claimed sanctity and 
authority, ignores all written record, and assumes to be kept in existence by 
oral tradition alone. 

3. Of the devanagari itself there are minor varieties, depending on 
differences of locality or of period, as also of individual hand (see examples 

Whitney, Grammar. 1 


in Weber's catalogue of the Berlin Sanskrit MSS., in Rajendralala Mitra's 
notices of MSS. in Indian libraries, in the published fac-similes of in- 
scriptions, and so on); and these are in some measure reflected in the type 
prepared for printing, both in India and in Europe. But a student who 
makes himself familiar with one style of printed characters will have little 
difficulty with the others, and will soon learn, by practice, to read the manu- 
scripts. A few specimens of types other than those used in this work will 
be given in an Appendix. 

On account of the difficulty of combining them with the smaller sizes 
of our Roman and Italic type, the devanagari characters will be used below 
only in connection with the first or largest size. And, in accordance with 
the laudable usage of recent grammars, they will, wherever given, be also 
transliterated in italic letters; while the latter alone will be used in the 
other sizes. 

4. The student may be advised to try to familiarize himself 
from the start with the devanagarl mode of writing. At the same 
time, it is not necessary that he should do so until, having 
learned the principal paradigms, he comes to begin reading and 
analysing and parsing ; and many will find the latter the more 
practical, and in the end equally or more effective, way. 

5. The characters of the devanagarl alphabet, and the 
European letters which will be used in transliterating them, 
are as follows : 

short. long. 

Vowels : 

diph- ( palatal 
thongs I labial 

$ l 

3 u 

^rd * 


^ e 

37 o 



, n or m (see 73) 








q p 

surd asp. 

is 1^ kh 

23 3T Ch 
2S % th 
33 5T th 

ss Cfi ph 


19 *T g 

24 Sf j 

29 3" d 
34 $ d 

39 ST b 

son. asp. 

o q gh 
5 3T jh 


26 3T n 

dh 31 HI n 

dh se ^ n 

bh 4i IT m 


42 T y 


43 ^ r 


44 ^T / 


4S Sf V 

[ palatal 
1 dental*" 

47 ^ s 




Aspiration 49 ^T h 

To these may be added a lingual I "3o , which in some of 
the Vedic texts takes the place of 3 d when occurring between 
two vowels (54). 

6. A few other sounds, recognized by the theories of the 
Hindu grammarians, but either having no separate characters to 
represent them, or only very rarely and exceptionally written, 
will be noticed below (71, 230). Such are the guttural and 
labial breathings, the nasal semivowels, and others. 

7. The order of arrangement given above is that in 
which the sounds are catalogued and described by the native 
grammarians ; and it has been adopted by European scholars 
as the alphabetic order, for indexes, dictionaries, etc. (to 
the Hindus, the idea of an alphabetic arrangement for such 
practical uses is wanting). 

In some works (as the Petersburg lexicon), a visarga which is regarded 
as equivalent to and exchangeable with a sibilant (172) is, though written 
as visarga, given the alphabetic place of the sibilant. 

8. Tbe theory of the devanagari, as of the other In- 
dian modes of writing, is syllabic and consonantal. 
That is to say, it regards as the written unit, not the simple 
sound, but the syllable (aksara); and further, as the sub- 
stantial part of the syllable, the consonant (or the consonants) 
preceding the vowel - - this latter being merely implied, 
or. if written, being written by a subordinate sign attached 
to tbe consonant. 

9. Hence follow these two principles: 

A. Tbe forms of the vowel-characters given in the 


alphabetic scheme above are used only when the vowel 
forms a syllable by itself, or is not combined with a pre- 
ceding consonant: that is, when it is initial, or preceded 
by another vowel. In combination with a consonant, other 
modes of representation are used. 

B. If more consonants than one precede the vowel, 
forming with it a single syllable, their characters must be 
combined into a single compound character. 

Ordinary Hindu usage does not divide the words of a sen- 
tence, any more than the syllables of a word ; a final consonant 
is combined into one syllable with the initial vowel or conso- 
nant of the next following word. 

10. Under A, it is to be noticed that the modes of 
indicating a vowel combined with a preceding consonant 
are as follows: 

a. The short ^ a has no written sign at all: the con- 
sonant-sign itself implies a following T a, unless some other 
vowel-sign is attached to it (or else the virama: 11). Thus 
the consonant-signs as given above in the alphabetic scheme 
are really the signs of the syllables ka, kha, etc. etc. (to ha]. 

b. The -long ^T a is written by a perpendicular stroke 
after the consonant: thus, 5fiT ka, EfT dha, ^T ha. 

c. Short ^ i and long ^ e, by a similar stroke, which 
for short i is placed before the consonant and for long I is 
placed after it, and in either case is connected with the 
consonant by a hook above the upper line: thus, f^R ki, 
3ft ki', Pr bhi, Ht bhi ; ft m', jft nl. 

The hook above, turning to the left or to the right, is historically the 
essential part of the character, having been originally the whole of it; the 
hooks were only later prolonged, so as to reach all the way down beside 
the consonant. In the MSS., they almost never have the horizontal stroke 
drawn across them above, though this is added in all the printed forms of 
the characters*. 

* Thus, originally dfi W, ofj it; in the MSS., jcfj, effj ; in print, 


d. The w-sounds, short and long, are written by hooks 
attached to the lower end of the consonant-sign: thus, Sfj 
ku, 3\ ku; I du, I du. On account of the necessities of 

6\ O SX 

combination, du and du are somewhat disguised: thus, If. 
^; and the forms with ^ r and ^T h are still more irre- 
gular : thus, "^\ ru, % ru ; <^T hu, f^ 1 hu. 

e. The r- vowels, short and long, are written by a sub- 
joined hook, single or double, opening toward the right: 
thus, ^\ kr, Sfj kf ; dr, ^ dr. In the /j-sign, the hooks 
are usually attached to the middle: thus, ^ hr, ^ hr. 

As to the combination of r with preceding r, see below, 14. 

f. The /-vowel is written with a reduced form of its 
full initial character: thus, efi kl: the corresponding long 
has no real occurrence (23), but would be written with a 
similar reduced sign. 

g. The diphthongs are written by strokes, single or 
double, above the upper line, combined, for Jt o and ETF 
aUj with the #-sign after the consonant : thus. 3ft ke, fi 
kai; ^t ko, 

In some devanagari MSS. (as in the Bengali alphabet), the single stroke 
above, or one of the double ones, is replaced by a sign like the a-sign 
before the consonant : thus, (off fee, \fi\ feat, fofil feo, |cftl feaw. 

11. A consonant -sign, however, is capable of being 
made to signify the consonant-sound alone, without an 
added vowel, by having written beneath it a stroke called 
the virama ('rest, stop'): thus, fi k^ < d, ^ h. 

Since, as was pointed out above, the Hindus write the words of a 
sentence continuously, like one word (9, end), the virama is in general called 
for only when a final consonant occurs before a pause. But it is also occasion- 
ally resorted to by scribes, or in print, in order to avoid an awkward or 
difficult combination of consonant-signs; and it is used freely in published 
texts which for the convenience of beginners have their words printed sepa- 

12. Under B, it is to be noticed that the consonant 
combinations are for the most part not at all difficult to 

6 I. ALPHABET. [12 

make or to recognise for one who is familiar with the 
simple signs. The characteristic part of a consonant-sign 
that is to be added to another is taken (to the exclusion 
of the horizontal or of the perpendicular framing-line, or 
of both), and they are put together according to conveni- 
ence, either side by side, or one above the other: in some 
combinations either arrangement is allowed. The consonant 
that is to be pronounced first is set before the other in the 
one order, and above it in the other order. 

Examples of the side-by-side arrangement are : ITf gga, 
ST jja, C?j pya, ZTJ nma, f^T tiha, ^T bhya, F^I ska, STU ma, 
f3T tka. 

Examples of the above-and-below arrangement are : 
^T Ma, sf cca, ^ nja, ^ dda, H pta, ^ tna. 

13. In some cases, however, there is more or less 
abbreviation or disguise of the independent form of a con- 
sonant-sign in combination. 

Thus, of 3\ k in "^7 kid, ^\ Ida; and in ^HT kna etc.; 

of rT t in fF tta; 

of ^ d in "% dga, ^ dna, etc.; 

of *T m and ZT y, when following other consonants : 
thus, ^T kya, 3R krna, ^ nma, 2T nya, ^ dma, t% dya, ^T 
hma, ^T hya. "5T chya, ^ dhya ; 

of 5T p, which generally becomes $T when followed 
by a consonant: thus, 31 pea, W praa, ^j $va, V3J $ya. The 
same change is usual when a vowel-sign is added below: 
thus, 5T pw, 5T cr. 

O (. 

Other combinations, of not quite obvious value, are 
ST nna, ^T //a, ^ ddha, ^ dbha, ^ sta, Tg stha; and the 
compounds of ^ h: as ^r 7m, ^ A^a. 

In a case or two, no trace of the constituent letters is 
recognisable : thus, ^ ksa, ^ jfia. 

14. The semivowel ^ r, in making combinations with 


other consonants, is treated in a wholly peculiar manner, 
analogous with that of the vowels. If pronounced before 
another consonant (or consonant-combination), it is written 
with a hook above, opening to the right (like the subjoined 
sign of r: 10 e): thus. R rka, ^ rsa (fP rtsna). If pro- 
nounced after another consonant (alone or in combination), 
it is written with a slanting stroke below: thus, El gra, 
Pf pra, T sra (and CET grya, T srva); and, with modifica- 
tions of the preceding consonant-sign like those noted above, 
"5T tra, "5T pr, ?T dra. 

When }T r is to be combined with a following 5fJ r, it 
is the vowel which is written in full, with its initial char- 
acter, and the consonant in subordination to it: thus, 
ft rr. 

15. Further combinations, of three, or four, or even 
five consonant-signs, are made according to the same rules. 
Examples are: 

of three consonants, f[ ttva, 1ST ddhya, $3 dvya, ^ 
dry a, SET dhrya, t3 psva, 5JT 9^ya, ^J sty a, ^1 hvya; 

of four consonants, ^J ktrya, ^T nksya, ^J strya, 
rF?I tsmya; 

of five consonants, fF^U rtsnya. 

The manuscripts, and the type-fonts as well, differ from one another 
more in their management of consonant combinations than in any other 
respect, often having peculiarities which one needs a little practice to under- 
stand. It is quite useless to give in a grammar the whole series of possible 
combinations (many of them excessively rare) which are provided for in any 
given type-font, or even in all. There is nothing which due familiarity 
with the simple signs and with the above rules of combination will not enable 
the student to analyse and explain. 

16. A sign called the avagraha ('separator') namely, 
vl - - is used in the manuscripts, sometimes in the manner 
of a hyphen, sometimes as a mark of hiatus, sometimes to 
mark the elision of initial 51 a after final ^ e or sqj o (135). 
In printed texts, especially European, it is ordinarily limited 


to the use last mentioned: thus. ?t ^^f^ te l bruvan, HT 
so 'bravit, for te abruvan, so abramt. 

The sign is used to mark an omission of something. 
In some texts, it has also the value of a hyphen. 

Signs of punctuation are I and II. 

17. The numeral figures are 
1 1, ^ 2, \ 3, 9 4, H 5, | 6, b 7. TT 8. $ 9, 0. 

In combination, to express larger numbers, they are 
used in precisely the same way with European digits : thus, 
^H 25, ^0 630, <(000 1000, ^T7b 1879. 

18. The Hindu grammarians call the different sounds, and 
the characters representing them, by a kara ('maker') added to 
the sound of the letter, if a vowel, or to the letter followed by 
a, if -a consonant. Thus, the sound or character a is called 
akara; k is kakara: and so on. But the kara is also omitted, 
and a, ka, etc. are used alone. The r, however, is never called 
rakara, but only ra or repha ('snarl' : the only example of a 
specific name for an alphabetic element of its class). The ami- 
svara and msarga are also known by these names alone. 



I. Vowels. 

19. THE a, i, and ^^-vowels. The Sanskrit has these 
three earliest and most universal vowels of Indo-European 
language, in both short and long form % a and TT a, 
^ t and ^ I. 3 u and 3T u. They are to be pronounced in 
the "Continental" or "Italian" manner as in far or father, 
pin and pique, pull and rule. 

20. The a is the openest vowel, an utterance from the ex- 

24] VOWELS. 9 

panded throat ; it stands, therefore., in no relation of kindred 
with any of the classes of consonantal sounds. The i and u are 
close vowels, made with marked approach of the articulating 
organs to one another : i is palatal, and shades through y into 
the palatal and guttural consonant-classes ; u is similarly related, 
through v, to the labial class, as involving in its utterance a 
narrowing and rounding of the lips. 

The Paninean scheme (commentary to Panini's grammar, i. 1. 9) classes a 
as guttural, but apparently only in order to give that series as well as the 
rest a vowel: no one of the Praticakhyas puts a into one class with k etc. 
All these authorities concur in calling the i and w-vowels respectively palatal 
and labial. 

21. The short a is not pronounced in India with the full 
openness of a, as its corresponding short, but usually as the 
"neutral vowel" (English so-called "short ", of but, son, blood, 
etc.). This peculiarity appears very early, being acknowledged 
by Panini and by two of the Praticakhyas (APr. i. 36 ; VPr. i. 
72), which call the utterance samvrta, 'covered up, dimmed'. 
It is, however, of course not original ; and it is justly wont to 
be ignored by Western scholars (except those who have studied 
in India). 

22. The a- vowels are the prevailing vowel-sounds of the 
language, being about twice as frequent as all the others (in- 
cluding diphthongs) taken together. The -vowels, again, are 
about twice as numerous as the w-vowels. And, in each pair, 
the short vowel is more than twice (2y 2 to 3 times) as common 
as the long. 

For more precise estimates of frequency, of these and of the other 
alphabetic elements, and for the way in which they were obtained, see 
below, 75. 

23. The r and /-vowels. To the three simple vo- 
wels already mentioned the Sanskrit adds two others, the 
r-vowels and the /-vowel, both of them plainly generated 
by the abbreviation of syllables containing a ^" r or ^T / 
along with another vowel : the TR r coming (almost always : 
see 237, 241-3) from T|" ar or f ra, the FT / from FT al. 

Some of the Hindu grammarians add to the alphabet also a long I ; 
but this is only for the sake of an artificial symmetry, since the sound does 
not occur in a single genuine word in the language. 

24. The vowel :fj r is simply a smooth or untrilled 
r-sound, assuming a vocalic office in syllable-making 



as, by a like abbreviation, it has done also in certain Sla- 
vonic languages. The vowel FT I is an -sound similarly 
uttered like the English /-vowel in such words as able, 
angle, addle. 

The modern Hindus pronounce these vowels as ri, ri, li 
;or even Iri), having long lost the habit and the facility of 
giving a vowel value to the pure r and ^-sounds. Their example 
is widely followed by European scholars ; and hence also the 
(distorting and quite objectionable) transliterations ri, n, li. 
There is no -real difficulty in acquiring and practising the true 

Some of the grammarians (see APr. i. 37, note) attempt to define more 
nearly the way in which, in these vowels, a real r or ^-element is combined 
with something else. 

25. Like their corresponding semivowels, r and I, these 
vowels belong respectively in the general lingual and dental class- 
es ; the euphonic influence of r and f (180) shows this clearly. 
They are so ranked in the Paninean scheme ; but the Praticakhyas 
in general strangely class them with the jihvcimuliya sounds, our 

26. The short r is found in every variety of word and of 
position, and is not rare, being just about as frequent as long u. 
Long f is very much more unusual, occurring only in certain 
plural cases of noun-stems in r (374, 378). The / is met with 
only in some of the forms and derivatives of a single not very 
common verbal root (kip). 

27. The diphthongs. Of the four diphthongs, two, 
the ^ e and ETT o, are in great part original Indo-European 
sounds. In the Sanskrit, they wear the aspect of being 
products of the increment or strengthening of ^ i and 3 u 
respectively; and they are called the corresponding guna- 
vowels to the latter (see below, 235). The other two, ^ ai 
and t au, are by the prevalent and preferable opinion held 
to be of peculiar Sanskrit growth (there is no certain trace 
of them to be found even in the Zend); they are also in 
general results of another and higher increment of ^ i and 
3 u, to which they are called the corresponding vrddhi- 
vowels (below, 235). But all are likewise sometimes gene- 


rated by euphonic combination (127); and TT o, especially, 
is common as result of the alteration of a final *3R as 175). 

28. The ^ e and 3TF o are, both in India and in Eu- 
rope, usually pronounced as they are transliterated that 
is, as long e (English "long ", or e in they] and o-sounds, 
without diphthongal character. 

Such they apparently already were to the authors of the 
Praticakhyas, which, while ranking them as diphthongs \san- 
dfa/afaara), give rules respecting their pronunciation in a manner 
implying them to be virtually unitary sounds. But their euphonic 
treatment (131-4) clearly shows them to have been still at the 
period when the euphonic laws established themselves, as they 
of course were at their origin, real diphthongs , ai (a -f- *) and 
au \a-\- tf). From them, on the same evidence, the heavier or 
vrddhi diphthongs were distinguished by the length of their a- 
element, as ai (a -{- i] and au (a -\- u). 

The recognisable distinctness of the two elements in the vrddhi-diph- 
thongs is noticed by the Praticakhyas (see APr. i. 40, note); but the relation 
of those elements is either defined as equal, or the a is made of less quan- 
tity than the i and u. 

29. The lighter or ywwa-diphthongs are much more frequent 
(6 or 7 times) than the heavier or vrddhi- diphthongs, and the 
e and ai than the o and au (a half more). Both pairs are 
somewhat more than half as common as the simple i and u- 
vowels . 

30. The general name given by the Hindu grammarians to the vowels 
is suara, 'tone'; the simple vowels are called samanaksara, 'homogeneous 
syllable', and the diphthongs are called sandhyaksara, i combination-syllable'. 
The position of the organs in their utterance is defined to be one of openness, 
or of non-closure. 

As to quantity and accent, see below. 76 ff. ; 80 ff. 

II. Consonants. 

31. The Hindu name for 'consonant' is vyan/ana, 'mani- 
fester'. The consonants are divided by the grammarians into 
sparca, i contact' or 'mute', antahstha, 'intermediate' or i semivowel', 
and usman, 'spirant'. They will here be taken up and described 
in this order. 

32. Mutes. The mutes, sparca, are so called as involving 
a complete closure or contact \sparca], and not an approximation 


only, of the mouth-organs by which they are produced. They are 
divided into five classes or series (varga], according to the organs 
and parts of organs by which the contact is made ; and each 
series is composed of five members, differing according to the 
accompaniments of the contact. 

33. The five mute-series are called respectively guttural, 
palatal, lingual (or cerebral), dental, and labial; and they 
are arranged in the order as just mentioned, beginning with 
the contact made furthest back in the mouth, coming for- 
ward from point to point, and ending with the frontmost 

34. In each series there are two surd members, two 
sonant, and one nasal (which is also sonant): for example, 
in the labial series, tj^p and Vf^ph, 3^b and ^bh, and ?T m. 

The members are by the Hindu grammarians called respectively 'first', 
'second', 'third', 'fourth 1 , and 'last' or 'fifth'. 

The surd consonants are known as aghosa, 'toneless', and the sonants 
as ghosavant, 'having tone' ; and the descriptions of the grammarians are in 
accordance with these terms. All alike recognise a difference of tone, and 
not in any manner a difference of force, whether of contact or of expulsion, 
as separating the two great classes in question. That the difference depends 
on vivara, 'opening', or sarhvara, 'closure' (of the glottis), is also recognised 
by them. 

35. The first and third members of each series are the 
ordinary corresponding surd and sonant mutes of European 
languages: thus, ^ k and \g, t^t and d, q p and ^ b. 

36. Nor is the character of the nasal any more doubtful. 
What q^ m is to ^p and 3[^, or ^n to t^t and < d, that 
is also each other nasal to its own series of mutes : a sonant 
expulsion into and through the nose, while the mouth- 
organs are in the mute-contact. 

The Hindu grammarians give distinctly this definition. The nasal 
(anunasika, 'passing through the nose') sounds are declared to be formed by 
mouth and nose together; or their nasality (anunasikya) to be given them 
by unclosure of the nose. 

37. The second and fourth of each series are aspirates : 
thus, beside the surd mute ^ k we have the corresponding 


surd aspirate ^ kh, and beside the sonant Tf.^, the corres- 
ponding sonant aspirate % gh. Of these, the precise char- 
acter is more obscure and difficult. 

That the aspirates, all of them, are real mutes or contact sounds, and 
not fricatives (like European th and ph and ch, etc.), is beyond question. 

It is also not doubtful in what way the surd M, for example, differs 
from the unaspirated t: such aspirates are found in many Asiatic languages, 
and even in some European : they involve the slipping-out of an audible bit 
of flatus or aspiration between the breach of mute-closure and the following 
sound, whatever it may be. They are accurately enough represented by the 
th etc., with which, in imitation of the Latin treatment of the similar ancient 
Greek aspirates, we are accustomed to write them. 

The sonant aspirates are generally understood and described as made 
in a similar way, with a perceptible ft-sound after the breach of sonant 
mute-closure. But there are insuperable theoretical difficulties in the way 
of accepting this explanation ; and some of the best phonetic observers (as 
A. J. Ellis) deny that the modern Hindu pronunciation is of such a character, 
and define the element following the mute as a "glottal buzz", rather, or an 
emphasized utterance of the beginning of the suceeding sound. The question 
is one of great difficulty, and upon it the opinions of the highest authorities 
are still much at variance. Sonant aspirates are still in use in India, in 
the pronunciation of the vernacular as well as of the learned languages. 

By the Pratic.akhyas, the aspirates of both classes are called sosman: 
which might mean either 'accompanied by a rush of breath' (taking usman 
in its more etymological sense), or 'accompanied by a spirant' (below, 59). 
And some authorities define the surd aspirates as made by the combination 
of each surd non-aspirate with its own corresponding surd spirant; and the 
sonant aspirates, of each sonant non-aspirate with the sonant spirant, the 
ft-sound (below, 65). But this would make the two classes of aspirates of 
quite diverse character, and would also make th the same as ts, th as ts, ch 
as cf which is in any measure plausible only of the last. Panini has no 
name for aspirates ; the scheme given in his comment (to i. 1 . 9) attributes 
to them mahaprana, 'great expiration 1 , and to the non-aspirates alpaprana^ 
'small expiration'. 

It is usual among European scholars to pronounce both 
classes of aspirates as the corresponding non-aspirates with 
a following h: for example, 2T th nearly as in English boat- 
hook, m ph as in haphazard, T dh as in madhouse, and so 
on. This is (as we have seen above) confessedly accurate 
only as regards the surd aspirates. 

38. The sonant aspirates are (in the opinion of most), or 
at least represent, original Indo-European sounds, while the surd 



aspirates are generally regarded as a special Indian development. 
The former are more than twice as common as the latter. The 
unaspirated (non- nasal) mutes are very much more frequent 
(5 times) than the aspirates (for the special frequency of bh and 
original gh, see 50 and 66) ; and among them the surds are more 
numerous (2Y 2 times) than the sonants. The nasals (chiefly n 
and m) are nearly as frequent as the surd non-aspirates. 
We take up now the several mute-series. 

39. Guttural series: ^ k, T^kh, JT^, \gh, 3" n. 
These are the ordinary European k and ^-sounds, with their 
corresponding aspirates and nasal (the last, like English ng 
in singing}. 

The gutturals are defined by the Pratic.akhyas as made by contact of 
the base of the tongue with the base of the jaw, and they are called, from 
the former organ, jihvamullya, 'tongue-root sounds'. The Paninean scheme 
describes them simply as made in the throat (kantha). From the euphonic 
influence of a k on a following s (below, 180), we may perhaps infer that 
in their utterance the tongue was well drawn back into the hinder mouth. 

40. The k is by far the commonest of the guttural series, 
occurring considerably more often than all the other four taken 
together. The nasal, except as standing before one of the others 
of the series, is found only as final (after the loss of a fol- 
lowing k), and in a very small number of words. 

41. The Sanskrit guttural series represents only a minority 
of Indo-European gutturals ; these last have suffered more and 
more general corruption than any other class of consonants. By 
processes of alteration which are proved to have begun in the 
Indo-European period, since the same words exhibit connected 
changes also in other languages of the family, the palatal mutes, 
the palatal sibilant c, and the aspiration h, have come from 
gutturals. See these various sounds below. 

42. Palatal series : r( c, ^ ch, s?F j, 37 jh, 3T n. This 
whole series is derivative, being generated by the corruption of 
original gutturals. The c comes from an original k as does 
also, by another degree of alteration, the palatal sibilant c (see 
below, 64). The /, in like manner, comes from a g ; but the 
Sanskrit j includes in itself two degrees of alteration, one cor- 
responding to the alteration of k to c, the other to that of k to g 
(see below, 219 : in the Zend, these two degrees are held dis- 
tinctly apart). The c is somewhat more common than the j 
(about as four to three). The aspirate ch is very much less fre- 
quent (a tenth of c), and comes from the original group sk. 
The sonant aspirate jh is excessively rare (occurring but once 


in the Vedic texts, and not half-a-dozen times in the Brahma- 
nas) ; where found, it is either onomatopoetic or of anomalous 
or not Indo-European origin iin the so-called root ujh, it comes 
from j and h}. The nasal, n, never occurs except immediately 
before or, in a small number of words, also after (201) 
one of the others of the series. 

43. Hence, in the euphonic processes of the language, the 
treatment of the palatals is in many respects peculiar. In some 
situations, the original unaltered guttural shows itself or, as 
it appears from the point of view of the Sanskrit, the palatal 
reverts to its original guttural. No palatal ever occurs as a final. 
The j is differently treated, according as it represents the one 
or the other degree of alteration. And c and j (except artificially, 
in the algebraic rules of the grammarians) do not interchange, 
as corresponding surd and sonant. 

44. The palatal mutes are by European scholars, as by 
the modern Hindus also, pronounced with the compound 
sounds of English ch and j (in church and judge}. 

Their description by the old Hindu grammarians, however, gives them 
a not less absolutely simple character than belongs to the other mutes. 
They are called talavya, 'palatal 1 , and declared to be formed against the 
palate by the middle of the tongue. They seem to have been, then, 
brought forward in the mouth from the guttural point, and made against the 
hard palate at a point not far from the lingual one (below, 45), but with 
the upper flat surface of the tongue instead of its point. Such sounds, in 
all languages, pass easily into the (English) ch and ./-sounds. The value 
of the ch as making the preceding vowel "long by position" (227), and its 
frequent origination from t -f- p (203), lead to the suspicion that it. at least, 
may have had this character from the beginning : compare 37, above. 

45. Lingual series: t, 3" th, Z d, ?o dh, HT n. The 
lingual mutes are by all the native authorities denned as 
uttered with the tip of the tongue turned up and drawn 
back into the dome of the palate (somewhat as the usual 
English smooth r is pronounced). They are called by the 
grammarians murdhanya, literally 'head-sounds, capitals, 
cephalics'; which term is in many European grammars 
rendered by 'cerebrals' . In practice, among European Sans- 
kritists, no attempt is made to distinguish them from the 
dentals : t is pronounced like rT t, J d like e[" d, and so 
w.ith the rest. 


46. The linguals are another non-original series of sounds, 
coming mainly from the phonetic alteration of the next series, 
the dentals, but also in part occurring in words that have no 
traceable Indo-European connection, and are perhaps derived from 
the aboriginal languages of India. The tendency to lingualization 
is a positive one in the history of the language : dentals easily 
pass into linguals under the influence of contiguous or neighbor- 
ing lingual sounds, but not the contrary ; and all the sounds 
of the class become markedly more frequent in the later litera- 
ture. The conditions of their ordinary occurrence are briefly 
these : a. s comes from s, much more rarely from q, j, ks, in 
euphonic circumstances stated below (180, 218, etc.) ; b. a 
dental mute following s is assimilated to it, becoming lingual 
(t, th, n) ; c. n is often changed to n after a lingual vowel or 
semivowel or sibilant in the same word (189 etc.) ; d. dh, which 
is of very rare occurrence, comes from assimilation of a dental 
after s (198 a) or h (222); e. t and d come occasionally by 
substitution for some other sound which is not allowed to stand 
as final (142, 145). When originated in these ways, the lingual 
letters may be regarded as normal ; in any other cases of their 
occurrence, they are either products of abnormal corruption, or 
signs of the non-Indo-European character of the words in which 
they appear. 

In a certain number of passages numerically examined (below, 75), 
the abnormal occurences of lingual mutes were less than half of the whole 
number (74 out of 159), and most of them (43) were of n: all were found 
more frequent in the later passages. In the Rig- Veda, only 15 words have 
an abnormal t; only 6, such a th; only 1, such a dh; about 20 (including 
9 roots, nearly all of which have derivatives) show an abnormal d, besides 
9 that have nd; and 30 (including 1 root) show a n. 

Taken all together, the linguals are by far the rarest class 
of mutes (about iy 2 per cent, of the alphabet) hardly half 
as frequent even as the palatals. 

47. Dental series: <^t, Z^tk, <[ d, V^dk, ^n. These 
are called by the Hindus also dantya, 'dental', and are 
described as formed at the teeth (or at the roots of the 
teeth), by the tip of the tongue. They are practically the 
equivalents of our European t, d, n. 

But the modern Hindus are said to pronounce their dentals with the 
tip of the tongue thrust well forward against the upper teeth, so that these 
sounds get a slight tinge of the quality belonging to the English and Modern 
Greek tft-sounds. The absence of that quality in the European (especially 


the English) dentals is doubtless the reason why to the ear of a Hindu the 
latter appear more analogous with his linguals, and he is apt to use the 
linguals in writing European words. 

48. The dentals are one of the three Indo-European original 
mute-classes. In their occurrence in Sanskrit they are just about 
as frequent as all the other four classes taken together. 

49. Labial series: q p Cfi ph ^ b. ^ bh, 3} m. 

X ^ -X * *X -X ^ ^X 

These sounds are called osthya, 'labial 1 , by the Hindu gram- 
marians also. They are, of course, the equivalents of our 
p, b, m. 

50. The numerical relations of the labials are a little pe- 
culiar. Owing to the absence (or almost entire absence) of b in 
Indo-European, the Sanskrit b also is greatly exceeded in fre- 
quency by bh, which is the most common of all the sonant 
aspirates, as ph is the least common of the surd. The nasal m (not- 
withstanding its frequent euphonic mutations when final: 212 ff.) 
occurs just about as often as all the other four members of the 
series together. 

51. Semivowels: f y, ^" r, s$ I, 3 v. The name given to this 
class of sounds by the Hindu grammarians is antahstha, "standing between' 
either from their character as utterances intermediate between vowel and 
consonant, or (more probably) from the circumstance of their being placed 
between the mutes and spirants in the arrangement of the consonants. 

The semivowels are clearly akin with the several mute series 
in their physical character, and they are classified along with 
those series though not without some discordances of view 
- by the Hindu grammarians. They are said to be produced 
with the organs "slightly in contact" (isatsprsta), or "in imperfect 
contact" (duhsprsta] . 

52. The ^ r is clearly shown by its influence in the 
euphonic processes of the language to be a lingual sound, 
or one made with the the tip of the tongue turned up into 
the dome of the palate. It thus resembles the English 
smooth r, and, like this,, seems to have been untrilled. 

The Paninean scheme reckons r as a lingual. None of the Praticakhyas, 
however, does so; nor are they entirely consistent with one another in its 
description. For the most part, they define it as made at 'the roots of the 
teeth'. This would give it a position like that of the vibrated r; but no au- 
thority hints at a vibration as belonging to it. 

Whitney, Grammar. 2 


In point of frequency, r stands very high on the list of 
consonants ; it is about equal with v, n, m, and y, and only 
exceeded by t. 

53. The T I is a sound of dental position, and is so 
defined and classed by all the native authorities. 

The peculiar character of an Z-sound, as involving expulsion at the 
sides of the tongue along with contact at its tip, is not noticed by any Hindu 

It is a disputed question whether r and I were distinguished from one 
another in Indo-European speech ; in the Sanskrit, at any rate, they are very 
widely interchangeable, both in roots and in suffixes: there is hardly a root 
containing an I which does not show also forms with r; words written with 
the one letter are found in other texts, or in other parts of the same texts, 
written with the other. In the later periods of the language they are more 
separated, and the I becomes decidedly more frequent, though always much 
rarer than the r (only as 1 to 7 or 8 or 10). 

54. Some of the Vedic texts have another -sound, written 
with a slightly different character (it is given at the end of the 
alphabet, 5), which is substituted for a lingual d (as also the 
same followed by h for a dh] when occurring between two vowels. 
It is, then, doubtless a lingual I, one made by breach (at the 
sides of the tongue) of the lingual instead of the dental mute- 

55. The <ET y in Sanskrit, as in other languages gene- 
rally, stands in the closest relationship with the vowel ^ i 
(short or long); the two exchange with one another in 
cases innumerable. 

And in the Veda (as the metre shows) an i is very often to be read 
where, in conformity with the rules of the later Sanskrit euphony, a y is 
written. Thus, the final i- vowel of a word remains i before an initial vowel ; 
that of a stem maintains itself unchanged before an ending; and an ending 
of derivation as ya, tya has i instead of y. Such cases will be con- 
sidered in more detail later, as they arise. The constancy of the phenomenon 
in certain words and classes of words shows that this was no merely optional 
interchange. Very probably, the Sanskrit y had everywhere more of an 
i-character than belongs to the corresponding European sound. 

56. The y is by its physical character a palatal utterance ; 
and it is classed as a palatal semivowel by the Hindu phonetists. 
It is one of the most common of Sanskrit sounds. 

57. The of v is pronounced as English or French v 
(German w) by the modern Hindus except when pre- 


ceded by a consonant in the same syllable, in which case 
it has the sound of English w; and European scholars follow 
the same practice (with or without the same exception). 

By its whole treatment in the euphony of the language, 
however, the v stands related to an w- vowel precisely as y to 
an /-vowel. It is, then, a v only according to the original 
Roman value of that letter that is to say, a z^-sound in the 
English sense : though (as was stated above for the y] it may 
well have been less markedly separated from u than English w, 
more like French ou in oui etc. But, as the original w has in 
most European languages been changed to v (English), so also 
in India, and that from a very early time : the Paninean scheme 
and two of the Praticakhyas (VPr. and TPr.) distinctly define 
the sound as made between the upper teeth and the lower lip 
which, of course, identifies it with the ordinary modern v-sound. 
As a matter of practice, the usual pronunciation need not be 
seriously objected to ; yet the student should not fail to note 
that the rules of Sanskrit euphony and the name of "semivowel" 
have no application except to a zp-sound in the English sense : 
a v-sound (German w) is no semivowel, but a spirant, standing 
on the same articulate stage with the English M-sounds and 

58. V is classed as a labial semivowel by the Hindu 
phonetical authorities. It has a somewhat greater frequency 
than the y. 

In the Veda, under the same circumstances as the y (above, 55), v is to 
be read as w. 

59. Spirants. Unjler the name usman (literally 'heat, 
steam, flatm\ which is usually and well represented by 
'spirant', some of the Hindu authorities include all the 
remaining sounds of the alphabet; others apply the term 
only to the three sibilants and the aspiration to which 
it will here also be restricted. 

The term is not found in the Paninean scheme; by different treatises 
the guttural and labial breathings, these and the visarga, or all these and anu- 
svara, are called usman (see APr. i. 31 note). The organs of utterance are 
described as being in the position of the mute-series to which each spirant 
belongs respectively, but unclosed, or unclosed in the middle. 

60. The H s. Of the three sibilants, or surd spirants, 
this is the one of plainest and least questioned character: 



it is the ordinary European s -- a hiss expelled between 
the tongue and the roof of the mouth directly behind the 
upper front teeth. 

It is, then, dental, as it is classed by all the Hindu 
authorities. It is the one primitive Indo-European sibilant. 
Notwithstanding the great losses which it suffers in Sanskrit 
euphony, by conversion to the other sibilants, to r, to visarga, 
etc., it is still very high among the consonants in the order of 
frequency, or considerably more common than both the other 
two sibilants together. 

61. The Ef *. As to the character of this sibilant, also, 
there is no ground for real question : it is the one produced 
in the lingual position, or with the tip of the tongue re- 
verted into the dome of the palate. It is, then, a kind of 
s^-sound; and by European Sanskritists it is pronounced 
as an ordinary sh (French ch, German sch\ no attempt 
being made (any more than in the case of the other lingual 
sounds : 45) to give it its proper lingual quality. 

Its lingual character is shown by its whole euphonic influence, 
and it is described and classed as lingual by all the Hindu author- 
ities (the APr. adds, i. 23, that the tongue in its utterance 
is trough-shaped) . In its audible quality, it is a sA-sound 
rather than a s-sound; and, in the considerable variety of 
sibilant-utterance, even in the same community, it may coincide 
with some people's sh. Yet the general and normal sh is palatal 
(see below, 63) ; and therefore the sign s, marked in accordance 
with the other lingual letters, is the only unexceptionable trans- 
literation for the Hindu character. 

In modern pronunciation in India, s is much confounded with kh; and 
the MSS. are apt to exchange the characters. Later grammatical treatises, 
too, take note of the relationship (see Weber's Pratijfia, p. 84). 

62. This sibilant (as was noticed above, 46, and will be 
more particularly explained below, 180 fF.) is no original sound, 
but a product of the lingualization of s under certain euphonic 
conditions. The exceptions are extremely few (9 out of 145 
noted occurrences : 75), and of a purely sporadic character. The 
Rig- Veda has (apart from y sah, 182) only twelve words which 
show a s under other conditions. 

The final s of a root has in some cases attained a more independent 

65 SPIRANTS. 21 

value, and does not revert to s when the euphonic conditions are removed, 
but shows anomalous forms (225 . 

63. The 5T g> This sibilant is by all the native author- 
ities classed and described as palatal, nor is there any- 
thing in its history or its euphonic treatment to cast doubt 
on its character as such. It is, then, made with the flat 
of the tongue against the forward part of the palatal arch 
that is to say, it is the usual and normal sA-sound. By 
European scholars it is variously pronounced more often, 
perhaps, as s than as sh. 

The two s/i-sounds, s and p, are made in the same part of the mouth 
(the s probably rather further back), but with a different part of the tongue ; 
and they are doubtless not more unlike than, for example, the two t-sounds, 
written t and t ; and it would be not less proper to pronounce them both as 
one sh than to pronounce the linguals and dentals alike. To neglect the 
difference of s and f is much less to be approved. The very near relationship 
of s and f is attested by their euphonic treatment, which is to a considerable 
extent the same, and by their not infrequent confusion by the writers 
of manuscripts. 

64. As was mentioned above (41), the r, like c, comes 
from the corruption of an original &-sound, by loss of mute- 
contact as well as forward shift of the articulating point. In 
virtue of this derivation, it sometimes (though less often than c) 
"reverts" to k that is, the original k appears instead of it ; 
while, on the other hand, as a s/j-sound, it is to a certain 
extent convertible to s. In point of frequency, it slightly 
exceeds the latter. 

65. The remaining spirant, ^ h y is ordinarily pronounced 
like the usual European surd aspiration h. 

This is not, however, its true character. It is defined by all the native 
authorities as not a surd element, but a sonant (or else an utterance inter- 
mediate between the two) ; and its whole value in the euphony of the language 
is that of a sonant: but what is its precise value is very hard to say. The 
Paninean scheme ranks it as guttural, as it does also a : this means nothing. 
The Pratic.akhyas bring it into no relation with the guttural class : one of them 
quotes the opinion of some authorities that "it has the same position with 
the beginning of the following vowel" (TPr. ii. 47) which so far identi- 
fies it with our h. There is nothing in its euphonic influence to mark it 
as retaining any trace of gutturally articulated character. By some of the 
native phonetists it is identified with the aspiration of the sonant aspirates 


with the element by which, for example, gh differs from g. This view 

is supported by the derivation of h from the aspirates (next paragraph), by 
that of l + h from dh (54), and by the treatment of initial h after a final 
mute (163). 

66. The h, as already noticed, is not an original sound, 
but comes in nearly all cases from an older gTi (for the few 
instances of its derivation from dh and bh, see below, 223). It 
is a vastly more frequent sound than the unchanged gh (namely, 
as 7 to 1): more frequent, indeed, than any of the guttural 
mutes except k. It appears, like j (219), to include in itself 
two stages of corruption of gh: one corresponding with that of 
k to c, the other with that of k to c; see below, 223, for the 
roots belonging to the two classes respectively. Like the other 
sounds of guttural derivation, it sometimes exhibits "reversion" 
to its original. 

67. The : h. or visarga (visarjamya, as it is uniformly 
called by the Prati^akhyas and by Panini, probably as 'be- 
longing to the end' of a word), appears to be merely a surd 
breathing, a final A-sound (in the European sense of h), 
uttered in the articulating position of the preceding vowel. 

One Praticakhya (TPr. ii. 48) gives just this last description of it. It 
is by various authorities classed with ft, or with h and a: all of them are 
alike sounds in whose utterance the mouth-organs have no definite shaping 

68. The visarga is not original, but always only a substi- 
tute for final s or r, neither of which is allowed to maintain 
itself unchanged. It is a comparatively recent member of the 
alphabetic system; the other euphonic changes of final s and r 
have not passed through visarga as an intermediate stage. And 
the Hindu authorities are considerably discordant with one an- 
other as to how far h is a necessary substitute, and how far a 
permitted one, alternative with a sibilant, before a following 
initial surd. 

69. Before a surd guttural or labial, respectively, some of 
the native authorities permit, while others require, conversion of 
final s or r into the so-called jihvamutiya and upadhmariiya spi- 
rants. It may be fairly questioned, perhaps, whether these two 
sounds are not pure grammatical abstractions, devised (like the 
long /-vowel : 23) in order to round out the alphabet to greater 
symmetry. At any rate, neither printed texts nor manuscripts 
(except in the rarest and most sporadic cases) make any account 
of them. Whatever individual character they may have must be, 




it would seem, in the direction of the (German) ch and ^sounds. 
When written at all, they are wont to be transliterated by % 
and (p. 

70. The - anusvara, n or w 7 is a nasal sound lacking 
that closure of the organs which is required to make a 
nasal mute (36); in its utterance there is nasal resonance 
along with some degree of openness of the mouth. 

71. There is discordance of opinion both among the Hindu phonetists 
and their modern European successors respecting the real character of this 
element : hence a little detail is necessary here with regard to its occurrence 
and their views of it. 

Certain nasals in Sanskrit are of servile character, always to be assi- 
milated to a following consonant, of whatever character that may be. Such 
are final m in sentence-combination (213), the penultimate nasal of a root, 
and a nasal of increment (255) in general. If one of these nasals stands 
before a contact-letter or mute, it becomes a nasal mute corresponding to the 
latter that is, a nasal utterance in the same position of the mouth-organs 
which gives the succeeding mute. If, on the other hand, the following con- 
sonant does not involve a contact (being a semivowel or spirant), the nasal 
element is also without- contact : it is a nasal utterance with unclosed mouth- 
organs. The question is, now, whether this nasal utterance becomes merely 
a nasal infection of the preceding vowel, turning it into a nasal vowel (as 
in French on, en, un, etc., by reason of a similar loss of a nasal mute); or 
whether it is an element of more individual character, having place between 
the vowel and the consonant; or, once more, whether it is sometimes the one 
thing and sometimes the other. The opinions of the Praticakhyas and Panini 
are briefly as follows : 

The Atharva-Pratic.akhya holds that the result is everywhere a nasalized 
vowel, except when n or m is assimilated to a following I , in that case, the 
n or m becomes a nasal I: that is, the nasal utterance is made in the 
^-position, and has a perceptible i-character. 

The other Praticakhyas teach a similar conversion into a nasal counter- 
part to the semivowel, or nasal semivowel, before y and I and v (not before 
r also). In most of the other cases where the Atharva-Pratic.akhya acknow- 
ledges a nasal vowel namely, before r and the spirants the others 
teach the intervention after the vowel of a distinct nasal element, called the 
anusvara, 'after-tone'. 

Of the nature of this nasal afterpiece to the vowel no intelligibly clear 
account is given. It is said (RPr.) to be either vowel or consonant; it is 
declared (RPr., VPr.) to be made with the nose alone, or (TPr.) to be nasal 
like the nasal mutes; it is held by some (RPr.) to be the sonant tone of 
the nasal mutes ; in its formation , as in that of vowel and spirant, there is 
(RPr.) no contact. As to its quantity, see farther on. 

There are, however, certain cases and classes of cases where these other 


authorities also acknowledge a nasal vowel. So, especially, wherever a final 
n is treated (208; as if it were ns (its historically older form); and also in 
a small number of specified words. They also meiitiou the doctrine of nasal 
vowel instead of anusvara as held by some (and TPr. is uncertain and incon- 
sistent in its choice between the one and the other). 

In Panini, finally, the prevailing doctrine is that of anusvara every- 
where; and it is even allowed in many cases where the Pratic.akhyas pre- 
scribe only a nasal mute. But a nasal semivowel is also allowed instead be- 
fore a semivowel, and a nasal vowel is allowed in the cases (mentioned above) 
where some of the Pratic,akhyas require it by exception. 

It is evidently a fair question whether this discordance and uncertainty 
of the Hindu phonetists is owing to a real difference of utterance in differ- 
ent classes of cases and in different localities, or whether to a different scho- 
lastic analysis of what is really everywhere the same utterance. If anu- 
svara is a nasal element following the vowel, it cannot well* be any thing 
but either a prolongation of the same vowel-sound with nasality added, or a 
nasalised bit of neutral-vowel sound (in the latter case, however, the altering 
influence of an i or w-vowel on a following s ought to be prevented, which 
is not the case: see 183). 

72. The assimilated nasal element, whether viewed as 
nasalized vowel, nasal semivowel, or independent anusvara, has 
the value of something added, in making a heavy syllable, or 
length by position (79). 

The Praticakhyas (VPr., RPr.) give determinations of the quantity of 
the anusvara combining with a short and with a long vowel respectively to 
make a long syllable. 

73. Two different signs, 1 and -, are found in the MSS., 
indicating the nasal sound here treated of. Usually they are 
written above the syllable, and there they seem most naturally 
to imply a nasal affection of the vowel of the syllable, a nasal 
(anunasika) vowel. Hence some texts (Sama and Yajur Vedas), 
when they mean a real anusvara, bring one of the signs down 
into the ordinary consonant-place ; but the usage is not general. 
As between the two signs, some MSS. employ, or tend to employ, 
the - where a nasalized (anunasika) vowel is to be recognized, 
and elsewhere the 1; and this distinction is consistently observed 
in many European printed texts; and the former is called the 
anunasika sign: but it is very doubtful whether the two are not 
originally and properly equivalent. 

It is^a very common custom of the manuscripts to write 
the anusvara-sign for any nasal following the vowel of a syllable, 
either before another consonant or as final (not before a vowel), 
without any reference to whether it is to be pronounced as nasal 
mute, nasal semivowel, or anusvara. Some printed texts follow 
this slovenly and undesirable'habit ; but most write a nasal mute 








whenever it is to be pronounced excepting where it is an 
assimilated m (213). 

It is convenient also in transliteration to distinguish the 
assimilated m by a special sign, m, from the anusvara of more 
independent origin, n; and this method will be followed in the 
present work. 

74. This is the whole system of sounds recognised by the 
written character; for certain transitional sounds, more or less 
widely recognised in the theories of the Hindu phonetists, see 
below, 230. 

75. The whole spoken alphabet, then, may be arranged 
in the following manner, so as to show, so far as is possible 
in a single scheme, the relations and important classifications 
of its various members : 


111 asp. 


b unasp . 


ph asp. 

, a 

. fc 

19-78 8-1P ^ 
'9 J $ 


V # 

i. I 

T,f I u, u 

4-85 1.19 

74 -01 -01 2.61 .73 


r I v 


5-05 -69 4.99 

n n 

n n m 

22 -35 

1-03 4-81 4-34 





gh jh 

15 .01 

ff J 

.82 .94 

kh ch 

13 .17 

k c 

1-99 1.26 

Gutt. Pal. 





















p unasp. 




The figures set under the characters give the average 
percentage of frequency of each sound, found by counting the 
number of times which it occurred in an aggregate of 10,OOC 
sounds of continuous text, in 10 different passages, ^of 1,000 
sounds each, selected from different epochs of the literature : 
namely, two from the Rig- Veda, one from the Atharva-Veda, 
two from different Brahmanas, and one each from Manu, Bha- 
gavad-Gita, Qakuntala, Hitopade9a, and Vasavadatta*. 

III. Quantity of sounds and syllables. 

76. The Hindu grammarians take the pains to define 
the quantity of a consonant (without distinction among 

consonants of different classes) as half that of a short vowel. 

77. They also define the quantity of a long (dirgha) 
vowel or diphthong as twice that of a short vowel making 
no distinction in this respect between the guna and the 
prefab-diphthongs . 

78. Besides these two vowel-quantities, the Hindus 
acknowledge a third, called pluta (literally 'swimming'), 
or protracted, and having three moras, or three times the 
quantity of a short vowel. A protracted vowel is marked 
by a following figure 3: thus, 5TT$ a 3. 

The protracted vowels are practically of rare occurrence (in 
RV., three cases; in AV., fifteen; in the Brahman a literature, 
rather more frequent). They are used in cases of questioning, 
especially of a balancing between two alternatives, and also of 
calling to a distance or urgently. The protraction is of the last 
syllable in a w r ord, or in a whole phrase ; and the protracted 
syllable has usually the acute tone, in addition to any other 
accent the word may have ; sometimes it takes also anusvara, or 
is made nasal. 

Examples are: adhdh svid asi3d updri svid asl3t (RV.), 'was it, forsooth, 
below? was it, forsooth, above?' iddm bhUydS id$3m Hi (AV.), 'saying, is this 
more, or is that?' dgndSi pdtmvdSh s6mam piba (TS.), 'oh Agni! thou with 
thy spouse ! drink the soma'. 

A diphthong is protracted by prolongation of its first or a-element: 
thus, e to a3i, o to a3u. 

* See J. A. 0. S., vol. X. 

82] QUANTITY. 37 

The sign of protraction is also sometimes written as the result of ac- 
centual combination, when so-called kampa occurs: see below, 90b. 

79. For metrical purposes, syllables (not vowels) are 
distinguished by the grammarians as 'heavy' (guru) or 'light' 
(laghu). A syllable is heavy if its vowel is long, or short 
and followed by more than one consonant ("long by po- 
sition"). Anusvara and visarga count as full consonants in 
making a heavy syllable. The last syllable of a pada (pri- 
mary division of a verse) is reckoned as either heavy or 

The distinction in terms between the difference of long and short in 
vowel-sound and that of heavy and light in syllable-construction is valuable, 
and should be retained.. 

IV. Accent. 

80. The phenomena of accent are, by the Hindu gram- 
marians of all ages alike, described and treated as depend- 
ing on a variation of tone or pitch; of any difference of 
stress involved, they make no account. 

81. The primary tones (svara) or accent-pitches are 
two : a higher (udatta, 'raised'), or acute ; and a lower 
(anudatta, 'not raised'), or grave. A third (called svarita : 
a term of doubtful meaning), is always of secondary origin, 
being (when not enclitic : see below, 85) the result of actual 
combination of an acute vowel and a following grave vowel 
into one syllable. It is also uniformly defined as compound 
in pitch, a union of higher and lower tone within the 
limits of a single syllable. It is thus identical in physical 
character with the Greek and Latin circumflex, and fully 
entitled to be called by the same name. 

82. Strictly, therefore, there is but one distinction of tone 
in the Sanskrit accentual system : the accented syllable is raised 
in tone above the unaccented ; while then further, in certain 
cases of the fusion of an accented and an unaccented element 


into one syllable, that syllable retains the compounded tone of 
both elements. 

83. The svarita or circumflex is only rarely found on a 
pure long vowel or diphthong, but almost always on a syllable 
in which a vowel, short or long, is preceded by a y or v re- 
presenting an originally acute t or w-vowel. 

In transliteration, in this work, the udatta or acute will be 
marked with the ordinary sign of acute, and the svarita or cir- 
cumflex (as being a downward slide of the voice forward) with 
what is usually called the grave accent : thus, d, acute , ya or 
va, circumflex. 

84. The Praticakhyas distinguish and name separately the circumflexed 
tones arising by different processes of combination : thus, the circumflex is 

a. Ksaipra ('quick'), when an acute i or w-vowel (short or long) is con- 
verted into y or v before a dissimilar vowel of grave tone : thus, vyhpta 
from vi-apta, apsvantdr from apsu antdr. 

b. Jatya ('native') or nitya ('own'), when the same combination lies 
further back, in the make-up of a stem or form, and so is constant, or 
belongs to a word in all circumstances of its occurrence: thus, kva (from fcwa), 
svhr (stiar), nybk (nfak), budhnya (budhnfa), kanyh (fcanla), nadyas (nadf-as), 
tanvh (tanU-a). 

The words of both these classes are in the Veda, in the great majority 
of cases, to be read with restoration of the acute vowel as a separate syllable : 
thus, apsu antdr, suar, nadias, etc. . In some texts, part of them are 
written correspondingly : thus, suvar, tanuva, budhnfya. 

c. Praflista, when the acute and grave vowels are of such character that 
they are fused into a long vowel or diphthong (128): thus divi 'va (RV. 
and AV.), from dM iva; shdgata (TS.), from su-udyata; nai 'vh J friiyat 
(B.), from nd evd apniyat. 

d. Abhinihita, when an initial grave a is absorbed by a final acute e 
or 6 (135): thus, te 'bruvan, from te abruvan; t> 'bravit, from so abravlt. 

85. But further, the Hindu grammarians agree in de- 
claring the (naturally grave) syllable following an acute, 
whether in the same or in another word, to be svarita or 
circumflex - - unless, indeed, it be itself followed by an 
acute or circumflex; in which case it retains its grave tone. 
This is called by European scholars the enclitic or depend- 
ent circumflex. 

Thus, in tena and te ca, the syllable na and word ca are 
regarded and marked as circumflex : but in tena te and te ca 
svar they are grave. 


This seems to mean that the voice, which is borne up at the higher 
pitch to the end of the acute syllable, does not ordinarily drop to grave pitch 
by an instantaneous movement, but descends by a more or less perceptible 
slide in the course of the following syllable. No Hindu authority suggests 
the theory of a middle or intermediate tone for the enclitic, any more than 
for the independent circumflex. For the most part, the two are identified 
with one another, in treatment and designation. The enclitic circumflex is- 
likewise divided into a number of sub-varieties, with different names: they 
are of too little consequence to be worth reporting. 

86. The essential difference of the two kinds of circum- 
flex is shown clearly enough by these facts : a. the independent 
circumflex takes the place of the acute as the proper accent of 
a word, while the enclitic is the mere shadow following an acute,. 
and following it in another word precisely as in the same word ; 
b. the independent circumflex maintains its character in all 
situations, while the enclitic before a following circumflex or 
acute loses its circumflex character, and becomes grave ; more- 
over, c. in many of the systems of marking accent (below, 88), 
the two are quite differently indicated. 

87. The accentuation is marked in manuscripts only of the 
older literature : namely, in the different Vedic texts, in two of 
the Brahmanas (Taittirlya and Qatapatha), and in the Taittiriya- 
Aranyaka. There are a number of methods of writing accent, 
more or less different from one another ; the one found in MSS . of 
the Rig- Veda, which is most widely known, and of which most of 
the others are only slight modifications, is as follows : the acute 
syllable is left unmarked ; the circumflex, whether independent 
or enclitic, has a brief perpendicular stroke above; and the grave 
next preceding an acute or (independent) circumflex has a brief 
horizontal stroke below. Thus 

stcfjtri juhdti; rp^T tanvh; WT kva. 

The introductory grave stroke below, however, cannot be given if an acute 
syllable is initial, whence an unmarked syllable at the beginning of a word 
is to be understood as acute ; and hence also, if several grave syllables precede 
an acute at the beginning of a sentence, they must all alike have the grave 
sign. Thus, 

^: mdrah; ft te ; cflf^fH karisyasi ; rT%TfTT tuvijatd. 

All the grave syllables, however, which follow a marked circumflex are left 
unmarked, until the occurrence of another accented syllable causes the one 
which precedes it to take the preparatory stroke below. Thus, 

sudfcikasamdrk ; 


but H^lfctiH^J JNIH sudfftkasamdrg gdvam 


88. The other methods it is not worth while to attempt to set forth. 
They may be found illustrated in the different texts, and explained by the 
editors of them. In part, their peculiarities consist in other forms or places 
given to the grave and circumflex signs. ' In some methods, the acute is itself 
marked, by a slight stroke above. In several, thte independent circumflex is 
distinguished from the enclitic. The most peculiar systems are the scanty 
and imperfect one of the Qatapatha-Brahmana, with a single sign, written 
below ; and the highly intricate one of the Sama-Veda, with a dozen different 
signs, written above. 

89. In this work, as everything given in the devanagari char- 
acters is also given in transliteration, it will in general be 
unnecessary to mark the accent except in the transliterated form ; 
where, however, the case is otherwise, there will be adopted the 
method* of marking only the accented syllables, the acute 
and the independent circumflex : the latter by the usual svarita- 
sign, the former by a small u (for udatta] above the syllable : 

*^3\ indra, Mi4 dgne, 

These being given, everything else which the Hindu theory recognises 
as dependent on and accompanying them can readily be understood as im- 

90. The theory of the Sanskrit accent, as here given (a consistent and 
intelligible body of phenomena), has been overlaid by the Hindu theorists, 
especially of the Praticakhyas, with a number of added features, of a much 
more questionable character. Thus : 

a. The unmarked grave syllables, following a circumflex (either at the 
end of a sentence, or till the near approach of another acute), are declared 
to have the same high tone with the (also unmarked) acute. They are called 
pracaya or pracita ('accumulated': because liable to occur in an indefinite 
series of successive syllables). 

b. The circumflex, whether independent or enclitic, is declared to begin 
on a higher pitch than acute, and to descend to acute pitch in ordinary cases : 
the' concluding instant of it being brought down to grave pitch, however, in 
the case of an independent circumflex which is immediately followed by 
another ascent of the voice to higher pitch (in acute or independent cir- 

This last case, of an independent circumflex followed by acute or cir- 
cumflex, receives peculiar written treatment. In the Rig- Veda method, a 
fignre 1 or 3 is set after the circumflexed vowel, according as it is short or 
long, and the signs of accent are thus applied: 

* Introduced by Bohtlingk, and used in the Petersburg lexicon and elsewhere. 

93] ACCENT. 31 

: apsv alntdh from apsu antdh; 

raybS 'vdnih from rayo avdnih . 
The other methods, more or less akin with this, need not be given. 

In the scholastic utterance of such a syllable is made a peculiar quaver 
or roulade of the voice, which is called kampa or vikampana. 

C. Panini gives the ambiguous name of eka$ruti ('monotone') to the pra- 
cita syllables, and says nothing of the uplifting of the circumflex to a higher 
plane: he teaches, however, a depression below the grave pitch for the mark- 
ed grave syllable before acute or circumflex, calling it sannatara (otherwise 

91. The system of accentuation as marked in the Vedic texts has assum- 
ed in the traditional recitation of the Brahmanic schools a peculiar and 
artificial form, in which the designated syllables, grave and circumflex 
(equally, the enclitic and the independent circumflex), have acquired a con- 
spicuous value, while the undesignated, the acute, has sunk into insigni- 
ficance *. 

92. The Sanskrit accent taught in the native grammars and 
represented by the accentuated texts is essentially a system of 
word-accent only. No general attempt is made (any more than 
in the Greek system) to define or mark a sentence-accent, the 
effect of the emphasis and modulation of the sentence in mo- 
difying the independent accent of individual words. The only 
approach to it is seen in the treatment of vocatives and personal 

A vocative is usually without accent except at the beginning 
of a sentence : for further details, see the chapter on Declension. 

A personal verb-form is usually accentless in an independ- 
ent clause, except when standing at the beginning of the clause : 
for further details, see the chapter on Conjugation. 

93. Certain other words also are, usually or always, without 

a. The particles ca, vd, u, sma, iva, cid, svid, fta, are always without 

b. The same is true of certain pronouns and pronominal stems : md, me, 
ndu, na<, tvd, te, i?am, vas, ena-, tva-. 

c. The cases of the pronominal stem a are sometimes accented and some- 
times accentless. 

An accentless word is not allowed to stand at the begin- 
ning of a sentence : also not of a pada or primary division of 
a verse ; a pada is, in all matters relating to accentuation, treat- 
ed like an independent sentence. 

* Hang, Wedischer Accent, in Abh. d. Bayr. Akad., vol. XIII, 1874. 


94. Some words have more than a single accented syllable. 
Such are : 

a. Dual collective compounds : as fndravdrunau. 

b. A few other compounds, in which each member irregularly retains 
its own accent : as tdnundpat, vdnaspdti, brhaspdti. In a rare case or two, 
also their further compounds, as brhaspdtipramitta. 

C. Infinitive datives in tavdf: as etavaf, 

d. A word naturally barytone, but having its final syllable protracted : 
see above, 78. 

e. The particle vdvd (in the Brahmanas). 

95. On the place of the accented syllable in a Sans- 
krit word there is no restriction whatever depending upon 
either the number or the quantity of the preceding or 
following syllables. The accent rests where the rules of 
inflection or derivation or composition place it, without 
regard to any thing else. 

Thus, indre, agnau, indrena, agnina, agriintim, bahucyuta, 
dnapacyuta, parjdnyajinvita, abhimatisahd } dnabhimlatavarna, abhicas- 
ticatana, hiranyavacimattama . 

96. Since the accent is marked only in the older litera- 
ture, and the statements of the grammarians, with the 
deduced rules of accentuation, are far from being sufficient 
to settle all cases, the place of the stress of voice for a 
considerable part of the vocabulary is undetermined. Hence 
it is a general habit with European scholars to pronounce 
Sanskrit words according to the rules of the Latin accent. 

97. In this work, the accent of each word and form will 
in general be marked, so far as there is authority determining its 
place and character. Where specific words and forms are quoted, 
they will only be so far accentuated as they are found with 
accent in accentuated texts. 

103] 33 



98. THE individual elements composing a language as 
actually used are its words. These are in part uninflected 
vocables (indeclinables, particles) ; in the main, they are in- 
flected forms. 

99. The inflected forms are analy sable into inflective en- 
dings, of declension or of conjugation, and inflected stems to 
which those endings are added. 

100. The inflected stems, again, are for the most part 
as are also in part the uninflected words analysable into 
derivative endings or suffixes, and roots, to which, either directly 
or through more primary stems, those endings are added. 

But, not a few stems and particles are irreducible to roots ; and, on the 
other hand, roots are often used directly as inflected stems, in declension as 
well as in conjugation. 

101. The roots are, in the condition of the language as 
it lies before us, the ultimate attainable elements ; to a great 
extent not actually ultimate, but, where otherwise, the result of 
processes of development too irregular and obscure to be made 
the subject of treatment in a grammar. 

102. The formative processes by which both inflectional 
forms and derivative stems are made, by the addition of endings 
to bases and to roots, are more regular and transparent in San- 
skrit than in any other Indo-European language, and the gram- 
matical analysis of words into their component elements is 
correspondingly complete. Hence it became the method of the 
native grammarians, and has continued to be that of their Euro- 
pean successors, to teach the language by presenting the endings 
and stems and roots in their analysed forms, and laying down 
the ways in which these are to be combined together to make 
words. And hence a statement of the euphonic rules which 
govern the combination of elements occupies in Sanskrit grammar 
a more prominent and important place than in other grammars. 

103. Moreover, the formation of compound words, by the 
putting together of two or more stems, is a process of very 
exceptional frequency in Sanskrit ; and this kind of combination 
also has its own euphonic rules. And once more, in the form 

Whitney, Grammar. 3 


in which the language is handed down to us by the litera- 
ture, the words composing a sentence or paragraph are adapted 
to and combined with each other by nearly the same rules which 
govern the making of compounds, so that it is impossible to 
take apart and understand the simplest sentence in Sanskrit 
without understanding those rules. Hence also a greatly added 
degree of practical importance belonging to the subject of 
euphonic combination. 

This euphonic interdependence of the words of a sentence, which is 
unknown to any other language in anything like the same degree, is shown 
to be at least in considerable measure artificial, implying an erection into 
necessary and invariable rules of what in the living language were only 
optional practices, by the evidence of the older dialect of the Vedas and the 
younger Prakritic dialects, in both of which these rules (especially as regards 
hiatus: 113) are very often violated. 

104. We have, therefore, in the first place to consider the 
euphonic principles and laws which govern the combination of 
the elements of words (and the elements of the sentence) ; and 
then afterward to take up the subject of inflection, under the 
two heads of declension and conjugation : to which will succeed 
some account of the classes of uninnected words. 

105. The formation of conjugatioixal stems (tense and 
mode-stems, etc.) will be taught, as is usual, in connection 
with the processes of conjugational inflection ; that of uninflected 
words, in connection with the various classes of those words. 
But the general subject of derivation, or the formation of de- 
clinable stems, will be taken up by itself later for a brief pre- 
sentation ; and it will be followed by an account of the formation 
of compound stems. 

Although, namely, the general plan of this series of grammars excludes 
the subject of derivation, yet, because of the comparative simplicity and 
regularity of the principal processes of derivation in Sanskrit, and the import- 
ance to the student of accustoming himself from the beginning to trace those 
processes, in connection with the analysis of derived forms, back to the root, 
an exception will be made in regard to the subject in the present work. 

106. We assume, then, for the purposes of the present 
chapter, the existence of the material of the language in a 
grammatically analysed condition, in the form of roots, stems, 
and endings. 

107. What is to be taken as the proper form of a root or 
stem is not in all cases clear. Very many of both classes show 
m a part of their derivatives a stronger and in a part a weaker 
form (260). This is, in most cases, the only difficulty affecting 


stems whether, for example, we shall speak of derivatives in 
mat or in mant, of comparatives in yas or in yam, of a perfect 
participle in vat or in vaiis or in us. The Hindu grammarians 
usually give the weaker form as the normal one, and derive the 
other from it by a strengthening change ; some European author- 
ities adopt the one form and some the other : the question is an 
unessential one, giving rise to no practical difficulty. 

108. As regards the roots, the difficulty is greater, partly 
because complicated with other questions, arising from practices 
of the Hindu grammarians, which have been more or less widely 
followed by their European successors. Thus : 

a. More than half of the whole number of roots given by the Hindu 
authorities (which are over 2000) have never been found actually used in 
the literature; and although some of these may yet come to light, or may 
have existed without finding their way into any of the preserved literary 
documents, it is certain that most are fictitious, made in part for the ex- 
planation of words claimed to be their derivatives, and in part for other and 
perhaps unexplainable reasons. Of the roots unauthenticated by traceable 
use no account will be made in this grammar or, if at all considered, 
they will be carefully distinguished from the authenticated. 

b. Those roots of which the initial n and s are regularly converted to 
n and * after certain prefixes are by the Hindu grammarians given as be- 
ginning with n and s: no European authority follows this example. 

c. A number of roots ending in a which is irregularly treated in the 
inflection of the present-system are written in the Hindu lists with diph- 
thongs e or at or o; and so, after this example, by many Western scholars. 
Here they will be regarded as a-roots : compare below, 251. The o of such 
roots, especially, is purely arbitrary ; no forms made from the root justify it. 

d. The roots showing interchangeably r, ar, and ir and Ir or ur and ur 
forms are written by the Hindus with r, or with f, or with both. Here also 
the f is arbitrary and indefensible. As between r and ar, even the latest 
European authorities are at variance, and it may be left to further research to 
settle whether the one or the other is alone worthy to be accepted. Here (mainly 
as a matter of convenience : compare below, 237) the r-forms will be used. 

e. In the other cases of roots showing a stronger and a weaker form, 
choice is in great measure a matter of minor consequence unless further 
research and the settlement of pending phonetic questions shall show that 
the one or the other is decidedly the truer and more original. From the 
point of view of the Sanskrit alone, the question is often impossible to 

f. The Hindus classify as simple roots a number of derived stems : 
reduplicated ones, as didhi, jagr, daridra ; present-stems, as urnu ; and 
denominative stems, as avadhir, kumar, sabhaj, mantr, santv, arth, and the 
like. These are in European works generally reduced to their true value. 

g. But it is impossible to draw any definite line between these cases 


and others in which root-forms evidently of secondary origin have attained a 
degree of independent value in the language which almost or quite entitles 
them to rank as individual roots. Even the weak and strong forms of the 
same root as vad and vand, cit and cint, mah and mahh may have 
such a difference of use that they count as two ; or a difference of inflection 
combined with a difference of meaning in a root has the same effect as 
in vr vrnoti and vr vrnlte, in ha jahati and ha jihite; or an evident present- 
stem becomes a separate root as jinv and pinv. Not a few roots occur 
in more or less clearly related groups, the members of which are of various 
degrees of independence. Thus, a considerable class of roots show an added a; 
and such as mna and dhma are reckoned only as side-forms of man and 
dham; while Jra, pra, pya, psa, and others, presumably made in the same 
manner, figure as separate from their probable originals. Many final con- 
sonants of roots have the value of "root-determinatives", or elements of 
obscure or unknown origin added to simpler forms. A class of derivative 
roots show signs of reduplication, as caks, jaks, dudh; or of a desiderative 
development, as bhaks and bhiks, ?rus, afes, naks. Yet another class seem 
)o contain a preposition fused with a root, as vyac, ap } and the later ujh 
and vyas. 

With most of such cases it is not the part of a Sanskrit grammar, but 
rather of a general Indo-European comparative grammar, to deal according 
to their historical character. We must be content to accept as roots what 
elements seem to have on the whole that value in the existing condition of 
the language; with due recognition of authorized differences of opinion on 
many points, as well as of the fact that further knowledge will set many 
things now doubtful in a clearer light. 

Principles of Euphonic Combination. 

109. The rules of combination are in some respects 
different, according as they apply 

a. to the internal make-up of a word, by the addition 
of derivative and inflectional endings to roots and stems; 

b. to the more external putting together of stems to 
make compound stems, and the yet looser and more acci- 
dental collocation of words in the sentence. 

Hence they are usually divided into rules of internal 
combination (or sandhi, -putting together'), and rules of 
external combination. 

110. In both classes of cases, however, the general princi- 
ples of combination are the same and likewise, to a great 
extent, the specific rules. The differences depend in part on 


the occurrence or non-occurrence of certain combinations in the 
one class or the other ; in part, on the difference of treatment of 
the same sound as final of a root or of an ending, the former 
being much more persistent than the latter ; in part, on the 
occurrence in external combination of certain changes which 
are apparently phonetic but really historical ; and, most frequent 
and conspicuous of all, on the fact that (157) vowels and se- 
mivowels and nasals exercise a sonantizing influence in external 
combination, but not in internal. Hence, to avoid unnecessary 
repetition as well as the separation of what really belongs to- 
gether, the rules for both kinds of combination will be given below 
in connection with one another. 

111. Moreover, before case-endings beginning with bh and 
* (namely, bhis, bhyas, bhyam, su}, the treatment of the finals 
of stems is in general the same as in the combinations of words 
(pada) with one another whence those endings are sometimes 
called jt?ae?a-endings, and the cases they form are known as pada- 
cases. And with some of the suffixes of derivation the same is 
the case. 

The importance of this distinction is somewhat exaggerated by the 
ordinary statement of it. In fact, dh is the only sonant mute initial of an 
ending occurring in conjugation, as bh in declension ; and the difference of 
their treatment is in part owing to the one coming into collision usually 
with the final of a root and the other of an ending, and in part to the fact 
that dh, as a dental, is more assimilable to palatals and linguals than bh. 
A more marked and problematic distinction is made between su and the 
verbal endings si, sva, etc., especially after palatal sounds and s. 

112. The leading rules of internal combination are those which are of 
highest and most immediate importance to a beginner in the language, since 
his first task is to master the principal paradigms of inflection; the rules of 
external combination may better be left untouched until he comes to dealing 
with words in sentences, or to translating. Then, however, they are indis- 
pensable, since the proper form of the words that compose the sentence is 
not to be determined without them. 

The general principles of combination underlying the 
euphonic rules, and determining their classification, may be stated 
as follows : 

113. Hiatus. A hiatus is avoided. 

There are but two or three words in the language which 
in their accepted written form exhibit successive vowels forming 
different syllables : they are titau, 'sieve' (perhaps for titasu, 
BE,.) and prdilga, 'wagon-pole' (for pray uga ?} ; and, in RV., 
suuti. For the not infrequent instances of composition and 
sentence combination where the recent loss of s or y or v 
leaves a permanent hiatus, see below, 132ff., 175 b, 177. 


It has been already noticed that in the Veda, as the metrical form of 
the hymns plainly shows, there is no avoidance of hiatus, either as between 
the stem-finals and endings of words, between the members of compounds, 
or between the words composing a sentence. In cases innumerable, a y 
and v (especially after two consonants, or a long vowel and consonant) are 
to be read as i and u. But also a long vowel is sometimes to be resolved into 
two syllables oftenest, a into a-a : this resolution is sometimes historical, 
but ordinarily purely metrical. For details, see below. It is with regard 
to the hiatus that the rules of the grammatically regulated classical Sanskrit 
are most demonstrably and conspicuously different from the more living usages 
of the sacred dialect. 

114. Deaspiration. An aspirate mute is liable to 
lose its aspiration, being allowed to stand unchanged only 
before a vowel or semivowel or nasal. 

115. Assimilation. The great body of euphonic 
changes in Sanskrit, as elsewhere, falls under the general 
head of assimilation which takes place both between 
sounds which are so nearly alike that the difference between 
them is too insignificant to be worth preserving, and between 
those which are so diverse as to be practically incompat- 

116. In part, assimilation involves the conversion of 
one sound to another of the same series, without change of 
articulating position ; in part, it involves a change of position, 
or transfer to another series. 

117. Of changes within the series, the most frequent and 
important occur in the adaptation of surd and sonant sounds to 
one another ; but the nasals and I have also in certain cases 
their special assimilative influence. Thus: 

a. In the two classes of non-nasal mutes and spirants, surd and sonant 
are wholly incompatible ; no surd of either class can either precede or follow 
a sonant of either. 

A mute, surd or sonant, is assimilated by being changed to its corres- 
pondent of the other kind ; of the spirants, the surd s is the only one having 
a sonant correspondent, namely r, to which it is convertible in external 

b. The nasals are more freely combinable: a nasal may either precede 
or follow a mute of either kind, or the sonant spirant h; it may also follow 
a surd spirant (sibilant); no nasal, however, ever precedes a sibilant in the 


interior of a word (it is changed instead to anusvara); and in external com- 
bination their concurrence is usually avoided by insertion of a surd mute. 

c. A semivowel has still less sonantizing influence ; and a vowel least 
of all * : both are freely preceded and followed by sounds of every other 
class, in the interior of a word. 

Before a sibilant, however, is found, of the semivowels, only r and very 
rarely I. Moreover, in external combination, r is often changed to its surd 
corrspondent 5. 


d. In composition and sentence-collocation, initial vowels and semivowels 
and nasals also require the preceding final to be sonant. And 

e. Before a nasal and i, the assimilative process is sometimes carried 
further, by the conversion of a final mute to a nasal or I respectively. 

118. Of conversions involving a change of articulate posi- 
tion, the most important are those of dental sounds to lingual, 
and, less often, to palatal. Thus : 

a. The dental s and n are "very frequently converted to and n by the 
assimilating influence of contiguous or neighboring lingual sounds : the , 
even by sounds namely, i and w-vowels and k which have themselves 
no lingual character. 

b. A non-nasal dental mute is (with a few exceptions in external 
combination) made lingual when it comes into collision with a lingual sound. 

c. The dental mutes and sibilant are made palatal by a contiguous palatal. 
But also: 

d. A m (not radical) is assimilated to a following consonant, of what- 
ever kind. 

119. The euphonic combinations of the palatal mutes, the 
palatal sibilant, and the aspiration, as being sounds derived by 
phonetic alteration from more original gutturals (42 ff.), are 
made peculiar and complicated by two circumstances : their rever- 
sion to a guttural form (or the appearance of the unaltered 
guttural instead of them^ ; and the different treatment of/ and h 
according as they represent one or another degree of alteration 
the one tending, like c, more to the guttural reversion, the 
other showing, like c, a more sibilant and lingual character. 

120. The lingual sibilant s, also of derivative character 
(from dental s} : shows as radical final a peculiar and problematic 
mode of combination. 

121. Extension and abbreviation of conso- 
nant-groups. The native grammarians allow or require 
certain extensions, by duplication or insertion, of groups of 
consonants. And, on the other hand, abbreviation of cer- 

* In conformity with general phonetic law: see Sievers, Lautphysiologie, p. 140. 


tain other groups is allowed, and found often practised in 
the manuscripts. 

122. Permitted Finals. The permitted occurrence 
of consonants at the end of a word is quite narrowly re- 
stricted. In general, only one consonant is allowed after 
the last vowel; and that must be neither the aspiration, 
nor a sibilant, nor a semivowel (save rarely 5J^, nor an 
aspirate mute, nor a sonant mute if not nasal, nor a palatal. 

123. Increment and Decrement. Besides these 
more or less regular changes accompanying the combination 
of the parts that make up words, there is another class of 
a somewhat different character, not consisting in the mutual 
adaptations of the parts, but in strengthening or weakening 
changes of the parts themselves. 

124. It is impossible to carry through a perfectly systematic 
arrangement of the detailed rules of euphonic combination, be- 
cause the different varieties of euphonic change more or less 
overlap and intersect one another. The order followed below 
will be as follows : 

1. Rules of vowel combination, for the avoidance of hiatus. 

2. Rules as to permitted finals (since these underlie the 
farther treatment of final consonants in external combination). 

3. Rules for loss of aspiration of an aspirate mute. 

4. Rules of surd and sonant assimilation, including those 
for final s and r. 

5. Rules for the conversion of dental sounds to lingual 
and palatal. 

6. Rules for the changes of final nasals, including those in 
which a former final following the nasal reappears in combi- 

7. Rules regarding the special changes of the derivative 
sounds the palatal mutes and sibilant, the aspiration, and 
the lingual sibilant. 

8. Rules as to extension and abbreviation of consonant 

9. Rules for strengthening and weakening processes. 
Everywhere, rules for more sporadic and less classifiable cases 

will be given in the most practically convenient connection ; and 
the Index will render what help is needed toward finding them. 


Rules of Vowel Combination. 

125. The concurrence of two vowels, or of vowel and 
diphthong, without intervening consonant, is forbidden by 
the euphony of the later or classical language. It is avoided, 
according to the circumstances of the case, either by fusion 
of the two concurrent sounds into one. by the reduction of 
one of them to a semivowel, or by development of a semi- 
vowel between them. 

The texts of the older or Vedic dialect are written according to the 
euphonic rules of the later, although in them (as was pointed out above, 113) 
the hiatus is really of very frequent occurrence. Hence they are not to be 
read as written, but with almost constant reversal of the processes of vowel 
combination which they have artificially undergone. 

The rules of vowel combination, as regards both the result- 
ing sound and its accent, are nearly the same in internal and 
in external sandhi. 

126. Two similar simple vowels, short or long, coalesce 
and form the corresponding long vowel: thus, two a-yowels 
(either or both of them short or long) form 5TT a ; two ^-vow- 
els, ^ l\ two w-vowels. T37w; and, theoretically, two r-vow- 
els form ^ f , but it is questionable whether the case ever 
practically occurs. Examples are: 

: sa ca 'prajah (ca -f- aprajah); 
ail \a (ati -\-iva) ; 
suktam (su-uktam); 

cga "sit (raja -\-asit) ; 
i adhiqvarah (adhi-~i$varah). 

As the above examples indicate, it will be the practice everywhere in 
this work, in transliteration (but not in the devanagari text), to separate 
independent words ; and if an initial vowel of a following word has coalesced 
with a final of the preceding, this will be indicated by an apostrophe 
single if the initial vowel be the shorter, double if it be the longer, of the 
two different initials which in every case of combination yield the same result. 

127. An a-vowel combines with a following *- vowel to 
e / with an w-vowel, to R o ; with ^J r, to STT ar ; with 


^T theoretically', to 3^al; with ^ e or ^ ai, to ^ a*/ with 
5JT or 5t au, to 3T #M. Examples are: 
I?RT rajendra (raja-indra); 

\\ hitopadegah (hita-upadec,ah).; 
maharsih (maha - rsih); 
va ( sa } cvaj; 

rajaigvaryam (raja - aigvaryam); 
divaukasah (diva - okasah); 
jvarausadham (jvara - ausadham) . 
In some of the Vedic texts, the vowel r is written unchanged after the 
a- vowel, which, if long, is shortened: thus, maharsih instead of maharsiti. 
The two vowels, however, are usually pronounced as one syllable. 

When successive words like indra a ihi are to be combined, the first 
combination, to indra, is made first, and the result is indre " '/u (not indrai " 
'hi, from indra e 'hi). 

128. As regards the accent of these vowel combinations, it is 
to be noticed that, a. as a matter of course, the union of acute 
with acute yields acute, and that of grave with grave yields grave : 
that of circumflex with circumflex cannot occur ; b. a circumflex 
with following acute yields acute, the final grave element of the 
former being raised to acute pitch ; a grave with following acute 
does the same, as no upward slide of the voice on a syllable is 
acknowledged in the language ; but, c. when the former of the 
fused elements is acute and the latter grave, we might expect 
the resulting syllable to be in general circumflex, to represent 
both the original tones. Panini in fact allows this accent in 
every such case; and in a single accentuated Brahmana text (QB.), 
the circumflex is regularly written. But the language shows, on 
the whole, an indisposition to allow the circumflex to rest on 
either long vowel or diphthong as its sole basis, and the acute 
element is suffered to raise the other to its own level of pitch, 
making the whole syllable acute. The only exception to this, 
in most of the texts, is the combination of i and t, which be- 
comes I: thus, div\ 'va, from divi iva; in the Taittiriya texts 
alone such a case follows the general rule, while u and u, in- 
stead, make ^: thus, shdgata from sii-udgata. 

> 129. The e-vowels, the w-vowels, and ft r, before a 
dissimilar vowel or a diphthong, are regularly converted 
each into its own corresponding semivowel, ET y or cf v or 
!f r. Examples are : 


ity aha (iti -\- aha) ; 
madhv iva (madku -f- wa); 

duhitrarthe (duhitr-arthe); 
stry asya (strl-\-asya); 
^tf vadhvdi (vadhu-ai). 

But in internal combination (never in external) the i and 
w-vowels are not seldom changed instead to iy and uv and 
this especially in monosyllables, or after two consonants, where 
otherwise a group of consonants difficult of pronunciation would 
be the result. The cases will be noticed below, in explaining 
inflected forms. 

A radical a-vowel is converted into y even before i in per- 
fect tense-inflection : so ninyima (nirit -|- ima) . 

130. As regards the accent here, as in the preceding 
case (128), the only combination requiring notice is that of an 
acute i or w-vowel with a following grave : the result is circum- 
flex ; and such cases of circumflex are many times more frequent 
than any and all others. Examples are : 

vyusti (vt-usti); ER^ffcT abhyarcati ; 
fr nadyau (nadi-au); 

svuta (su-istaj; H*W tanvas (tanti-as). 
Of a similar combination of acute r with following grave, only a single 
case has been noted in accented texts : namely, vijndtr etdt (B. xiv. 6. 8 n ) : 
the accentuation is in accordance with the rules for i and u. 

131. Of a diphthong, the final i or w-element is chang- 
ed to its corresponding semivowel. ^ y or of 0, before any 
vowel or diphthong : thus, ^ e (really ai : 28) becomes %m ay, 
and EJT o (that is, au) becomes 5(Sf av ; ^ ai becomes SHIT ay, 

and f au becomes 35R av. 


No change of accent, of course, occurs here; each original 
syllable retains its syllabic identity, and hence also its own tone. 
Examples can be given only for internal combination, since in external 
combination there are further changes: see the next paragraph. Thus, 

^ naya (ne-a); ^TFT naya (nai-a); 

H^ bhava (bko-a); >TR bhava (bhau-a). 
^ 132. In external combination, we have the important 
additional rule that the semivowel resulting from the con- 
version of the final element of a diphthong is in general 


dropped; and the resulting hiatus is left without further 

133. That is to say. a final ^ e (by far the most fre- 
quent case) becomes simply 3f a before an initial vowel 
(except % a: see 135, below), and both then remain un- 
changed; and a final ^ ai, in like manner, becomes (every- 
where) a. Thus, 

t ta agatah (te -\- agatah); 
nagara iha (nagare -f- 7^/ 

adadat (tasmai -f- adadat); 
uktam (striyai-{~uktam). 

The later grammarians allow the i/ in such combinations to be either 
retained or dropped; but the uniform practice of the MSS., of every age, 
in accordance with the strict requirement of the Vedic grammars (Pratigakhyas), 
is to omit the semivowel and leave the hiatus. 

The persistence of the hiatus caused by this omission is a plain indi- 
cation of the comparatively recent loss of the intervening consonantal sound. 
Instances of the combination of the remaining final and initial are not unknown, 
but they are of sporadically rare occurrence. 

/ 134. The diphthong o (except as phonetic alteration of 
final as: see 175 a) is an unusual final, appearing only in the 
stem go (356), in the voc. sing, of w-stems, in words of which 
the final a is combined with the particle u, as atho, and in a 
few interjections. In the last two classes it is uncombinable 
(below, 138) ; the vocatives sometimes retain the v and sometimes 
lose it (the practices of different texts are too different to be 
briefly stated); go (in composition only) does not lose its final 
element, but remains gav or go. A final as becomes a, with 
following hiatus, before any vowel save a (for which, see the 
next paragraph). 

The of v of STR av from aft au is usually retained : thus, 
fiic<ef tav eva (tau -\-eva) ; 

3H"i(e(-*ii4l ubhav indragm (ubhau -\- indragm) . 
In some texts, however, it is lost before an w-vowel, the a alone 
remaining, with hiatus; in at least one text (Kathaka), it is dropped before 
every vowel. The later grammarians allow it to be either retained or dropped. 

135. After final ^ e or ^ o. an initial ?J a disappears. 

The resulting accent is as if the a were not dropped, but 
rather absorbed into the preceding diphthong, having its tone 


duly represented in the combination. If, namely, the e or o is 
grave or circumflex and the a acute, the former becomes acute ; 
if the e or o is acute and the a grave, the former becomes cir- 
cumflex, as usually in the fusion of an acute and a grave ele- 
ment. If both are acute or both grave, no change, of course, 
is seen in the result. Examples are : 

te 'bruvan (te abruvan); 
so 'bravit (sdh abravit); 

t hihsitavyd 'gnih (hinsitavydh agnih); 
y aa> indro 'bravit (ydd indrah dbramt); 
yddrajanyo 'bramt (yddrajanydh dbravit). 

As to the use of the avagraha sign in the case of such an elision, see 
above, 16. In transliteration, the reversed apostrophe, or rough breathing, 
will be used in this \vork to represent it. 

This elision or absorption of initial a after final e or o, which in the 
later language is the invariable rule, is in the Veda only an occasional 
occurrence; and there is no close accordance with regard to it between the 
written and the spoken form of the Vedic texts. In the Atharvan, for ex- 
ample, the a is omitted in writing in about one third of the cases, but is 
to be omitted in reading in less than one fifth (including a number in which 
the written text preserves it). See APr. iii. 54, note. 

To the rules of vowel combination, as above stated, there 
are certain exceptions. Some of the more isolated of these will 
be noticed where they come up in the processes of inflection 
etc. ; a few require mention here. 

136. In internal combination : 

a. The augment a makes with the initial vowel of a root 
the combinations at, au, ar (vrddhi-vowels), instead of e, o, ar 
(ywraz-vowels), as required by 127. 

b. The final o of a strengthened stem (238 b) becomes av before the 
suffix ya (originally za); 

C. The final vowel of a stem is often dropped when a secondary suffix 
is added. 

For the weakening and loss of radical vowels, and for certain insertions, 
see below, 249 ff., 2578. 

137. In external combination : 

a. The final a or o of a preposition, with initial r of a 
root, makes ar instead of ar. 

b. The final a of a preposition before roots beginning with e or o is 
usually omitted. 

C. A final a in composition may be cut off before otu and ostka. 

d. The form uh from yvah sometimes makes the heavier (vrddhi) 


diphthongal combination with a preceding a ; thus, prauha. praudhn, akmuhini 
(from pra-uha etc.). 

138. Certain final vowels, moreover, are uncombinable 
(pragrhya) , or maintain themselves unchanged before any 
following vowel. Thus, 

a. The vowels /, u and e as dual endings, both of declen- 
sional and of conjugational forms. Thus, landliu asate imau ; 
giri arohatam. 

b. The pronoun ami, (nom. pi.); and the Vedic pronom- 
inal forms asme, yitsme, tve. 

c. A final o made by combination of a final o-vowel with the particle u : 
thus, atho, mo, no. 

d. A final I or u of a Vedic locative case. 

e. A protracted final vowel (78). 

f. The final, or only, vowel of an interjection, as aho, he, a, i, u. 

Permitted Finals. 

139. The sounds allowed to occur as finals in Sanskrit 
words, standing by themselves (not in euphonic combination 
with something following), are closely limited, and those 
which would etymologically come to occupy such a position 
are often variously altered, in general accordance with their 
treatment in other circumstances, or are sometimes omitted 

The variety of consonants that would ever come at the end of either 
an inflected form or a derivative stem in the language is very small : namely, 
in forms, only t (or dj, n, m, s; in derivative stems, only t, d, n, r, s (and, 
in a few rare words, j). But almost all consonants occur as finals of roots; 
and every root is liable to be found, alone or as last member of a compound, 
in the character of a declined stem. 

140. All the vowel sounds, both simple and diphthongal, 
may be sounded at the end of a word. 

But neither f nor I ever actually occurs; and r is very rare (only as 
neuter sing, of a stem in r or or, or as final of such a stem in composition). 

Thus, indra, civdya, dkari, nadk, d&tu, camu, janayitfj dyne, 
civQyai, vnyo, agnau. 

141. Of the non-nasal mutes, only the first in each 
series, the non-aspirate surd, is allowed; the others surd 


aspirate, and both sonants whenever they would etymo- 
logically occur, are converted into this. 

Thus, aynimdf for agnimdth, su/ift for suhfd, vlrut for vlnid/t. 

In a few roots, when their final sonant aspirate) thus 
loses its aspiration, the original sonant aspiration of the 
initial reappears: compare cj k. below. 147. 

Tims, dagh becomes dhak, budh becomes Wiut, and so on. 

The roots exhibiting this change are stated below, 155. 

There was some question among the Hindu grammarians as to whether 
the final mute is to be estimated as of surd or of sonant quality; but the 
great weight of authority, and the invariable practice of the MSS., favor 
the surd. 

142. The palatals, however, form here (as often else- 
where) an exception to the rules for the other mutes. No 
palatal is allowed as final. The 3T c reverts to its original 
efi k: thus, oTTfi vak, Jbjc^j^e^ anhomuk. The $5" ch (extremely 

*X *S O ^s. L 

rare : perhaps only in the root 5f^ prach) becomes T, t : thus, 
"SnTT prat. The sf / either reverts to its original guttural or 
becomes , in accordance with its treatment in other combi- 
nations (219): thus, pTERI bhisdk, virat. The sfi/^ does not occur, 
but is by the native grammarians declared convertible to t. 

143. Of the nasals, the *T m and ^ n are extremely com- 
mon, especially the former (IT m and H s are of all final 
consonants the most frequent); the HI n is allowed, but is 
quite rare ; 3" n is found (remaining after the loss of a fol- 
lowing ^1 k) in a very small number of words; 31 n 
never occurs. 

But the final m of a root is changed to n (compare 212. 
below) : thus, dgan from gam, dnan from nam. 

144. Of the semivowels, the FT I alone is an admitted 
final, and it is very rare. The ^ r is (like its nearest surd 
correspondent, H s: 145) changed as final to visarga. Of 
ET y and of v there is no occurrence. 

x 17 -v. 

145. Of the sibilants, none may stand unaltered at the 
end of a word. The R s (which of all final consonants 

-x \ 


would otherwise be the commonest) is like ^" r changed to a 
breathing, the visarga. The $T^? either reverts to its original 
5R , or, in some roots, is changed to ^ t (in accordance 
with its changes in inflection and derivation ; see below, 218). 
The *T s is likewise changed to J. 

The change of to t is of very rare occurrence : see below, 226. 

Final radical s is said by the grammarians to be changed to t: thus, 
dhvat from dhvas : but no example of the conversion appears to occur : 
see 168. 

146. The compound ^{ ks is prescribed to be treated as 
simple *T (not becoming sfj k by 150, below). But the case 
is a rare one. and its actual treatment in the older language 

In the only RV. cases where the ks has a quasi-radical character namely 
andk from anafcs, and dmyak from ymyaks the conversion is to k. Also 
of forms of the s-aorist (see this aorist below) we have adhak, asrafc, araife, 
etc. for (adhdks-t etc.); but also aprat, dbhrat, ayat (for apraks-t etc.). 
And AV. has in two cases srah (i. e. sras), apparently for asraks-s, from 
ysrj (wrongly referred by BR. to j/srcms). 

The numeral a, 'six', is perhaps better to be regarded as safes, with 
its ks treated as s, according to the accepted rule. 

147. The aspiration ^ h is not allowed to maintain 
itself, but (like j^ / and $F g] either reverts to its original 
guttural form, appearing as 3T k, or is changed to Z t 
both in accordance with its treatment in inflection; see be- 
low, 222. And, also as in inflection, the original sonant 
aspiration of a few roots (given at 155) reappears when their 
final thus becomes unaspirated. Where the ^ h is from 
original V^dh (223), it becomes rT t. 

148. The visarga and anusvara are nowhere etymolog- 
ical finals ; the former is only the substitute for an original 
final H^5 or ^ r ; the latter occurs as final only so far as 
some later grammarians allow it to be substituted for *T m- 

149. Apart from the vowels, then, the usual finals, 
nearly in the order of their frequency, are : #, q m, ^ n, 





rf t, ^ k, ^ p, t ; those of only sporadic occurrence are 
3" n, ^l, tIT n,- and, by permitted substitution, - m. 

150. In general, only one consonant, of whatever kind, 
is allowed to stand at the end of a word; if two or more 
would etymologically occur there, the last is dropped, and 
again the last, and so on, till only one remains. 

Thus, tudants becomes tudant, and this tudan ; udanc-s becomes 
udank, and this udan; and achdntst (s-aor., 3d sing., of V chand] 
is in like manner reduced to achan. 

But a non-nasal mute, if radical and not suffixal, is re- 
tained after r : thus, urk from urj, vdrk from )A>r/, dmart from 
ymrj, suhurt from suhdrd. The case is not a common one. 

For relics of former double finals, preserved by the later language under 
the disguise of apparent euphonic combinations, see below, 207 ff. 

151. Anomalous conversions of a final mute to one of another class are 
occasionally met with. Examples are: 

a. Of final t to fc, in a few words that have assumed a special value 
as particles: thus, jytik, tdjdk (beside tajdt], prthak, rdhak ; but also now 
and then in a verbal form, as savisak (AV. and VS. Kan.), avisyak (Parask.); 
and in the feminines in km (as asiknl beside asita). SV. has once 
prks'i for prtsu. 

b. Of k or j to f, in an isolated example or two, as samydt and dsrt 
(TS., K.). 

c. In Taittiriya texts, of the final of anustubh and tristubh to a guttural : 
as, anustuk ca, tristiigbhis (Weber, Ind. St., xiii. 109 ff.). 

d. Of a labial to a dental : in kakud for and beside kaktibh ; in saihsfdbhis 
(TS.) from ysrp, an d in adbMs, adbhyds, from ap or ap (chap. V). Excepting 
the first, these look like cases of dissimilation ; yet examples of the com- 
bination bbh are not unknown in the older language: thus, kakubbhyam, 
tristubbhis, anustub bhi. 

152. For all the processes of external combination that 
is to say, in composition and sentence-collocation a stem- 
final or word-final is in general to be regarded as having, not 
its etymological form, but that given it by the rules as to per- 
mitted finals. From this, however, are to be excepted the s 
and r : the various transformations of these sounds have nothing 
to do with the visarga to which as finals before a pause they 
have come doubtless at a comparatively recent period of 
phonetic history to be reduced. Words will everywhere in 
this work be written with final s or r instead of h; and the 
rules of combination will be stated as for the two more original 
sounds, and not for the visarga. 

Whitney, Grammar. 4 



153. An aspirate mute is changed to its corresponding 
non-aspirate before another non-nasal mute or before a sib- 
ilant; it stands imaltered only before a vowel or semi- 
vowel or nasal. 

Such a case can only arise in internal combination, since the processes 
of external combination presuppose the reduction of the aspirate to a non- 
aspirate surd (152). 

Practically, also, the rules as to changes of aspirates concern almost only 
the sonant aspirates, since the surd, being of later development and rarer 
occurrence, are hardly ever found in situations that call for their application. 

154. Hence, if such a mute is to be doubled, it is 
doubled by prefixing its own corresponding non-aspirate. 

But in theMSS., both Vedic and later, an aspirate mute is not seldom 
found written double especially, if it be one of rare occurrence : for 
example (BV.), akhkhali, jdjhjhatl. 

155. In a few roots, when a final sonant aspirate (f 
gh, U dh, *T bh ; also ^ A, as representing an original 5J gh) 
thus loses its aspiration, the initial sonant consonant (JJ g 
or ^" d or Sf V) becomes aspirate. 

That is to say, the original initial aspirate of such roots is restored, 
when its presence does not interfere with the euphonic law, of comparatively 
recent origin, which (in Sanskrit as in Greek) forbids a root to both begin 
and end with an aspirate. 

The roots which show this peculiar change are : 

in gh dagh; 

in h (for original gh} dah, dih, duh, druh, drhh, guh; and grah (in 
the later desiderative jighrksa)] 

in dh bandh, badh, budh ; 

in bh dabh (but only in the later desiderative dhipsa, for which the 
older language has dipsa). 

The same change appears when the law as to finals causes the loss of 
the aspiration at the end of the root: see above, 141. 

But from dah, duh, and guh are found in the Veda also forms without 
the restored initial aspirate: thus, daksnt; aduksat, duduksa etc.; juguksa. 

The same analogy is followed by dadh, the abbreviated substitute of the 
present-stem dadha, from ydha, in some of the forms of conjugation; thus, 
dhatthas from dadh-+-thas, adhatta from adadh + ta, etc. 


Surd and Sonant Assimilation. 

156. Under this head, there is especially one very mark- 
ed and important difference between the internal combi- 
nations of a root or stem with suffixes and endings, and 
the external combinations of stem with stem in composition 
and of word with word in sentence-making: namely, 

157. In internal combination, the initial vowel or se- 
mivowel or nasal of an ending of derivation or inflection 
exercises .no altering influence upon a final consonant of the 
root or stem to which it is added. 

To this rule there are a few exceptions only : thus, a reverted palatal 
sometimes before a nasal (216. 4,5) ; d before the participle-suffix na (161) 5 
a final consonant before may a (161). 

In external combination, on the other hand, an initial 
sonant of whatever class, even a vowel or semivowel or 
nasal, requires the conversion of a final surd to sonant. 

It has been pointed out above (152) that in the rules of external com- 
bination only admitted finals, along with s and r, need be taken account of, 
all others being regarded as reduced to these before combining with initials. 

158. Final vowels, nasals, and ^T / are nowhere lia- 
ble to change in the processes of surd and sonant assimi- 

The r, however, has a corresponding surd in s, to which it 
is changed in external combination under circumstances that favor 
a surd utterance. 

159. With the exceptions above stated, the collision 
of surd and sonant sounds is avoided in combinations 
and, regularly and usually, by assimilating the final to the 
following initial. 

Thus, in internal combination : dtsi, dtti, atthds, attd (yad 
-\- si etc.); qagdhi, cagdhvdm (}/cak -\-dhi etc.); in external 
combination, dbhud ay dm, jyog jiva, sad aqitdyah, tristub dpi; 
diggaja, sadahd, brhddbhdnu, abjd. 

160. If, however, the final sonant aspirate of a root is 
followed by rT t or 5T th of an ending, the assimilation is in 



the other direction: the combination is made sonant, and 
the aspiration of the final (lost according to 153, above) is 
transferred to the initial of the ending. 

Thus, gh with t or th becomes gdh; dh with the same be- 
comes ddh, as baddhd (ybadh -f- ta), runddhds (^rundh -f- thus or 
tas); bh with the same becomes bdh, as laMhd (yiabh-\-ta), 
labdhvh, (Vldbh -\-foa). 

Moreover, h, as representing original gh, is treated in the 
same manner : thus, dugdhd, dogdhum from duh and compare 
rudhd and lldhd from ruh and Uh etc., 222. 

In this combination, as the sonant aspiration is not lost but transferred, 
the restoration of the initial aspiration (155) does not take place. 

In dadh from ydha (155, end), the more normal method is followed ; the 
dh is made surd, and the initial aspirated : thus, dhatthas, dhattas. And RV. 
has dhaktam instead of dagdham from ydagh. 

161. Before a nasal in external combination, a final 
mute may be simply made sonant, or it may be still fur- 
ther assimilated, being changed to the nasal of its own class. 

Thus, either Idd ndmas or t&n ndmas, vag me or vhn me, 
bdd mahan or ban mahan, tristub nundm or tristum nundm. 

In practice, the conversion into a nasal is almost invariably made in 
the MSB., as, indeed, it is by the Praticakhyas required and not permitted 
merely. Even by the general grammarians it is required in the compound 
sdnnavati, and before matra, and the suffix maya (really a noun in com- 
position): thus, vanmdya, mrnmdya. 

Even in internal combination, final d of a root becomes n before the 
participle-suffix na : thus, bhinnd, sannd, tunnd. 

162. Before I, a final t is not merely made sonant, but 
fully assimilated, becoming I: thus, tdl labhate, uttuptam. 

163. Before ^ h (the case occurs only in external com- 
bination), a final mute is made sonant; and then the h may 
either remain unchanged or be converted into the sonant 
aspirate corresponding with the former: thus, either rrf^ 
tad hi or rrf% tad dhi. 

In practice, the latter method is almost invariably followed; and the 
grammarians of the Prati$akhya period are nearly unanimous in requiring it. 
The phonetic difference between the two is very slight. 

Examples are: vag ghutdh, sdddhota (sat -\-hota), tod- 
dhita (tat -\-hita), anustub bhi. 

169] FINAL s AND r. 53 

Combinations of final s and r. 

164. The euphonic changes of H s and ^" r may best 
be considered together, because of the practical relation of 
the two sounds, in composition and sentence-collocation, 
as corresponding surd and sonant : in a host of cases H s 
becomes ^T r in situations requiring or favoring the occur- 
rence of a sonant ; and, less often, ^~ r becomes H s where 
a surd is required. 

In internal combination, the two are far less exchangeable 
with one another : and this class of cases may best be taken up 

165. Final r radical or quasi-radical (that is, not belonging 
to an ending of derivation) remains unchanged before both surd 
and sonant sounds, and even before su in declension : thus, pi- 
parsi, caturthd, cattirsu, pursu. 

166. Final radical s remains before a surd in general, and 
usually before s, as in cassi, cassva, aclssu (but the last is also 
written ackhsu: 172): it is however dropped in dsi (for assi : 
yas-\-si). Before a sonant (that is, bh) in declension, it is 
treated as in external combination : thus, acirbhis. Before a so- 
nant (that is, dh) in conjugation, it is dropped : thus, cadhi, 
adhvdm (but cases like the latter may be by abbreviation [232] 
for addhvam): in edhi (as -\-dhi), the preceding vowel is anoma- 
lously altered. 

167. In a very few cases, final radical s before s is changed to t (per- 
haps by dissimilation) : they are, from yvas, the future vatsydmi and aorist 
dvatsam; from yghas, the desiderative stem jighatsa. 

168. According to the grammarians, the final s of certain other roots, 
used as noun-stems, becomes t at the end of the word, and before bh and 
su: thus, dhvat, dhvadbhis, sradbhyas, sratsu. But genuine examples of such 
change do not appear to have been met with in use. 

Sporadic cases of a like conversion are found in the Veda : namely, 
madbhts and madbhyds from ma's ; usddbhis from usds ; svdtavadbhyas from 
svdtavas ; svdvadbhis etc. (not found in use), from svdvas. But the reality 
of the conversion here is open to grave doubt; it rather seems the substi- 
tution of a f-stem for a s-stem. The same is true of the change of vans 
to vat in the declension of perfect participles (chap. V.). 

In the compounds ducchuna (dus-funa) and pdrucchepa (parus-cepa), the 
final s of the first member is treated as if a t (203). 

For t as apparent ending of the 3d sing, in s-verbs, see chap. VIII. 

169. As the final consonant of derivative stems and of in- 
flected forms, both of declension and of conjugation, s is extreme- 


ly frequent; and its changes form a subject of first-rate im- 
portance in Sanskrit euphony. The r, on the other hand, is 
quite rare, being found only in certain forms of r-stems and 
in a few particles. 

The euphonic treatment of s and r yielding precisely the same result 
after all vowels except a and a, there are certain forms with regard to which 
it is uncertain whether they end in s or r, and opinions differ respecting 
them. Such are us (or ur} of the gen.-abl. sing, of r-stems, and us (or ur} 
of the 3d plur. of verbs. 

170. a. The H s, as already noticed (145), becomes 
visarga before a pause. 

b. It is retained unchanged only -when followed by 
rT t or % thj the surd mutes of its own class. 

c. Before the palatal and lingual surd mutes ^ c and 
^ ch, It and ~& th it is assimilated, becoming the sibilant 
of either class respectively, namely ST Q or Ef s . 

d. Before the guttural and labial surd mutes 3\ k 
and Ir M, Q p and Cfi ph it is also theoretically assimil- 
ated, becoming respectively the jihvamuliya and upadhma- 
niya spirants (60); but in practice these breathings are 
unknown, and the conversion is to visarga. 

Examples are: to b. tatas te, caksus te ; to c. tataq ca, ta- 
syag chaya; to d. nalah kamam, purusah khanati ; yacah prapa, 
vrksah phalavan. 

171. The first three of these rules are almost universal ; to 
the last there are numerous exceptions, the sibilant being re- 
tained (or, by 180, converted into s), especially in compounds ; 
but also, in the Veda, even in sentence combination. 

In the Veda, the retention of the sibilant in compounds is the general 
rule, the exceptions to which are detailed in the Vedic grammars. 

In the later language, the retention is mainly determined by the intimacy 
or the antiquity and frequency of the combination. Thus, the final sibilant 
of a preposition or a word filling the office of a preposition before a verbal 
root is wont to be preserved; and that of a stem before a derivative of ykr, 
before pati, before kalpa and kama, and so on. Examples are namaskara, 
vacaspati, ayuskama, payaskalpa. 

The Vedic retention of the sibilant in sentence-collocation is detailed in 
full in the Praticakhyas. The chief classes of cases are: a. the final of a 
preposition or its like before a verbal form ; b. of a genitive before a govern- 
ing noun: as divas putrdh, idds pade ; c. of an ablative before part: as 


himdvatas part; d. of other less classifiable cases: as dyaus pitd, tris putvd, 
yds pdtih, paridhfs pdtati, etc. 

172. Before an initial sibilant STc, ^ s, H s H s is 

^JT> ^.7 ^ ^ 

either assimilated, becoming the same sibilant, or it is 
changed into visarga. 

The native grammarians are in some measure at variance (see APr. ii. 40, 
note) as to which of these changes should be made, and in part they allow 
either at pleasure. The usage of the MSS. is also discordant; the conversion 
to visarga is the prevalent practice, though the sibilant is also not infre- 
quently found written. European editors generally write visarga ; but the 
later dictionaries and glossaries make the alphabetic place of a word the 
same as if the sibilant were read instead. 

Examples are : manuh svayam or manus svayam ; indrah curah 
or indrac qurah. 

173. There are one or two exceptions to these rules: 

a. If the initial sibilant has a surd mute after it, the final s may be 
dropped altogether and by some authorities is required to be so dropped. 
Thus, vayava stha or v&yavah stha,' catustanam or catuhstanam. 

With regard to this point the usage of the different MSS. and editions 
is greatly at variance. 

b. Before (s, the s is allowed to become visarga, instead of being 

174. Before a sonant, either vowel or consonant (ex- 
cept ^ r : see 179), H s is changed to the sonant ^ r 
unless, indeed, it be preceded by [ a or TT a. 

Examples are : devapatir iva, crir iva ; manur gacchati, tanur 
apsu ; fay or adrstakamah ; sarvair gunaih agner manve. 
For a few cases like duda$a, duna$a, see below, 199. 

The endings 3^q as and ^TH as (both of which are ex- 
tremely common) follow rules of their own, as follows : 

175. a. Final 5R7 as, before any sonant consonant and 
before short ^ a, is changed to m o and the 5f a after 

it is lost. 

The resulting accentuation, and the fact that the loss of a is only oc- 
casional in the older language of the Veda, have been pointed out above, 135. 

Examples are : nalo nama, brahmanyo vedavit; hantavyo 'smi. 

b. Final 3^f as before any other vowel than ^ a loses 
its H s, becoming simple ^ a ; and the hiatus thus occa- 
sioned remains. 


That is to say, as is treated as an original o, or an e, would be treated 
in the same situation: see 132 4. 

Examples are : vrhadacva uvaca, aditya iva, ndmaiikti. 

176. Exceptions to the rules as to final as are: 

a. The pronouns sas and eshas (also syas in the Veda) lose 
their s before any consonant : thus, sa dadarca, esha purusah ; but 
sada tu sah, so 'bravit. The exclamation Ihos loses its s before 
all vowels and all sonant consonants. 

b. In the Veda, and more rarely in the later language, the rule for the 
maintenance of the hiatus is sometimes violated, and the remaining con- 
tiguous vowels are combined into one : for example, se 'd ague, se 'morn, 
saw J sadhih (for sa id ague, sa imam, sa osadhih). 

c. A few instances are found (almost all Vedic) of s apparently changed 
to r after a, as after other vowels : but in nearly every case there is to be 
assumed, rather, a stem in ar beside that in as, evidences of the former 
being sometimes found in the kindred languages : thus, in forms of udhas 
and dhas (see chap. V.); in amnas (no occurrence); in bhiivas (second of the 
trio bhus, bhuvas, svar), except in its oldest occurrences; in avds (once, in 
RV.); in usds (voc., and in usarbudh}; in vddhar and vadhary (RV.); in 
vanargu, dnarvi$, vasarhdn, sabardugha, and one or two other more doubtful 
words; and in a series of words in a single passage of TS. and K., viz. 
jinvdr, ugrdr, bhimdr, tvesdr, frutdr, bhutdr, and (K. only) putdr. 

In aharpdti (VS.), and vanarsdd and vanarsdd (RV.), we see the same 
change even before a surd consonant. 

d. Final as is once changed to o in RV. before a surd consonant: thus, 
ado pito. 

177. Final 5TTCT as before any sonant, whether vowel or 
consonant, loses its *T s, becoming simple 5TT a ; and the 
hiatus thus occasioned remains. 

The maintenance of the hiatus in these cases, as in that of o and e 
and ai (above, 1334), seems to indicate a recent loss of the intermediate 
sound. Opinions are divided as to what this should have been. Some of 
the native grammarians assimilate the case of as to that of ai, assuming the 
conversion to ay in both alike but probably only as a matter of formal 
convenience in rule-making. 

178. Final ^" r, in general, shows the same form which 

H s would show under the same conditions. But 

a. Original final r, after a or a, maintains itself unchang- 
ed before a sonant : thus, punar eti, pratar-jitj dhar dkmna, dkar 

b. Also before a surd, r is preserved in a few Vedic compounds : thus, 
suarcanas, svarcaksas, yvarpati, svarsd, svbrsati; dhursdd, dhurs&h; purpati. 

181] CONVERSION OF s TO s. 57 

vdrkaryd, dfirpada, punartta. In some of these, the r is optionally retained 
even in the later language. 

c. On the other hand, r is lost, like s, in one or two Vedic cases: 
aksd fnduh, aha evd. 

179. A double r is nowhere admitted: if such would occur, 
either by retention of an original r or by conversion of * to r, 
one r is omitted, and the preceding vowel, if short, is made 
long by compensation. 

In some Vedic texts, however (Yajur-Veda), ar becomes o before ini- 
tial r; thus, svb rohdva. 

Conversion of 3U to 

v 1 

180. The dental sibilant H s is changed to the lingual 
ET 5, if immediately preceded by any vowel save Ef a and 
m a, or by fi k or ^" r unless the H s be final, or fol- 
lowed by ^~ r. 

The assimilating influence of the preceding lingual vowels and semi- 
vowel is obvious enough; that of fc and the other vowels appears to be due 
to a somewhat retracted position of the tongue in the mouth during their 
utterance, causing its tip to reach the roof of the mouth more easily at a 
point further back than the dental one. 

The general Hindu grammar prescribes the same change after a I also; 
but the Pratic.akhyas give no such rule, and phonetic considerations, the I 
being a dental sound, are decidedly against it. Actual cases of the com- 
bination, if they occur at all, are excessively rare. 

The vowels that cause the alteration of s to s may be called 
for brevity's sake "alterant" vowels. 
As a consequence of this rule, 

181. In the interior of a Sanskrit word, the dental s is 
not usually found after any vowel save a and a, but, instead 
of it, the lingual s. But 

a. A following r prevents the conversion : thus, usra, tisras, 
tamisra. And it is but seldom made in the forms and derivatives 
of a root containing an r- element (whether r or r), whatever 
the position of that element : thus, sisarti, sisrtam, sarisrpd, tistire, 
parisrut. To this rule there are a few exceptions, as visfard, 
nistrta, vispardhas, gdvisthira, etc. In ajusran the final s of a 
root is preserved even immediately before r. 

This dissimilating influence of a following r, as compared with the in- 
variable assimilating influence of a preceding r, is peculiar and problematical. 

b. The recurrence of s in successive syllables is sometimes avoided by 


leading the former s unchanged: thus, sisaksi, but sisakti ydsislsthds, but 
ydsislmahi. Similarly, in certain desiderative formations: see bel'-w, 184c. 

C. Other cases are sporadic : RV. has the forms sisice and sisicus (but 
sisicatus), and the stems rbisa, kistd, bfsa, busd, bfsaya; a single root pis, 
with its derivative pesuka, is found once in (JR. ; for puns and the roots- 
nins and hins, see below, 183. 

182. On the other hand (as was pointed out above, 62), 
the occurrence of s in Sanskrit words is nearly limited to cases 
falling under this rule : others are rather sporadic anomalies 
except where s is the product of c or ks before a dental, as in 
drastum, caste, tvastar : see 218, 221). Thus, we find: 

a. Four roots, kas, Zas, bhas, bha--, of which the last is common and 
is found as early as the Brahmanas. 

b. Further, in RV. , dsa, kavdsa, casala, casa, jdldsa, pdsyh, baskdya, 
vdsat (for vaksat?), kdsthd (for kaksta, Fick); and, by anomalous alteration 
of original s, -sdh (turdsdh, etc.), dsadha, upasttit, and probably apdsthd and 
asthivdnt. Such cases grow more common later. 

The numeral sas, as already noted, is more probably safes. 

183. The nasalization of the alterant vowel or, in other 
words, its being followed by anusvara does not prevent its 
altering effect upon the sibilant : thus, havihsi, paruhsi. And 
the alteration takes place in the initial s of an ending after the 
final s of a base, whether the latter be regarded as also changed 
to s or as converted into visarga : thus, havissu or havihsu, pa- 
russu or paruhsu. 

But the s of puns (chap. V.) remains unchanged, apparently on account of 
the retained sense of its value as pums , also that of yhihs, because of its 
value as hins (hinasti etc.); ynins (RV. only) is more questionable (perhaps 
nims, from nam). 

184. The principal cases of alteration of s in internal com- 
bination are : 

a. In endings, inflectional or derivative, beginning with s 
su; si, se, sva ; s of sibilant-aorist, future, and desiderative ; 
suffixes sna, snu, sya, etc. after a final alterant vowel or 
consonant of root or stem, or a union-vowel : thus, juhosi, $ese, 
anaisam, bhavisyami, cucruse, desna, jisnu, viksu, akarsam. 

b. The final s of a stem before an ending or suffix : thus, 
havisa, Jwvisas, etc., from havis ; caksusmant, coctska, manusa, 
manusya . 

Roots having a final sibilant (except p) after an alterant vowel are 
with the exception of fictitious ones and pis, nins, hins regarded as end- 
ing in s, not s; and concerning the treatment of this s in combination, see 
below, 225 ff. 

188] CONVERSION OF s TO s. 59 

c. The initial s of a root after a reduplication : thus, si- 
syade, susvapa, sisasati, coskuyate, sanisvanat. 

Excepted is in general an initial radical s in a desiderative stem, when 
the desiderative-sign "becomes s: thus, sisanisati from ysan, sisanksati from 

185. But the same change occurs also, on a considerable 
scale, in external combination, especially in composition : thus, 

Both in verbal forms and in derivatives, the final i or 
M of a preposition or other like prefix ordinarily lingualizes the 
initial s of the root to which it is prefixed ; since such combi- 
nations are both of great frequency and of peculiar intimacy, 
analogous with those of root or stem and affix : thus, abhisttc, 
pratisthh, ntsikta, visita; anusvadhdm, suseka. 

The principal exceptions are in accordance with the principles already 
laid down : namely, when the root contains an r-element, and when a recur- 
rence of the sibilant would take place. But there are also others, of a more 
irregular character; and tlie complete account of the treatment of initial 
radical s after a prefix would be a matter of great detail, and not worth 
giving here. 

In a few cases, the initial , usually altered after a certain prefix, 
retains the altered sibilant even after an interposed a of augment or reduplic- 
ation : thus, abhy astham, pary asasvajat, vy asahanta, ny asadama, abhy 
asincan, vy astabhnat; vi tasthe, vi tasthire. 

Much more anomalous is the occasional alteration of initial radical s 
after an a-element of a prefix. Such cases are ava stambh (against ni stambh 
and prati stambh) arid (according to the grammarians) ava svan. 

186. In other compounds, the final alterant vowel of the 
first member not infrequently (especially in the Veda) lingualizes 
the initial s of the second : for example, yudhisthira, pitrsvasr, 
gosthd, agnistomd, annstubh, trisandhi, divisdd, paramesthin, abhi- 
send, pitrsdd, purustutd. 

A very few cases occur of the same alteration after an a-element: thus, 
savyasthd, apasthd, upastut; also }/sa/i, when its final, by 146, becomes t: 
thus, satrdsdt (but satrasdham) . 

187. The final s of the first member of a compound often 
becomes s after an alterant vowel : thus, the s of a prepositional 
prefix, as nissidhvan, dustdra ffor dusstdra), aviskrta; and, regu- 
larly, a s retained instead of being converted to visarga before 
a labial or guttural mute (171) , as havispa, jyotiskft ; tapuspa, 

188. Once more, in the Veda, the same alteration, both of an initial 
and of a final s, is not infrequent even between the words composing a 
sentence. The cases are detailed in the Prati^akhya belonging to each text, 
and are of very various character. Thus, 

a. The initial s, especially of particles : as u su, M ma, kdm u svft; 


also of pronouns : as hi sdh ; of verb-forms, especially from yas : as 

hf sthd, divf stha; and in other scattering cases: as u stuhi, nu sthirdm, 
tri sadhdsthd. 

b. A final s, oftenest before pronouns (especially toneless ones): as 
(ignis tva, nis te, lyus te, cucis tvdm, sddhis tdva; but also in other cases, 
and wherever a final s is preserved, instead of being turned into visarga, 
before a guttural or labial (171): as iris putvd, dyus krnotu, vdstos pdtih, 
dyaus pits. 

Conversion of n to n. 

189. The dental nasal ^ n, when immediately followed 
by a vowel or by ^ n or JT m or Tf y or cf v, is turned in- 
to the lingual HI n if preceded in the same word by the 
lingual sibilant or semivowel or vowels that is to say, 
by *T s, ~$ r. OY ft r or ty f : and this, not only if the 
altering letter stands immediately before the nasal, but at 
whatever distance from the latter it may be found: unless, 
indeed, there intervene (a consonant moving the front of 
the tongue : namely) a palatal (except T y) , a lingual, or a 

We may thus figure to ourselves the rationale of the process: in the 
marked proclivity of the language toward lingual utterance, especially of the 
nasal, the tip of the tongue, when once reverted into the loose lingual position 
by the utterance of a non-contact lingual element, tends to hang there and 
make its next nasal contact in that position: and does so, unless the pro- 
clivity is satisfied by the utterance of a lingual mute, or the organ is thrown 
out of adjustment by the utterance of an element which causes it to assume 
a different posture. This is not the case with the gutturals or labials, which 
do not move the front part of the tongue (and, as the influence of k on 
following s shows, the guttural position favors the succession of a lingual): 
and the y is too weakly palatal to interfere with the alteration (as its next 
relative, the i-vowel, itself lingualises a s). 

This is a rule of constant application ; and (as was pointed 
out above) the great majority of occurrences of n in the language 
are the result of it. 

190. The rule has force especially 

a. When suffixes, of inflection or derivation, are added to roots or 
stems containing one of the altering sounds : thus, rudrena, rudrdndm, vdrine, 
v&rinl, vdrini, datfni, hdrani, dveshani, krindmi, prntiti, ksubhand, ghrnd, 
kdrna, vrknd, rugnd, drdvina, isdni, purand, reknas, cdksana, cfkirsamana, 

194] CONVERSION OF n TO n, 61 

b. When the final n of a root or stem comes to be followed, in inflec- 
tion or derivation, by such sounds as allow it to feel the effect of a preceding 
altering cause: thus, from }/ran, rdnanti, rdnyati, rdrana, ardnisus; from 
brahman, brdhmand, brdhmani, brdhmand, brahmanya, brdhmanvant. 

191. This rule (like that for the change of s to si applies 
strictly and especially when the nasal and the cause of its alter- 
ation both lie within the limits of the same integral word ; 
but (also like the other) it is extended, within certain limits, 
to compound words and even, in the Veda, to contiguous 
words in the sentence. 

192. Especially, a preposition or similar prefix to a root, 
if it contain r or end in euphonic r for s (174), very often 
lingualizes the n of a root or of its derived stems and forms. 

a. The initial n of a root is usually and regularly so altered, in all 
forms and derivatives, after para, pari, pra, nir (for nis), antar, dur (for 
dus) : thus, para naya, pan myate, prd nudasva ; pardnutti, parindma, pranava, 
nirnfj, durndfa. 

Roots suffering this change are written with initial n in the native root- 
lists. The only exceptions of importance are nrt, nabh, nand (very rare), 
and nay when its p becomes * (as in prdnasta], 

b. The final n of a root is lingualized in some of the forms of an arid 
han: thus, prd 'nitz, prdnd, prd hanyate, prahdnana. 

c. The class-signs nu and na are altered after the roots hi and mi: 
thus, pari hinomi, prd minanti (but the latter not in the Veda). 

d. The 1st sing. impv. ending ani is sometimes altered : thus, prd 

e. Derivatives by suffixes containing n sometimes have n by influence 
of a preposition: thus, prayana. 

f. The n of the preposition ni is sometimes altered, like the initial of 
a root, after another preposition : thus, pranipdta, pranidhi. 

193. In compound words, an altering cause in one member sometimes 
lingualizes a n ot the next following member either its initial or final n, 
or n in its inflectional or derivative ending. The exercise of the altering 
influence can be seen to depend in part upon the closeness or frequency of 
the compound, or its integration by being made the base of a derivative. 
Examples are : grdmani, trindman, urunasd ,- vrtrahdnam etc. (but vrtraghnd 
tc. : 195), nrmdnas, drughand; pravdhana, nrpdna, purydna, pitrydna; 
svargena, durgdni, usrdydmne, tryangdnam. 

194. Finally, in the Veda, a n (usually initial) is lingualized even by 
an altering sound in another word. The toneless pronouns nas and ena- are 
oftenest thus affected : thus, pdri nas. prai 'nan, {ndra enam ; but also the 
particle nd, 'like': thus, vdr nd; and a few other cases, as vdr ndma, punar 
nayamasi, agner dvena. 


195. The immediate combination of a n with a preceding guttural or 
labial seems in some cases to hinder the conversion to n: thus, vrtraghnS 
etc. ksubhnati, trpnoti (but in Veda trpnu). 

Conversion of dental mutes to linguals and palatals. 

196. When a dental mute comes in contact with, a 
lingual or palatal mute or sibilant, the dental is usually 
assimilated, becoming lingual or palatal respectively. 

The cases are the following: 

197. A dental surd mute or nasal, or the dental sibilant, 
when immediately preceded by a s, is everywhere converted into 
the corresponding lingual. 

Under this rule, the combinations at, sth, and sn are very common ; aa 
is rarely so written, the visarga being put instead of the former sibilant: 
thus, jydtihsu instead of jy6tissu. 

Those cases in which final a becomes t before su (226 b) do not, of 
course, fall under this rule. 

198. In the other (comparatively infrequent) cases where 
a dental is preceded by a lingual in internal combination, the 
dental (except of su loc. pi.) becomes lingual. Thus : 

a. A dh after d from a of a root or stem (226 b) : as ddviddhvam etc. 

b. Only a very few other instances occur: itte and ditto, from }/id; 
saddhd (also saddhd and sodha], and sanndm (aaa -(- nam : anomalous gen. 
pi. of aaa) ; trnna (ytrd-\-na). A small number of words follow the same 
rule in external combination: see below, 199. 

But tadhi (Vedic: ytad -\-dhi] shows loss of the final lingual after assi- 
milation of the dental, and compensatory lengthening. 

Some of the cases of abnormal occurrence of d are explained in a simi- 
lar way, as results of a lingualized and afterward omitted sibilant before d : 
thus nlda from nisda, ypld from pisd, j/mrd from mrsd (Zend marezhdd). 
For words exhibiting a like change in composition, see below, 199 b. 

199. In external combination, 

a. A final t is directed to be assimilated to an initial lingual mute : 
thus, tat-tlka, tad dayate, tat-thalirii, tad dhaukate : but the case never 
occurs in the older language, and very rarely in the later. For final n be- 
fore a lingual, see 205. 

b. An initial dental after a final lingual usually remains 
unchanged ; and su of the loc. pi. follows the same rule : thus, 
satsu y ratsu. 

Exceptions are: a few compounds with aaa, 'six': namely, sdnnavati, 
sannabhi (and one or two others not quotable from the literature). 


In a few compounds, moreover, there appears a lingualized dental, with 
compensatory lengthening, after a lost lingual sibilant or its representative : 
namely, in certain Vedic compounds with dus : duddbha, duddf, dudhi, du~ 
ndfa, dundfa; and, in the language of every period, certain compounds of 
sas, with change of its vowel to an alterant quality (as in vodhum and 
sodhum: 224 b: sddafa, sodhd (also saddhd and saddhd], sodant. 

C. Between final t and initial s, the insertion of a t is permitted 
or, according to some authorities, required: thus, sdt sahdsrah or salt sahdsrah. 

200. The cases of assimilation of a dental to a contiguous 
palatal occur almost only in external combination, and before 
an initial palatal. There is but one case of internal combina- 
tion, namely : 

201. A. ^ n coming to follow a palatal mute in inter- 
nal combination is itself made palatal: 

Thus, yacnh, (the only instance after c) , yajna, jajne, ajnata, 
rkjna, rhjni. 

202. An final rT t before an initial palatal mute is as- 
similated to it, becoming ^ c before rf c or ^ ch, and sT / 

before sf / (^f\ jh does not occur). 
sr ^y 

A final ^T n is assimilated before 5f j, becoming 3T n. 

All the grammarians, of every period, require this assimilation of n; 
but it is more often neglected, or only sporadically made, in the MSS. 
For n before a surd palatal, see below, 208. 

203. Before the palatal sibilant 5T p, both rT t and ^ n 
are assimilated, becoming respectively rj c and oT n; and 
then the following ST $ may be, and in practice almost 
always is, converted to ^ ch. 

Some authorities regard the conversion of p to ch as everywhere oblig- 
atory, others as only optional ; some except, peremptorily or optionally, a 
f followed by a mute. And some require the same conversion after every 
mute save m, reading also vtpat chutudrl, dnat chuci, anustup charadi, puk 
chuci. The MSS. generally write c/, instead of ccA, as result of the com- 
bination of t and p. 

Combinations of final n. 

204. Final radical n is assimilated in internal combination 
lo a following sibilant, becoming anusvara. 

Thus, vdnsi, vdhsva, vdhsat, mahsydte, tfghahsati. 

According to the grammarians, it is treated before bh and su in declen- 


sion as in external combination. But the cases are extremely rare, and RV. 
has rdnsu and vdhsu (the only Vedic examples). 

Final n of a derivative suffix is regularly and usually dropped before a 
consonant in inflection and composition in composition, even before a 
vowel; and a radical n occasionally follows the same rule. 

For assimilation of n to a preceding palatal, see 201. 
The remaining cases are those of external combination. 

205. The assimilation of n in external combination to a 
following sonant palatal and the palatal sibilant c have been 
already treated (2O2, 2O3). 

The n is also declared to be assimilated (becoming n) be- 
fore a sonant lingual (d, dh, n), but the case hardly ever 

206. A n is also assimilated to a following initial I, be- 
coming (like m : 21 3 c) a nasal I. 

The MSS. in general attempt to write the combination in accordance 
with this rule. 

207. Before the lingual and dental sibilants, s and s, final 
n remains unchanged ; but a t may also be inserted between 
the nasal and the sibilant : thus, thn sdt or thnt sdt ; mahhn sdn 
or mahhnt sdn. 

According to most of the grammarians of the Praticakhyas (not RPr.), 
the insertion of the t in such cases is a necessary one. In the MSS. it is 
very frequently made, but not uniformly. It is probably a purely phonetic 
phenomenon, a transition-sound to ease the double change of sonant to surd 
and nasal to non-nasal utterance although the not infrequent cases in 
which final n stands for original nt (as bharan, abharan, agnimari) may have 
aided to establish it as a rule. Its analogy with the conversion of n p into 
nch (203) is palpable. 

208. Before the surd palatal, lingual, and dental mutes, 
there is inserted after final n a sibilant of each of those class- 
es respectively, before which the n becomes anusvara : thus, 
hq c; nc ch; ns t; ns th; hs t; ns th. 

This rule, which in the classical language has established itself in the 
form here given, as a phonetic rule of unvarying application, really involves 
a historic survival. The large majority of cases of final n in the language 
(not far from three quarters: see APr. ii. 26, note) are for original ns ; and 
the retention of the sibilant in such, cases, when once its historical ground 
had been forgotten, was extended by analogy to all others. 

Practically, the rule applies only to n before c and t, since cases in- 
volving the other initials occur either not at all, or only with extreme rarity 
(the Veda does not present an example of any of them). In the Veda, the 
insertion is not always made, and the different texts have with regard to it 
different usages, which are fully explained in their Praticakhyas; in general, 


it is less frequent in the older texts. When the p does not appear between 
n and c, the n is assimilated, becoming n (as before j : 202). 

209. The same retention of original final s after a nasal, 
and consequent treatment of (apparent) final an, in, un, fn as 
if they were ans, Ins, uns, fns (long nasalized vowel with final s}, 
shows itself also in other Vedic forms of combination, which, 
for the sake of unity, may be briefly stated here together : 

a. Final an becomes an (nasalized a) before a following vowel : that is 
to say, ans, with nasal vowel, is treated like as, with pure vowel (177): 
thus, devdn e 'hd, upabaddhdn ihd, mahdn asi. This is an extremely com- 
mon case, especially in RV. Once or twice, the s appears as h before p: 
thus, svdtavdnh pdyuh. 

b. In like manner, s is treated after nasal z, u, f as it would be after 
those vowels when pure, becoming r before a sonant sound (174), and (much 
more rarely) h before a surd (170j : thus, rafminr iva, sununr yuvanyuhr ut, 
nfnr abhi; nfnh patram. 

210. The nasals n, n, n, occurring as finals after a short 
vowel, are doubled before any initial vowel : " thus, pratydnn ud 
esi, udydnn aditydh. 

This is also to be regarded as a historical survival, the second nasal 
being an assimilation of an original consonant following the first. It is 
always written in the MSS., although the Vedic metre seems to show* that 
the duplication was sometimes omitted. 

211. The nasal n and n before a sibilant are allowed to 
insert respectively k and t as n (207) inserts t: thus, praty- 
dnk somah. 

Combinations of final m. 

212. Final radical JT m, in internal combination, is as- 
similated to a following mute or spirant in the latter case, 
becoming anusvara ; in the former, becoming the nasal of 
the same class with the mute. 

Before m or v (as when final: 143), it is changed to n; thus, from 
|/gram, dganma, aganmahi, ganvahi, jaganvans (which appear to be the only 
quotable cases). According to the grammarians, the same change is made in 
the inflection of root-stems, before bh and su : thus, prafdnbhis, prafdnsu 
(from pra$am: pra-f- j/fam). No derived noun-stem ends in m. 

QB. has once the anomalous kdmvant, from the particle kdm. 

213. Final f m in external combination is a servile 
sound, being assimilated to any following consonant . Thus : 

* Kuhn, in his Beitrage etc., iii. 125. 
Whitney, Grammar. 


a. It remains unchanged only before a vowel or a labial 

But also, by an anomalous exception, before r of the root raj in samrtij 
and its derivatives samrdjnl and samrajya. 

b. Before a mute of any other class than labial, it becomes 
the nasal of that class. 

c. Before the semivowels y, I, v it becomes, according to 
the Hindu grammarians, a nasal semivowel, the nasal counter- 
part of each respectively (see 71). 

d. Before r, a sibilant, or h, it becomes anusvara (see 71). 
The MSS. and the editions in general make no attempt to distinguish 

the nasal tones arising from the assimilation of m before a following semivowel 
from that before a spirant. 

e. But if h be immediately followed by another consonant (which can only 
be a nasal or semivowel), the m is allowed to be assimilated to that following 

This is because the h has no position of the mouth-organs peculiar to 
itself, but is uttered in the position of the next sound. The Praticakhyas 
do not take any notice of the case. 

It has been pointed out above (73) that the assimilated 
m is generally represented in texts by the anusvara-sign, and 
that in this work it is transliterated by m (instead of a nasal 
mute or w). Also, that the general grammarians allow m to be 
pronounced before any and every consonant as anusvara. 

The palatal mutes and sibilant, and h. 

214. These sounds show in some situations a reversion to 
the original gutturals from which they are derived. The treat- 
ment of j and /j, also, is different, according as they represent 
the one or the other of two different degrees of alteration from 
their originals. 

215. The palatals and h are the least stable of alphabetic 
sounds, undergoing, in virtue of their derivative character, alter- 
ation in many cases where other similar sounds are retained. 

216. Thus, in derivation, even before vowels, semivowels, 
and nasals, reversion to guttural form is by no means rare. The 
cases are the following : 

1. Before a of suffix a, final o becomes k in ankd, arkd, pakd, vakd, 
parka, markd, vfka, prdtika etc., rc/ca, seka, moka, rokd, fo'fca, tokd, mrokd, 
vraskd: final j becomes g in tyagd, bhdga, bhagd, yaga, bhangd, sanga, 
varga, marga, mrgd, sarga, vega, bhoga, yugd, yoga, roga : final h becomes 
gh in aghd, maghd, arghd, dlrghd (and drdghiyas, drdghistha) , meghd, dogha. 
drtigha, mdgha; and in dughana. 

In another series of derivatives with a, the altered sound appears : 
examples are ajd, yaja, pucd, foca, vrajd, vevijd, yuja, urjd, doha. 


Before the suffixes as and ana, the guttural only rarely appears : namely, 
in dnkas, okas, rofcas, ptf/cas, bhdrgas; and in rogana. 

2. Before an i-vowel, the altered sound appears (except in the doubtful 
abhogi): e. g. o/f, tuji, rtici, fdci, vivid, rocisnu. 

3. Before w, the guttural reappears, as a rule (the cases are few): thus, 
anku, vanku, reku, bhrgu. 

4. Before n, the examples of reversion are few, except of j (becoming g) 
before the participial ending na : thus, reknas, vagnu (with the final also 
made sonant) ; and participles bhagnd, rugnd, etc.; and apparently prgna from 

5. Before m (of wa, man, min), the guttural generally appears: thus, 
rukmd, tigmd, fagmd (with sonant change) ; vdkman, sdkman, rukmdn; rgmtn 
and vagmm (with sonant change) : but djman, ojman, bhujmdn. 

6. Before r/, the altered sound is used : thus, pacya, yajya, yajyu, yujya, 
bhujyu. Such cases as bhogya and yogya are doubtless secondary derivatives 
from bhoga and yoga. 

7. Before r, the cases are few, and the usage apparently divided : thus, 
fukrd, ugrd, mrgra; but vdjra and pajrd(?). 

8. Before v (of the suffixes va, van, vin, etc., and participial vans] 
the guttural is regularly preserved : thus, rkvd, pakvd ; fkvan, rikvan, fukvan, 
mrgvan, yugvan; vagvfa (with further sonant change); ririkvahs, rurukvdhs, 
CUfukvdns; fupukvand, pupukvani: also before the union-vowel i in okivdns 
(RV., once). An exception is ydjvan. 

The reversion of h in derivation is exhibited only before the suffix a 
(and in the participle dughana, RV.). The final j which is analogous with 
f (219) shows much less proclivity to reversion than that which corresponds 
with c. 

9. A like reversion shows itself also to some extent in conjugational 
stem-formation and inflection. Thus, the initial radical becomes guttural 
after the reduplication in the present or perfect or desiderative or intensive 
stems of the roots ci, cit, ji, hi, han; and han becomes ghn on the elision of 
a. The RV. has vivakmi from yvac. And before ran etc. of 3d pi. mid. we 
have g for radical j in asrgran, asrgram, asasrgram (all in RV.). 

217. Final rf c of a root or stem, if followed in in- 
ternal combination by any other sound than a vowel or 
semivowel or nasal, reverts to its original guttural value, 
and shows everywhere the same form which a 5fi k would 
show in the same situation. 

Thus, vakti, uvdktha, vaksi, vaksykmi, vagdhi; vagbhis, vaksu ; 
ulita, ukthd, vaktdr. 

And, as final c becomes k (above, 142), the same rule applies 
also to c in external combination : thus, vak ca, vhg api, van me. 


Examples of c remaining unchanged in inflection are : ucydte, 
riricre, vaci, mumucmdhe. 

218. Final 5T ? reverts to its original 3\ &, in internal 
combination, only before trie "Q^s of a verbal stem or ending 
(whence, by 180, ^f ks]-, before rT t and 5T ih, it everywhere 
becomes Ef s (whence, by 197, ^ sz and "$" s% before VI dh, 
H bh, and H sw of the loc. pi., as when final (145), it re- 
gularly becomes the lingual mute ( t or 3 d). 

Thus, dviksata, veksykmi ; vdsti, vista, didestu; dididdhi, 

But a few roots exhibit the reversion of final c to k before 
bh and su, and also when final (145): they are die, drc, sprc, 
and optionally nac (always, in V.); and vif has in V. always 
viksu, loc. pl. ; but vit, vidbhis, etc. Examples are diksamcita, 
drgbhis, hrdisprk, ndk. 

Examples of c remaining unchanged before vowels etc. are: 
vici, vivicyas, avicran, acnomi, vacmi, ucmdsi. 

\ $ remains irregularly unchanged before p in the compound vifpdti. 

219. Final sf / is in one set of words treated like rf c, 

sT -v 7 

and in another set like ST f. 

Thus, from yvj ': dyukthas, dyukta, yunkte, yukli, yoktra, 
yoksyami, yuksu ; yungdhi, dyugdhvam, yugbhis. 

Again, from mrj etc.: dmrksat, sraksy&mi; mtirsti, mrstd, srsti, 
rastrd ; mrddhi, mrddhvdm, radbhis, ratsu, rat. 

To the former or yuj-cl&ss belong (as shown by their quotable forms) 
about twenty roots and radical stems: namely, bhaj, saj, tyaj (riot V.), raj 
'color', svaj, majj, nij, tij, vij, i and Ibhuj, yuj, ruj, vrj, anj, bhanj, fify ; 
urj, sraj, bhisdj, dsrj ; also, stems formed with the suffixes aj and y 
;383. 5), as trsndj, vanij : and rtvij, though containing the root yaj. 

To the latter or mrj-class belong only about one third as many : namely, 
yaj, bhrajj, vraj, raj, bhraj, mrj, srj. 

A. considerable number of j-roots are not placed in circumstances to ex- 
hibit the distinction; but such roots are in part assignable to one or the 
other class on the evidence of the related languages*. The distinction 
appears, namely, only when the j occurs as final, or is followed, either in 
inflection or in derivation, by a dental mute (t, th, dh), or, in noun-inflec- 
tion, by bh or au. In derivation (above, 216) we find a g sometimes from 
the mrj-class: thus, mdrga (in apamarga) and sarga; and before r of Vedic 
3d pi. mid. endings, asrgran, asrgram, asasrgram (beside saarjrire) while 
from the ywj-class occur only yuyujre, ayujran, bubhujrire, with j. 

* See H&bschmann, in KZ. xxiii. 384 ff. 

222] COMBINATIONS OF FINAL ch, Jcs, h. 69 

220. Final ch falls under the rules of combination almost 
only in the root prac/i, in which it is treated as if it were c 
(and prac is perhaps its more original form) : thus, praksy&mi, 

prstd; and also the derivative pracnd. As final and in noun- 
inflection (before bh and su], it is directed to be changed to the 
lingual mute. 

Murtd is called the participle of murcfc, and a gerund murtva is given 
to the same root. They (with mUrti] must doubtless come from a simpler 
form of the root. 

Of jh there is no occurrence : the grammarians declare it 
to be treated like c. 

221. The compound ks is not infrequent as final of a root 
(generally of demonstrably secondary origin), or of a tense-stem 
(s-aorist : see below, chap. XI.) ; and, in the not very frequent cases 
of its internal combination, it is treated as if a single sound, 
following the rules for c: thus, cdkse (caks -j- se ) , cdksva ; caste, 
dcasta, dsrastam, dsrsta, tvdstar. As to its treatment when final, 
see 146. We are taught by the grammarians to make such 
forms as gordt, gorddbhis, gordtsu (from gordks] ; and we actually 
have sdt, sadbhts, satsu from saks or sas (146, end). 

In the single anomalous root vrapc, the compound fc is said to follow 
the rule for simple f : thus, vraksydti, dvraksam; vrdstum, vrasta. Its c. 
reverts to fe in vraska. Its participle is vrknd ; it has a Vedic gerund vrktin. 

222. The roots in final ^ A, like those in ^j, fall into 
two classes, exhibiting a similar diversity of treatment, ap- 
pearing in the same kinds of combination. 

In the one class, as duh, we have a reversion of h (as of c] 
to a guttural form, and its treatment as if it were still its 
original gh: thus, ddhuksam, dhoksy&mi; dugdham, dugdhd; ddhok. 
d/iuk, dhugbhis, dhuksu. 

In the other class, as ruh and sah, we have a guttural rever- 
sion (as of c) only before 5 in verb-formation and derivation : thus, 
druksat, roksyftmi, saksiyd, saksdni. As final, in external combi- 
nation, and in noun-inflection before bh and su, the h (like c] 
becomes a lingual mute : thus, turasnt, prtanasad ayodhydh, 
turasndbhis, turashtsu. But before a dental mute (t t th, dh) in 
verb-inflection and in derivation, its euphonic effect is peculiarly 
complicated : it turns the dental into a lingual (as would c] ; but 
it also makes it sonant and aspirate (as would dh: see 160); 
and further, it disappears itself, and the preceding vowel, if 
short, is lengthened by way of compensation : thus, from ruh 
with ta comes rudhd, from leh with ti comes Udhi, from guh 
with tar comes gudhdr, from meh with turn comes medhum, from 


lih with tas or thas comes lid/ids, from lih with d/warn comes 
lidhvum, etc. 

This is as if we had to assume as transition sound a sonant aspirate 
lingual sibilant zh, with the euphonic effects of a lingual and of a sonant 
aspirate (160), itself disappearing under the law of the existing language 
which admits no sonant sibilant. 

223. The roots of the two classes, as shown by their forms 
found in use, are : 

of the first or dw/j-class : dah, dih, duh, druh, muh, snih 
(and the final of umih is similarly treated) ; 

of the second or rttA-claSB : yah, sah, mih, rih or li/i, guh, 
ruh, drh/i, trnh, brJi, bank, sprh(?). 

But muh forms also (not in RV.) the participle mudha and agent-noun 
mudhdr, as well as mugdhd and mugdhdr ; and druh and snih are allowed by 
the grammarians to do likewise: such forms as drudha and sriidha, however, 
do not appear to have been met with in use. 

From roots of the rwft-class we find also in the Veda the forms gartaruk, 
nom. sing., and pranadhrk and dadhfk; and hence puruspfk (the only occur- 
rence) does not prove ysprh to be of the dwft-class. 

A number of other ft-roots are not proved by their occurring forms to 
belong to either class ; they, too, are with more or less confidence assigned 
to the one or the other by comparison with the related languages*. 

In derivation, before the suffix a, we have (216.1) meghd and dlrghd 
from roots of the rwft-class. Before the r verb-endings, we have examples 
only from dw/t, with h : thus, duduhre etc. 

The root nah comes from original dh instead of gh, and its reversion is 
accordingly to a dental instead of a guttural mute: thus, natsydmi, naddhd, 
upanddbhis, upanad-yuga. 

224. Irregularities of combination are : 

a. The vowel r is not lengthened to compensate for the loss of the 
fe-element : thus, drdhd, trdhd, brdhd (the only cases ; and in the Veda their 
first syllable has metrical value as heavy or long). 

b. The roots vah and sah change their vowel to- o instead of lengthening 
it: thus, vodhdm, vodhdm, vodhdr, sddhum. But from sah in the older 
language forms with a are more frequent: thus, sadhd, dsadha (also later), 
sadhar. The root trhh changes the vowel of its class-sign na into e instead 
of lengthening it: thus, trnedhi, trnedhu, atrnet (the grammarians teach also 
trnehmi and trneksi: if such forms are actually in use, they must have been 
made by false analogy with the others). 

These anomalous vowel-changes seem to stand in connection with the 
fact that the cases showing them are the only ones where other than an 

* See Hubschmann, as above. 


alterant vowel (180) comes before the lingualised sibilant representative of 
the h. Compare sodafa etc. 

Apparently by dissimilation, the final of vah in the anomalous compound 
anadvah is changed to d instead of d: see 404. 

The lingual sibilant s. 

225. Since the lingual sibilant, in its usual and normal 
occurrences, is (182) the product of lingualization of s after 
certain alterant sounds, we might expect final radical s, when 
(in rare cases) it comes to stand where a s cannot maintain it- 
self, to revert to its original, and be treated as a s would be 
treated under the same circumstances. That, however, is only 
true in a very few instances. 

Namely, in the prefix dus (evidently identical with ydus); in sajus 
(adverbially used case-form from yjus); in (RV.) vives and dvives, from yvis; 
in aiyes (RV.), from yis; and in atfs, from ci as secondary form of y$as. 
All these, except the first two, are more or less open to question. 

226. In general, final lingual T s is treated in the same 
manner as palatal 5T $. Thus : 

a. Before t and th it remains unchanged, and the latter 
are assimilated : e.g. dvistas, dvisthas, dvestum. 

This is a common and perfectly natural combination. 

b. Before dh, bh, and su, as also in external combination 
(145), it becomes a lingual mute ; and dh is made lingual after 
it : thus, pinddhi, viddhi, vivid dhi, dvidflkvam, dvidbhis, dvitsu. 

The same holds good of the altered s of a tense-sign : thus, dstoddhvam 
(from astos-dhvam}. 

The conversion of * to * (or d) as final and before &/t and su is parallel 
with the like conversion of p , and of j and h in the mrj and ruh classes of 
roots, and perhaps with the occasional change of s to t (167 8). It is a very 
infrequent case, occurring (save as it may be assumed in the case of acts) only 
once in RV. and once in AV. (-dvit and -prut), although those texts have 
more than 40 roots with final s; in the Brahmanas, moreover, has been noticed 
further only -flit. From pins, RV. has the irregular form pinak (2d and 
3d sing., for pinas-s and pinas-t}. 

c. Before s in internal combination (except su of loc. pi.) 
it becomes k: thus, dveksi, dveksy&mi, adviksam. 

This change is of anomalous phonetic character, and difficult of ex- 
planation. It is also practically of very rare occurrence. The only RV. 
examples (apart from pinak, above) are viveksi, from yvis, and the desid. 
stem ririksa from yris ; AV. has only dviksat and dviksata, and the desid. stem 
fifliksa from y^lis. Other examples are quotable from yykra (B. etc.), fis 
(QB.), and pis (Khand. Up.); and they are by the Hindu grammarians pre- 
scribed to be formed from about half-a-dozen other roots. 


Extension and Abbreviation. 

227. As a general rule, ch is not allowed by the gramma- 
rians to stand in that form after a vowel, but is to be doubled, 
becoming cch (which the MSS. sometimes write chcfy. 

The various authorities disagree with one another in detail as to this 
duplication. According to Panini, ch is doubled within a word after either 
a long or a short vowel; and, as Initial, necessarily after a short and after 
the particles d and md, and optionally everywhere after a long. In RV., 
initial ch is doubled after a long vowel of <J only, and certain special cases 
after a short vowel are excepted. For the required usage in the other Vedic 
texts, see their several Pratoakhyas. The Kathaka writes for original ch 
(not ch from combination of t and n with c: 203) after a vowel every- 
where cch. 

Opinions are still at variance as to how far this duplication has an 
etymological ground, and how far it is only an acknowledgment of the fact 
that ch makes a heavy syllable even after a short vowel (makes "position": 79). 

228. After r, any consonant (save a spirant before a vowel) 
is by the grammarians either allowed or required to be doubled 
(an aspirate, by prefixing the corresponding non-aspirate : 154). 

Some of the authorities include, along with r, also h or I or -u, or more 
than one of them, in this rule. 

A doubled consonant after r is very common in MSS. and inscriptions, 
as also in native text-editions and in the earlier editions prepared by Euro- 
pean scholars in later ones, the duplication is universally omitted. 

The first consonant of a group whether interior, or initial after 
a vowel of a preceding word is by the grammarians either allowed or 
required to be doubled. 

This duplication is allowed by Panini and required by the Praticakhyas 
in both, with mention of authorities who deny it altogether. For certain 
exceptions, see the Praticakhyas : the meaning of the whole matter is too 
obscure to justify the giving of details here. 

230. Other cases of extension of consonant-groups, required 
by some of the grammatical authorities, are the following : 

a. Between a non-nasal and a nasal mute, the insertion of so-called 
yamas ('twins'), or nasal counterparts, is taught by the Prati^akhyas (and 
assumed in Panini's commentary): see APr. i. 99, note. 

b. Between h and a following nasal mute the Praticaktiyas teach the 
insertion of a nasal sound called nasikya : see APr. i. 100, note. 

c. Between r and a following consonant the Praticakhyas teach the 
insertion of a svardbhakti or 'vowel-fragment': see APr. i. 101 2, note. 

Some authorities assume this insertion only before a spirant; the others 
regard it as twice as long before a spirant as before any other consonant 


namely, a half or a quarter mora before the former, a quarter or an 
eighth before the latter. One (VPr.) admits it after I as well as r. It is 
variously described as a fragment of the vowel a or of r (or I). 

The RPr. puts a svarabhakti also between a sonant consonant and a 
following mute or spirant; and APr. introduces an element called sphotana 
( 'distinguished) between a guttural and a preceding mute of another class. 

For one or two other cases of yet more doubtful value, see the Prati^akhyas. 

231. After a nasal, the former of two non-nasal mutes 
may be dropped, whether homogeneous only with the nasal, or with 
both : thus, yundhi for yungdhi, yundhvdm for yungdhvdm, antdm for 
anktdm, chintkm for chintt&m, bhinthd for bhintthd, indhe for inddhe. 

This abbreviation, allowed by Panini, is required by APr. (the other 
Praticakhyas take no notice of it). It is the usual practice of the MSS., 
though the full group is also often written. 

232. In general, a double mute (including an aspirate 
which is doubled by the prefixion of a non-aspirate) in combi- 
nation with any other consonant is by the manuscripts written 
as simple. 

That is to say, the ordinary usage of the MSS. makes no difference 
between those groups in which a phonetic duplication is allowed by the rules 
given above (228, 229) and those in which the duplication is etymological. 
As every tv after a vowel may also be properly written ttv, so dattvff and 
tattvd may be, and almost invariably are, written as datvd and tatvd. As 
kartana is also properly kdrttana, so karttika (from krtt i] is written as kartika. 
So in inflection, we have always, for example, majfid etc., not majjnd, from. 
majjdn. Even in composition and sentence-collocation the same abbreviations 
are made: thus, hrdyotd foxhrddyotd; chindty asya for chindtty asya. Hence 
it is impossible to determine by the evidence of written usage whether we 
should regard adhvam or addhvam (from yas), ddvidhvam or ddviddhvam 
(from ydvis) as the true form of a second person plural. 

233. Among occasional omissions of an etymologically justified member 
of a consonant- group, is of importance enough to be here noticed* that 

A s is sometimes lost (perhaps after assimilation) between 
two surd mutes : thus, 

a. The initial s of the roots stha and stabh after the preposition ut: 
thus, utthdtum for utsthdtum, uttabhnoti for utstabhnoti. 

b. The tense-sign 8 of the s-aorist (chap. XI.) after a final consonant 
of a root before the initial consonant of an ending: thus, achantta (and for 
this, by 231, achanta] for achantsta, fapta for fapsta, taptam for tapstam. 

c. Other instances are only sporadic : thus, the compound rkthd (re -{-stha: 
PB.); the collocations tasmat tute (for stute) and puroruk tuta- (for stuta-: K.). 

Strengthening and Weakening Processes. 

234. Under this head, we take up first the changes that 
affect vowels, and then those that affect consonants adding, 


for convenience's sake, in each case, a brief notice of the vowel 
and consonant elements that have come to bear the apparent 
office of connectives. 

Guna and Vrddhi. 

^ 235. The so-called guna and vrddhi-changes are the 
most regular and frequent of vowel-changes, being of con- 
stant occurrence both in inflection and in derivation. 

A guna-vowel differs from the corresponding simple 
vowel by a prefixed ^-element, which is combined with the 
other according to the usual rules; a vrddhi-vowel, by the 
further prefixion of a to the ffwia-vowel. Thus, of 5 i or 
| the corresponding guna is (a-\-i=) ^ e; the correspond- 
ing vrddhi is (a -f- e ==) ^7 at. But in all gunating processes 
5f a remains unchanged or, as it is sometimes expressed, 
51 a is its own guna; m a, of course, remains unchanged 
for both guna and vrddhi. 

{ 236. The series of corresponding degrees is then as 
follows : 

simple vowel a a it u u r I 
Guna a a e o ar al 

Vrddhi a ai au ar 

There is nowhere any occurrence of f in a situation to undergo either 
guna or vrddTu-change ; nor does I (26) ever suffer change to vrddhi. Theo- 
retically, f would have the same changes as r; and the vrddhi of I would 
be al. 

237. The historical relations of the members of each vowel-series are 
still matters of some difference of opinion. From the special point of view 
of the Sanskrit, the simple vowels wear the aspect of being in general the 
original or fundamental ones, and the others of being products of their 
increment or strengthening, in two several degrees so that the rules of 
formation direct a, i, u, r, I to be raised to guna or vrddhi respectively, 
under specified conditions. But r is so clearly seen to come by abbreviation 
or weakening from an earlier ar (or ra) that many European grammarians 
prefer to treat the puna-form as the original and the other as the derivative. 
Thus, for example : instead of assuming certain roots to be bhr and vrdh, 
and making from them bharati and vardhati, and bhrta and vrddha, by the 
same rules which from bhu and nl and from budh and cit form bhavati and 
nayati, bodhati and cetati, bhuta and mta, buddha and citta they assume 

241 1 GUNA AND VRDDHI. 75 

bhar and vardh to be the roots, and give the rules of formation for them in 
reverse. Both methods have their advantages, and the question between 
them is one of minor consequence, which may fairly be settled by considera- 
tions of convenience. Mainly from such considerations, the r-form is gen- 
erally assumed in this work, though' not without constant recognition of 
the other. 

238. The ^ma-increment is an Indo-European phenomenon, 
and is in many cases seen to occur in connection with an accent 
on the increased syllable. It is found : 

a. In root-syllables : either in inflection, as dvfati from 
ydvis, dohmi from yduh; or in derivation, as dvesa, dohas, 
dvestum, dogdhum. 

b. In formative elements : either conjugational class-signs, 
as tanomi from tanu; or suffixes of derivation, in further deri- 
vation or in inflection, as agnaye from agni, bhandvas from bhcinu, 
pitdram from pitf, hantavyd from hdntit. 

239. The WfflUttt-increment is specifically Indian (being 
unshared, save in a few doubtful cases, even by the Zend), and 
its occurrence is more infrequent and irregular. It is found : 

a. In root and suffix-syllables, instead of guna : thus, stauti 
from ystu, sdkhayam from sdkhi, dnaisam from )/m, dkarsam and 
Mrdyati and karya from ykr (or kar), dataram from datr. 

b. Especially often, in initial syllables in secondary deri- 
vation : thus, manasd from mdnas, vaidyutd from vidyut, bhaumd 
from bhumiy parthiva from prthivl. 


240. The ^ma-increment does not usually take place in a 
heavy syllable ending with a consonant : that is to say, the 
rules prescribing guna in processes of derivation and inflection 
do not apply to a short vowel which is "long by position", nor 
to a long vowel unless it be final : thus, cetati from j/ceV, but 
nindati from ynind; ndyati from )/m, but jivati from yjlv. 

The vrddhi-increment is not liable to this restriction. 

Exceptions to -the rule are occasionally met with : thus, dideva from 
ydw; hedas from yhid; ohate from }/uh. 

And a few cases occur of prolongation instead of increment: thus, dusdyati 
from ydus, guhati from yguh. 

The changes of r (more original ar or ra) are so various 
as to call for further description. 

241. The increments of r are sometimes ra and ra, instead 
of ar and ar : namely, especially, where by such reversal a dif- 
ficult combination of consonants is avoided : thus, from ydrc, 
draksydmi and ddraksam ; but also prthu and pratk, prch and 
prach, krpa and dkrapista. 



242. In a number of roots (about a dozen quotable ones) 
ending in r (for more original ar), the r exchanges both with ar, 
and, more irregularly, in a part of the forms, with ir or also 
with ur (especially after a labial, in pr> m T> v ?> sporadically in 
others'! : which ir and ur, again, are liable to prolongation into 
ir and ur. Thus, for example, from tr (or tar}., we have tarafi, 
titarti, tatara, atarisam, by regular processes ; but also tt'rati, 
firyati, tirtva, -tirya, firna, and even (V.) turyama, tuturyat, tar- 
turana. The treatment of such roots has to be described in 
speaking of each formation. 

For the purpose of artificially indicating this peculiarity of treatment, 
such roots are by the Hindu grammarians written with long f , or with both r 
and f : no f actually appears anywhere among their forms. 

The (quotable) f-roots are 3kr 'strew', igr 'praise', Igr 'swallow', ijr 
'wear out', *r, Iff 'crush'*. 

The (quotable) r and f-roots are r, idr 'burst', Ipr 'fill', 2mr 'die', 
2w 'choose', str, hvr. 

243. In a few cases, r comes from the contraction of other syllables 
than ar and ra: thus, in trta and trtlya from ri; in prnu from ru; in trca, 
even from rir (or ryr). 

Vowel - lengthening. 

244. Vowel-lengthening has regard especially to i and u, 
since the lengthening of a is in part (except where in evident 
analogy with that of i and u) indistinguishable from its incre- 
ment, and r is made long only in certain plural cases of stems 
in r (or ar : chap. V.). Lengthening is a much more irregular and 
sporadic change than increment, and its cases will in general 
be left to be pointed out in connection with the processes of 
inflection and derivation : a few only will be mentioned here. 

245. a. Final radical i and u are especially liable to prolongation 
before y : as in passive and gerund and so on. 

b. Final radical ir and ur (from r-roots : 242) are liable to prolongation 
before all consonants except those of personal endings : namely, before y and 
tva and na: and in declension before bh and (392). Radical is and us 
have the same prolongation in declension. 

246. Compensatory lengthening, or absorption by a vowel of the time 
of a lost following consonant, is by no means common. Certain instances of 
it have been pointed out above (179, 198 b, 199b, 222). Perhaps such cases 
as pita for pitars and dhanl for dhanins (chap. V.) are to be classed here. 

247. The final vowel of a former member of a compound is often made 
long, especially in the Veda. Prolongations of final a, and before v. are most 

* Numbers prefixed to roots denote the order in which, there being more than 
one of the same form, the root intended is given in the Petersburg lexicon. 




frequent; but cases are found of every variety. Examples are devavi, 
vayundvfd, praWt, rtdvasu, mdrdvant, sadandsdd, fatamagha, vifvanara, 
ekdda$a; aplju, parlndh, virudh, tuvlmaghd, tvfsimant, fdktlvant; vasuju, 
anurudh, sumdya, puruvdsu. 

248. In the Veda, the final vowel of a word generally a, much less 
often i and u is in a large number of cases prolonged. Usually the 
prolongation takes place where it is favored by the metre, but sometimes 
even where the metre opposes the change (for details, see APr. iii. 16 note, 
Benfey, Abh. Gott. Ges. xix. xxi. [1874 6], and tlie various Prati^akhyas). 

Words of which the finals are thus treated are: 

a. Particles: namely, dthd, ddhd, evd~, utd, ghd, fca, ihd, ivd, cd, smd, 
angd, ktld, dtrd, ydtrd, tdtrd, ktitrd, anydtrd, ubhaydtrd, adyd, dchd, dpd, 
prd ; yddi, nahi, abhi ; u, tti, mi, sG, maksti. 

b. Case-forms : especially instr. sing. , as end, tend, t/ena, svend, aud 
others; rarely gen. sing., as asya, harindsyd. Cases besides these are few: 
so stmd (voc.); tanvi (loc.), and uru and (not rarely) puru. 

c. Verb-forms ending in a, in great number and variety : thus (nearly 
in the order of their comparative frequency), 2d sing. impv. act., as pibd, 
57/a, gamayd; 2d pi. act. in ta and tha, as sthd, attd, bibhrtd, jayatd, 
frn'ufa, anadatd, nayathd, jlvayathd (and one in tana: avistand); 1st pi. 
act. in ma, as vidmd, risdmd, rdhydmd, ruhemd, vanuydmd, cakrmd, mar- 
mrjmd; 2d sing. impv. mid. in sua, as yuksvd, idisvd, dadhisvd, vahasvd; 
1st and 3d sing. perf. act., as vedd, vive^d, jagrabhd; 2d sing. perf. 
act., vetthd; 2d pi. perf. act., ancya, cwfera. Of verb-forms ending in i, 
only the 2d sing. impv. act., as $rudhi, prnuhi, dldihi, jahi. 

To these may be added the gerund in ya, as abhigUryd, dcyd. 


249. The alteration of short a to an i or w-vowel in the 
formative processes of the language, except in r or ar roots (as 
explained above), is a sporadic phenomenon only. 

250. But the lightening of a long a especially to an /-vowel 
(as also its loss), is a frequent process : no other vowel is so 

a. Of the class-sign nd (of the fcn-class of verbs : chap. IX.), the d is in 
"weak" forms changed to i, and before vowel-endings dropped altogether. 
The final d of one or two roots is treated in the same manner : thus ma, hd. 
And from some roots, a and I or i-forms so interchange that it is difficult 
to classify them or to determine the true character of the root. 

b. Radical a is weakened to the semblance of the union-vowel i in 
certain verbal forms : as perfect dadima from ydd etc. ; aorist adhithds from 
ydhd etc. ; present jahimas from j/fta etc. 

C. Radical a is shortened to the semblance of stem-a in a number of 


reduplicated forms, as tistha, piba, dada, etc. : see chap. IX. ; also in a few 
aorists, as dhvam, dkhyam, etc. : see chap. XI. 

d. Radical a sometimes becomes e, especially before y : as, stheyasam, dey<i. 

251. Certain a-roots, because of their peculiar exchanges with * and 
i-forms, especially in forming the present stem, are given by the Hindu 
grammarians as roots ending in e or ai or o. Thus, from Qdha 'suck' (dhe) 
come the present dhdyati and participle and gerund dhltd, dhitva ; the other 
forms are made from dha, as dadhus, adhat, dhasyati, dhdtave, dhdpayati. 
From 2$ra 'sing' (gai) come the present g&yati, the participle and gerund 
gltd and gltvti, and passive glyate, and the other forms from ga. From 3 da 
'cut' (do) come the present dydti and participle ditd or dma, and the other 
forms from da. The iiregularities of these roots will be treated below, under 
the various formations. 

252. By a process of abbreviation essentially akin with that of ar or 
ra to r, the va (usually initial) of a number of roots becomes w, and the ya 
of a much smaller number becomes i, in certain verbal forms and derivatives. 
Thus, from vac come uvdca, ucydsam, uktva, ukta, uktf, ukthd, etc. ; from 
yaj come iydja, ijydsam, istvd, istd, tsti, etc. See below, under the various 

To this change is generally given by European grammarians the name 
of samprasarana, by adaptation of a term used in the native grammar. 

253. A short a, of root or ending, is not infrequently lost between 
consonants in a weakened syllable : thus, in verb-forms, ghndnti, dpaptam, 
jagmus, jajnus, djnata ; in noun-forms, rajne, rajni. 

254. Union-vowels. All the simple vowels come to assume in cer- 
tain cases the aspect of union-vowels, or insertions between root or stem 
and ending of inflection or of derivation. 

That character belongs oi'tenest to i, which is very widely used : a. before 
the s of aorist and future and desiderative stems, as in djlvisam, jlvisydmi, 
jijlvisami; b. in tense-inflection, especially perfect, as jijivima; occasionally 
also present, as dniti, roditi ; c. in derivation, as jivitd, khdnitum, janitf, 
rocisnu, etc. etc. 

Long I is used sometimes instead of short: thus, agrahlsam, grahlsydmi; 
bravlti, vavadlti; tarltr, savitr ; it is also often introduced before s and t of 
the 2d and 3d sing, of verbs: thus, cms, dslt. 

For details respecting these, and the more irregular and sporadic occur- 
rences of u and a-vowels in the same character, see below. 

Nasal Increment. 

255. Both in roots and in endings, a distinction of strong- 
er and \yeaker forms is very often made by the presence or 
absence of a nasal element, mute or anusvara, before a follow- 
ing consonant. In general, the stronger form is doubtless the 
more original; but in the present condition of the language, 


the nasal has come in great measure to seem, and to some 
extent also to be used, as an actually strengthening element, 
introduced under certain conditions in formative and inflective 

Examples are, of roots: ac and anc, grath and yranth, vid and vind, 
dac, and dan? , sras and mms, drh and drhh : of endings, bhdrantam and 
bhdrata, mdnasi and mdnahsi. 

256. A flrial n, whether of stem or of root, is less stable than any 
other consonant, where a weaker form is called for : thus, from rajan we 
have raja and rajabhis, and in composition raja; from dhanin, dharii and 
dhanibhis and dhani ; from yhan we have hathd and hatd t etc. A final rad- 
ical m (m does not occur as final of a stem) is sometimes treated in the 
same way: thus, from ygam, yaht, yatdm, yata, ydti. 

257. Inserted n. On the other hand, the nasal n has come to be 
used with great and, in the later history of the language, with increas- 
ing frequency as a union-consonant, inserted between vowels : thus, 
from ayni, aynina and ayninam ; from mddhu, mddhunas, mddhurii, md- 
dhuni; from fivd. fivena, $ivani, fivanam. 

258. Inserted y. After long , a y is not very infrequently found 
as apparently a mere union-consonant before another vowel: thus, in deri- 
vation, yaym, svadhayin, dhayas. sthayuka; in inflection, ddhayi, faydyati, 
and perhaps vivtiyas and gayati. 


259. Reduplication of a root (originating doubtless in its 
complete repetition) has come to be a method of radical incre- 
ment or strengthening in various formative processes : namely, 

a. in present-stem formation : as dddamt, bibhdrmi ; 

b. in aorist-stem formation: as adidharam, dcucyavam ; 

c. in perfect-stem formation, almost universally : as tathna, 
dadhau, cakara ; 

d. in intensive and desiderative-stem formation, through- 
out : as janghanti, johaviti, marmrjyate, pipasati, jighahsati ; 

e. in the formation of derivative noun-stems : as papri, 
cdrcara, sasa/ii, cikitu, malimlucd. 

Rules for the treatment of the reduplication in these several cases will 
be given in the proper connection below. 

260. As, by reason of the strengthening and weakening 
changes indicated above, the same root or stem not seldom 
exhibits, in the processes of inflection and derivation, varieties 
of stronger and weaker form, the distinction and description of 
these varieties forms an important part of the subjects hereafter 
to be treated. 

80 [261 



261. THE general subject of declension includes nouns, ad- 
jectives, and pronouns, all of which are inflected in essentially 
the same manner. But while the correspondence of nouns and 
adjectives is so close that they cannot well be separated in treat- 
ment, the pronouns, which exhibit many peculiarities, will be 
best dealt with in a separate chapter ; and the words designat- 
ing number, or numerals, also form a class peculiar enough to 
require to be presented by themselves. 

262. Declensional forms show primarily case and num- 
ber; but they also indicate gender since, though the 
distinctions of gender are made partly in the stem itself, 
they also appear, to no inconsiderable extent, in the chang- 
es of inflection. 

263. Gender. The genders are three, namely mascu- 
line, feminine, and neuter, as in the other older Indo-Euro- 
pean languages; and they follow in general the same laws 
of distribution as, for example, in Greek and Latin. 

The only words which show no sign of gender-distinction are the per- 
sonal pronouns of the first and second person (along with the numerals above 
'four': chap. VI.). 

264. Number. The numbers are three singular, 
dual, and plural. 

A few words are used only in the plural: as daras, 'wife', Spas, 
'water'; the numeral dva, 'two', is dual only; and, as in other languages, 
many words are, hy the nature of their use, found to occur only in the 

265. As to the uses of the numbers, it needs only to be 
remarked that the dual is found without the addition of the 
numeral dva, 'two', wherever the duality of the objects spoken 
of is a thing well understood : thus, agvinau, 'the two (horse- 
men) Acvins'; indrasya hdri, 'Indra's (two) bays'; devadattasya 
fastau, 'Devadatta's two hands'; but devadattasya dvav acvau stah, 
'Devadatta has two horses'. 


266. Case. The cases are (including the vocative) eight : 
nominative, accusative, instrumental, dative, ablative, gen- 
itive, locative, and vocative. 

The order in which they are here mentioned is that established for 
them by the Hindu grammarians, and accepted from these by Western schol- 
ars. The Hindu names of the cases are founded on this order : the nomi- 
native is called prathama, 'first', the accusative dvitlya, 'second', the gen- 
itive sasthi, 'sixth' (sc. vibhakti, 'division', i. e. 'case'), etc. The object sought 
in the arrangement is simply to set next to one another those cases which are 
to a greater or less extent, in one or another number, identical in form ; and, 
putting the nominative first, as leading case, there is no other order by 
which that object could be attained. The vocative is not considered and 
named by the native grammarians as a case like the rest; in this work, it 
will be given in the singular (where alone it is ever distinguished from the 
nominative otherwise than by accent) at the end of the series of cases. 

A brief compendious statement of the uses of the cases 
is given in the following paragraphs : 

267. Uses of the Nominative. The nominative is the 
case of the subject of the sentence, and of any word qualifying 
the subject, whether attributively, in apposition, or as predi- 

268. As somewhat peculiar constructions may be mentioned a predicate 
nominative with manye etc., 'think one's self to be', and with bruve etc., 'call 
one's self: thus, somam manyate papivdn (RV.), 'he thinks he has been drink- 
ing sorna'; sd manyeta puranavit (AV.), 'he may regard himself as wise in 
ancient things'; fndro brahmano bruvdnah (TS.), 'Indra calling himself (pre- 
tending to be) a Brahman'; and with rupdrh kr: thus, krsno rupdrh krtvd 
(TS.), 'taking on a black form' (making shape for himself as one that is 

269. Uses of the Accusative. The accusative is espe- 
cially the case of the direct object of a transitive verb, and of 
any word qualifying that object, as attribute or appositive or 
objective predicate. The construction of the verb is shared, of 
course, by its participles and infinitives ; but also, in Sanskrit, 
by a number of other derivatives, having a more or less parti- 
cipial or infinitival character, and even sometimes by nouns and 
adjectives. A few prepositions are accompanied by the accusative. 
As more indirect object, the accusative is construed especially 
with verbs of approach and address. It is found used yet more 
adverbially as adjunct of place or time or manner ; and a host 
of adverbs are accusative cases in form. Two accusatives are 
often found as objects of the same verb. 

270. The use of the accusative as direct object of a transitive verb 
Whitney, Grammar. 6 

g2 IV. DECLENSION. [270 

and of its infinitives and participles hardly needs illustration : an example 
or two are: agnim Ide, 'I praise Agni'; ndmo bhdrantah, 'bringing homage'; 
bh&yo d&tum arhasi, 'thou shouldst give more'. Of predicate words quali- 
fying the object, an example is tdm ugrdrh krnomi tarn brdhmdnam (RV.j, 
'him I make formidable, him a priest'. 

271. Of verbal derivatives having so far a participial character that 
they share the construction of the verb, the variety is considerable: thus, 

a. Derivatives in u from desiderative stems have wholly the character 
of present participles : thus, damayantim abhipsavah (MBh.), 'desiring to win 
Damayanti': rajdnam didrksuh (MBh.), 'desiring to see the king. 

b. So-called primary derivatives in in have the same character : thus, 
mam kaminl (AV.), 'loving me'; enam abhibhdsinl (MBh.), 'addressing him'. 

C. Derivatives in oka, in the later language : as, bhavantam abhivddakah 
(MBh.), 'intending to salute you'. 

d. Nouns in tar, very frequently in the older language, and as peri- 
phrastic future forms (chap. XII.) in the later: thus, hdntd y6 vrtrdm sdnito 
*ta vajam data maghdni (RV.), 'who slayeth the dragon, winneth booty, bestow- 
eth largesses'. 

e. The root itself, in the older language, used with the value of a 
present participle at the end of a compound: thus, yam yajndm paribhur 
dsi (RV.), 'what offering thou surroundest (protectest)'. 

f. The derivative in i from the (especially the reduplicated) root, in the 
older language: thus, babhrir vdjram papih somam dadir gdh (RV.), 'bearing 
the thunderbolt, drinking the soma, bestowing kine'. 

g. Derivatives in uka, very frequently in the Brahmana language : thus, 
vatsahc ca ghdtuko wrkah (AV.), 'and the wolf destroys his calves'; veduko 
vdso bhavati (TS.), 'he wins a garment'. 

h. Other cases are more sporadic : thus, derivatives in a, as fndro drdhd 
cid drujdh (RV.), 'Indra breaks up even what is fast'; in atnu, as vldu cid 
drujatnubhih (RV.), 'with the breakers of whatever is strong'; in ana, as 
tarn nivarane (MBh.), 'in restraining him'. 

272. Examples of an accu ative with an ordinary noun or adjective are 
only occasional: anuvrata is so used: thus, damayantim anuvratah (MBh.), 
'devoted to Damayanti': and kama, as mdm kdmena (AV.), 'through lov- 
ing me'. 

273. The direct construction of cases with prepositions is comparatively 
restricted in Sanskrit (see the subject of Prepositions, chap. XVI.). With the 
accusative is oftenest found prati, 'opposite to, in reference to', etc.; also anu, 
'after, in the course of; antar or antara, 'between'; rarely ati, 'across'; abhi, 
'against, to'; and others. Case-forms which have assumed a prepositional 
value are also often used with the accusative : as antarena, uttarena, daksin- 
ena, avarena, urdhvam, rte. 

274. The accusative is very often found also as object of verbs which 
in the related languages are not transitive. 

a. It stands especially as the goal of motion, with verbs of going, bring- 


ing, sending, and the like : thus, vidarbhdn agaman (MBh.), 'they went to 
Vidarbha'; divam yayuh (MBh.), 'they went to heaven'; vanagulmdn dhd- 
vantah (MBh.), 'running to woods and bushes'; apo dwam lid vahanti (AV.), 
'they carry up waters to the sky'; devd~n yaje (AV.), 'I make offering to the 

With verbs meaning 'go', this is an extremely common construction; 
and the use of such a verb with an abstract noun makes peculiar phrases 
of 'becoming': thus, samatdm eti, 'he goes to equality' (i. e. 'becomes equal'); 
sa gached badhyatdm mama (MBh.), 'he shall become liable to be slain by 
me'; sa pancatvam dgatah (H.), 'he was resolved into the five elements' (un- 
derwent dissolution, died). 

b. Verbs of speaking follow the same rule : thus, tarn abravit, 'he said 
to him'; prdkrocad uccdir ndisadham (MBh.), 'she cried out loudly to the Nisha- 
dhan'; yds tvo : vd~ca (AV.), 'who spoke to thee J . 

275. The cognate accusative, or accusative of the implied object, is, of 
course, found also in Sanskrit : as, ?apatham ?epe, 'he swore an oath'. 

276. As a yet more adverbial adjunct to a verb, the accusative is used 
to denote space traversed : as yojana$atam gantum (MBh.), 'to go a hundred 
leagues'; much more often, duration of time: as tisro ratrir dlksitdh sydt 
(TS.), 'let him be consecrated three nights'; sd samvatsardm urdhvb l tisthat 
(AV.), 'he stood upright for a year'; gatvd trin ahordtrdn (MBh.), 'having 
travelled three complete days'. But the point of place or time also is occasio- 
nally found represented by the accusative (instead of the locative). 

The same case is used adverbially to express manner or accompanying 
circumstance; and many adverbs have the accusative form (see Adverbs, 
chap. XVI.): this is especially true of compound words, even to such an 
extent as to forma class of adverbial compounds (below: chap. XVIII.). 

277. The accusative is freely used along with other cases as objects of 
the same verb, wherever the sense admits. And when it is usable with a 
verb in two different constructions, the verb may take two accusatives, one 
in each construction. Thus, especially, verbs of having recourse, appealing, 
asking : as, tvdm vayarh $aranam gatdh (MBh.), 'we have resorted to tbee 
for succor'; ap6 ydcdmi bhesajdm (RV.), 'I ask the waters for medicine'; 
and verbs of sending, bringing, imparting, saying : as, gurutvarh nararh 
nayanti (H.), 'conduct a man to respectability'; tarn idam abravit (MBh.), 
; she said this to her'; other less usual cases are, for example, tdm visdm 
evd J dhok (AV.), 'he milked from her poison'; jitvd rdjyarh nalam (MBh.), 
'having won the kingdom from (i. e. by beating) Nala'. 

A causative verb, naturally, takes a double accusative : thus, tdm grham 
prave?ayati, 'he makes her enter the house'. 

278. Uses of the Instrumental. The instrumental is 
originally the Svith'-case : it denotes adjacency, accompaniment, 
association passing over into the expression of means and 
instrument (by the same transfer of meaning which appears in 
our prepositions with and by}. 



Nearly all the uses of the case are readily deducible from this funda- 
mental meaning, and show nothing anomalous or difficult. 

279. The instrumental is often used to signify accompaniment : thus, 
agnir devebhir d gamat (RV.), 'may Agni come hither along with the gods'; 
dvdparena sahdyena kva ydsyasi (MBh.), 'whither wilt thou go, with Dvapara 
for companion?' kathayan ndisadhena, (MBh.), 'talking with the Nishadhan'. 
But the relation of simple accompaniment is more often helped to plainer 
expression by prepositions. 

280. The instrumental of means or instrument or agent is yet more 
frequent : thus, bhadrdm kdrnebhih prnuyama (RV.), 'may we hear with our 
ears what is propitious'; fastrena nidhanam (MBh.), 'death by the sword'; 
kecit padbhydrh hatd gajdih (MBh.), 'some were slain by the elephants with 
their feet'. And this passes easily over into the expression of occasion or 
reason (for which the ablative is more frequent) : thus, krpayd, 'through pity'; 
tena satyena, 'by reason of that truth'. 

281. Of special applications, the following may be noticed: 

a. Accordance, equality, likeness, and the like : thus, samdrn jyotih 
suryena (AV.), 'a brightness equal with the sun'; yeshdrh na pddarajasd 
tulyah (MBh.), 'to the dust of whose feet I am not equal'. 

b. Price (by which obtained): thus, dafdbhir krindti dhentibhih (RV.), 'he 
buys with ten kine'; gavdm sahasrena vikrinite, 'he sells for a thousand 

c. Medium, also space or distance or road, traversed; and hence also time 
passed through : thus, udnd na ndvam anayanta (RV.), 'they led [him] as it 
were a ship through the water'; e 'hd ydtam pathfbhir devaydnaih (RV.), 'come 
hither by god-travelled paths'; jagmur vihdyasd (MBh.), 'they went off through 
the air': vidarbhdn ydtum ichdmy ekdhnd (MBh.), 'I wish to go to Vidarbha 
in the course of one day'; kiyatd kdlena pradhdnatdm labhante (H.), 'in how 
long time do they obtain chief rank?' But the expression of 'within the 
space of, for time, is also sometimes made by the accusative or locative. 

d. The part of the body on (or by) which anything is borne is usually 
expressed by the instrumental : as, kukkurah skandheno 'hyate (H.), 'a dog is 
carried on the shoulder'; and this construction is extended to such cases as 
tulayd krtam (H.), 'put on (i. e. so as to be carried by) a balance'. 

282. The construction of a passive verb (or participle) with an 
instrumental of the agent is common from the earliest period, and becomes 
decidedly more so later, the passive participle with instrumental taking to 
no small extent the place of an active verb with its subject. Thus, yamena 
dattdh (RV.), 'given by Tama'; ffibhir idyah (RV.), 'to be praised by sages'; 
vyddhena jdlarh vistlrnam (H.), 'by the hunter a net [was] spread'; teno 
J ktam, 'by him [it was] said'; mayd gantavyam (H.), = 'I shall go'. 

The second object of a causative verb is sometimes put in the instru- 
mental instead of the accusative : as, paksibhih pinddn khddayati, 'he causes 
the cakes to be eaten by the birds'. 

283. Many instrumental constructions are such as call in translation for 


other prepositions than 'with' or 'by'; yet the true instrumental relation is 
usually to be traced, especially if the etymological sense of the words be 
carefully considered. 

More anomalously, however, the instrumental is used interchangeably 
with the ablative with words signifying separation : thus, vatsafr vfyutah 
(RV.), 'separated from their calves'; md~ 'ham atmdna vi radhisi (AV.), 'let 
me not be severed from the breath of life'; sa taya vyayujyata (MBh.j, 'he 
was parted from her' (compare English parted with}. 

284. The prepositions taking the instrumental are those signifying 
'with' and the like : thus, saha, with the adverbial words containing sa as an 
element, as sakam, sardham, saratham; and, in general, a word compounded 
with sa, sam, saha takes an instrumental as its regular and natural com- 

But also the preposition uma, 'without' (along with the compounds of 
vi spoken of in the preceding paragraph), takes sometimes the instrumental. 

285. Uses of the Dative. The dative is the case of 
the indirect object of that toward or in the direction of or 
in order to or for which anything is or is done (either intrans- 
itively or to a direct object). 

In more physical connections, the uses of the dative approach those of 
the accusative (the more proper 'to'-case), and the two are sometimes inter- 
changeable ; but the general value of the dative as the 'for'-case is almost 
everywhere distinctly to be traced. 

286. Thus, the dative is used with 

a. Words signifying give, share out, assign, and the like : thus, yo nd 
dddati sdkhye (RV.), 'who gives not to a friend'; ydcha 'smai fdrma (RV.), 
'bestow upon him protection'. 

b. Words signifying show, announce, declare, arid the like : thus, avtr 
ebhyo abhavat suryah (RV.), 'the sun was manifested to them'; rtuparnam 
bhlmaya pratyavedayan (MBh.), 'they announced Rituparna to Bhima'; tebhyah 
pratijnaya (MBh.), 'having promised to them'. 

c. Words signifying give attention, have a regard or feeling, and the 
like; thus, nive$aya memo dadhuh (MBh.), 'they set their minds upon en- 
camping'; mate *va putrebhyo mrda (AV.), 'be gracious as a mother to her 
sons'; kim asmdbhyam hrnlse (RV.), 'why art thou angry at us?' 

d. Words signifying inclination, obeisance, and the like : thus, mdhyam 
namantam pradfyac cdtasrah (RV.), 'let the four quarters bow themselves to 
me'; devebhyo namaskrtya (MBh.), 'having paid homage to the gods'. 

e. Words signifying hurling or casting: as ycna dudSfe dsyasi (AV.), 
'with which thou hurlest at the impious'. 

In some of these constructions the genitive and locative are also used : 
see below. 

287. In its more distinctive sense, as signifying for, for the benefit of, 
with reference to, and the like, the dative is used freely, and in a great 
variety of constructions. And this use passes over into that of the dative of 

gg IV. DECLENSION. [287 

end or purpose, which is extremely common. Thus, isurh krnvana dsanaya 
(AV.), 'making an arrow for hurling'; grhndmi te saubhagatvdya hdstam (RV.), 
'I take thy hand in order to happiness'; rdstrdya mdhyam badhyatdm sa- 
pdtnebhyah pardbhuve (AV.), 'be it bound on in order to royalty for me, in 
order to destruction for my enemies'. Such a dative is much used predicat- 
ively (and oftenest with the copula omitted), in the sense of 'makes for', 
l tends toward'; also 'is intended for, and so 'must'; or 'is liable to', and so 
'can'. Thus, upadeco murkhdndm prakopdya na cdntaye (H.), 'good counsel 
[tends] to the exasperation, not the conciliation, of fools'; sa ca tasydh sarh- 
tosdya na 'bhavat (H.), 'and he was not to her satisfaction'; sugopd asi na 
ddbhdya (RV.), thou art a good herdsman, not one for cheating' (i. e. not to 
be cheated). 

These uses of the dative are in the older language especially illustrated 
by the dative infinitives, for which see chap. XIII. 

288. The dative is not used with prepositions. 

289. Uses of the Ablative. The ablative is the 'from'- 
case, in the various senses of that preposition : it is used to 
express removal, separation, distinction, issue, and the like. 

290. The ablative is used where expulsion, removal, distinction, release, 
protection, and other kindred relations are expressed : thus, te sedhanti pathd 
vrkam (AV.), 'they drive away the wolf from the path'; md prd gdma pathdh 
(RV.), 'may we not go away from the path': are asmdd astu hetih (AV.), 'far 
from us be your missile'; pdtdm no vfkdt (RV.), 'save us from the wolf. 

291. The ablative is used where procedure or issue from something as 
from a source or starting-point is signified : thus, cukrd krsndd ajanista 
(RV.), 'the bright one has been born from the black one'; lobhdt krodhah 
prabhavati (MBh.), 'passion arises from greed'; vatdt te prdndm avidam (AV.), 
'I have won thy life-breath from the wind'; ye prdcyd died abhiddsanty 
asmdn (AV.), 'who attack us from the eastern quarter'; tac chrutvd sakhlgan- 
dt (MBh.), 'having heard that from the troop of friends'; vdyur antariksdd 
abhdsata (MBh.), 'the wind spoke from the sky'. 

Hence also, procedure as from a cause or occasion is signified by the 
ablative : this is especially frequent in the later language, and in technical 
phraseology is a standing construction; it borders on instrumental construc- 
tions. Thus, vdjrasya ciisndd daddra (RV.), 'from (by reason of) the fury 
of the thunderbolt he burst asunder; yasya dandabhaydt sarve dharmam 
anurudhyanti (MBh.), 'from fear of whose rod all are constant to duty'; afca- 
ramicritatvdd ekarasya (Tribh.), 'because e contains an element of a'. 

Perhaps by a further modification of this construction (the effect following 
the cause), the ablative also in technical language signifies 'after': thus, 
sakarat, 'after s\ 

292. One or two special applications of the ablative construction are 
to be noticed: 

a. The ablative with words implying fear (terrified recoil from): thus, 


tdsyd jdtdyah sdrvam abibhet (AV.), 'everything was afraid of her at her 
birth'; ydsmdd rejanta krstdyah (RV.), 'at whom mortals tremble 1 ; yusmdd 
bhiyd (RV.), 'through fear of you'. 

b. The ablative of comparison (distinction from): thus, prd ririce divd 
mdrah prthivydh (RV.), 'Indra is greater than the heaven and the earth'. 
With a comparative (or other word used in a kindred way), the ablative is 
the regular and almost constant construction : thus, svaddh svddiyah (RV.), 
'sweeter than the sweet'; kirh tasmdd duhkhataram (MBh.), 'what is more painful 
than that?" mitrdd anyah (H.), 'any other than a friend'; ndisadhad any am 
(MBh.), 'another than the Nishadhan' ; sa matto ma/tan, 'he is greater than 1' ; 
tad anydtra tvdn nf dadhmasi (AV.), 'we set this down elsewhere (away) 
from thee'; purva vfyvasmad bhtivanat (RV.), 'earlier than all beings'. 

Occasionally, a partitive genitive is used with the comparative (as with 
the superlative); or an instrumental (of holding together the things compared). 

293. The ablative is used with a variety of prepositions and words 
sharing a prepositional character; but all these have rather an adverbial value, 
as strengthening or denning the 'from'-relation, than any proper governing 
force. Thus : 

a. In the Veda, ddhi and pdri are much used as directing and strength- 
ening adjuncts with the ablative : as, jato himdvatas pdri (AV.), 'born from 
the Himalaya (forth)'; samudrdd ddhi jajnise (AV.), 'thou art bom from 
the ocean'; cdrantam pdri tastMsah (RV.), 'moving forth from that which 
stands fast'. 

b. Also purd (and purds], in the sense of 'forward from 1 , and hence 
'before': as, purd jdrasah (RV.), 'before old age' : and hence also, with words 
of protection and the like, 'from': as fafamdndh purd niddh (RV.), 'setting 
in security from ill-will'. 

c. Also d, in the sense of 'hither from, all the way from': as, d m&lad 
dnu fusyatu (AV.), 'let it dry completely up from the root' ; tdsmdd d nadyb 
ndma siha (AV.), 'since that time ye are called rivers'. But usually, and 
especially in the later language, the measurement of interval implied in d, 
is reversed in direction, and the construction means 'all the way to. until' : 
as, a 'syd yajndsyo 'drca/i (VS.;, 'until the end of this sacrifice'; d sodayat 
(M.), 'till the sixteenth year'; a pradandt ((,.), 'until her marriage'. 

d. Other prepositional constructions offer little subject for remark: vina 
takes the ablative as well as instrumental (284); rte in the older language 
usually has the ablative, in the later the accusative. 

294. Uses of the Genitive. The proper value of the 
genitive is adjectival ; it belongs to and qualifies a noun, de- 
signating something relating to the latter in a manner which 
the nature of the case, or the connection, defines more nearly. 
Other genitive constructions, with adjective or verb or prepo- 
sition, appear to arise out of this by a more or less distinctly 
traceable connection . 

The use of the genitive has become much extended, espe- 

88 IV. DECLENSION. [294 

cially in the later language, by attribution of a noun-character 
to the adjective, and by pregnant verbal construction, so that it 
often bears the aspect of being a substitute for other cases as 
dative, instrumental, locative. 

295. The genitive in its normal adjective construction is classifiable 
into the usual varieties, as : genitive of possession or appurtenance, including 
the complement of implied relation this is, as elsewhere, the commonest 
of all; the so-called partitive genitive; the subjective and objective genitives; 
and so on. Genitives of appellation ('city of Rome'), material, and charac- 
teristic ('man of honor'), do not occur. Examples are: indrasya vajrah, 
'India's thunderbolt'; pita putranam, 'father of sons'; putrdh pituh, 'son of 
the father' ; pituh kamah putrasya, 'the father's love of the son' ; ke nah, 
'which of us'; fatarh dasmam, 'a hundred female slaves'. 

The expression of possession etc. on the part of pronouns is made 
almost entirely by the genitive case, and not by a derived possessive ad- 
jective (516). 

296. The genitive is dependent on an adjective: 

a. In part, by a construction similar to that of verbs which take a 
genitive object: thus, abhijna rajadharmanam. 'understanding the duties of 
a king'. 

b. In great part, by a transfer of the possessive genitive from noun to 
adjective, the adjective being treated as if it had noun-value: thus, tasya 
samah or anurupah or sadrfah, 'resembling him' (i. e. his like); tasya priya, 
'dear to him' (his dear one); tasya 'viditam, 'unknown to him' (his unknown 
thing); havya? carsariindm (RV.), 'to be sacrificed to by mortals' (their object 
of sacrifice); Ipsito naranarinam (MBh.), 'desired of men and women' (their 
object of desire); yasya kasya prasutah (H.), 'of whomsoever born' (his son). 

c. The so-called partitive genitive with a superlative, or another word 
of similar value, is a matter of course : thus, frestham viranam, "best of 
heroes'; vlriidhaih vlryavatl (AV.), 'of plants the mighty (mightiest) one'. 

d. Adjectives meaning 'capable', 'worthy', 'full', and a few others, take 
the genitive by a more original and proper right. 

297. The genitive as object of a verb is: 

a. A possessive genitive of the recipient, by pregnant construction, with 
verbs signifying give, impart, communicate, and the like : thus, varan pra- 
daya 'sya (MBh.), 'having bestowed gifts upon him' (made them his by 
bestowal); rajfto niveditam (H.), 'it was made known to the king' (made his 
by knowledge). 

This construction, by which the genitive becomes substitute for a dative 
or locative, abounds in the later language, and is extended sometimes to 
problematic and difficult cases. 

b. A (in most cases, probably) partitive genitive, as a less complete or 
less absolute object than an accusative : thus, with verbs meaning partake 
(eat, drink, etc.), as piba sutdsya (AV.), 'drink (of) the Soma'; mddhvah 
payaya (RV.), 'cause to drink the sweet draught'; with verbs meaning 




impart (of the thing imparted) etc., as d'idata no amftasya (RV.), 'bestow 
upon us immortality'; with verbs meaning enjoy, be satisfied or filled 
with: as mdtsy dndhasah (RV.), 'do thou enjoy the juice'; -- with verbs 
meaning to perceive, note, care for, regard with feelings of various kinds : 
as, vdsisthasya stuvatd indro a$rot (RV.), Indra listened to Vasishtha who was 
praising him'; ydtha mama smdrat (AV.), 'that he may think of me'; tasya 
cukopa (MBh.), 'he was angry at him'; bibhlmas tava (MBh.), 'we are afraid 
of thee'. 

c. A genitive of more doubtful character, with verbs meaning to rule 
or have authority, as tvdm i$ise vdsunam (RV.), 'thou art lord of good 
things'; ydtha 'ham esam virdjani (AV.), 'that I may rule over them'; 
with verbs meaning throw at, injure, as yds ta dsyat (AV.), 'whoever hurl- 
ed at thee'; and with some others. 

298. A genitive in its usual possessive sense is often found as predi- 
cate, and not seldom with the copula omitted : thus, ydthd 'so mama keva- 
lah (AV.), 'that thou mayest be wholly mine'; sarvah sampattayas tasya 
samtustam yasya manasam (H.), 'all good fortunes are his who has a con- 
tented mind'. 

299. The prepositional constructions of the genitive are for the most 
part with such prepositions as are really noun-cases, and have the govern- 
ment of such : thus, agre, arthe, tote, and the like ; also with other prepo- 
sitional words which, in the general looseness of use of the genitive, have 
become assimilated to these. A few more real prepositions take the gen- 
itive : either usually, like updri, 'above', or occasionally, like adhds, antdr, dtL 

A genitive is used in the older language with certain adverbs of time : 
thus, sakfd dhnah (RV.), 'once a day 1 ; trih samvatsarasya, 'thrice a year'; 
idanim dhnah (RV.), 'at this time of the day'. 

300. The genitive is used adverbially hardly at all; a few genitives of 
time occur in the older language: as aktos, 'by night', vastos, 'by day'. 

A genitive of accompanying circumstance, with a qualifying word, is 
sometimes used absolutely, instead of a locative (303 b); but this construction 
is unknown in the earlier language, and rare in the later. It is said by 
the grammarians to convey an implication of contempt: thus, vaidarbhyaJi 
preksamanayah panakalam amanyata (MBh.), 'he thought it a time for stak- 
ing at play, while the Vidarbhan was looking on' (or, in spite of it) ; but it 
is found without any such implication: thus, samkhyasyami phalany asya 
pavyatas te (MBh.), 'I will count its fruits while you look on 1 . 

301. Uses of the Locative. The locative is properly 
the ''irT-case, the case expressing situation or location ; but its 
sphere of use has been somewhat extended, so as to touch and 
overlap the boundaries of other cases, for which it seems to be 
a substitute. 

Unimportant variations of the sense of 'in' are those of 
'amid' or 'among', 'on , and 'at\ Of course, also, situation in 
time as well as place is indicated by the case ; and it is applied 

90 IV. DECLENSION. [301 

to yet less physical relations, to sphere of action and feeling 
and knowledge, to state of things, to accompanying circumstance ; 
and out of this last grows the frequent use of the locative as 
the case absolute. 

Moreover, by a pregnant construction, the locative is used 
to denote the place of rest or cessation of action or motion 
('into' or 'on to' instead of 'in' or 'on'; German in with accu- 
sative instead of dative : compare English there for thither], 

302. The locative of situation in space hardly needs illustration. An 
example or two are : ye, devd divt sthd (AV.), 'which of you gods are in heav- 
en'; na devesu na yaksesu tddrk (MBh.), 'not among gods or Yakshas is such 
a one'; pdrvatasya pfsthe (RV.), 'on the ridge of the mountain'; viddthe 
santu devdh (RV.), 'may the gods be at the assembly'; dacame pade (MBh.), 
'at the tenth step'. 

The locative of time indicates the point of time at which anything 
takes place : thus, asyd usdso vyitstdu (RV.), 'at the shining forth of this 
dawn'; etasminn eva kale (MBh.), 'at just that time'; dvddace varshe (MBh.), 
'in the tenth year'. But the accusative is occasionally used in this sense, 
instead of the locative. 

303. The locative of sphere or condition or circumstance is of very 
frequent use: thus, made dhim jaghdna (RV.), 'in fury Indra slew 
the dragon'; mitrdsya sumatdu sydma (RV.), 'may we be in the favor of 
Mitra'; te vacane ratam (MBh.), 'delighted in thy words'. 

a. This construction is, on the one hand, generalized into an expres- 
sion for 'in the matter or case of, or 'with reference to, respecting', and 
takes in the later language a very wide range, touching upon genitive and 
dative constructions : thus, e J mdm bhaja grame devesu gosu (AV.), 'be gen- 
erous to him in retainers, in horses, in cattle'; tdm it sakhitvd Imahe (RV.), 
'him we beg for friendship'; nydyo l yam mayd drsta dnayane tava (MBh.), 
'this means was devised by me for (with reference to) bringing thee hither'; 
satltve kdranam striydh (H.), 'the cause of (in the case of) a woman's chast- 
ity'; na cakto 'bhavan nivdrane (MBh.), 'he was not capable of preventing'. 

b. On the other hand, the expression by the locative of a condition of 
things in which anything takes place, or of a conditioning or accompanying 
circumstance, passes over into a well-marked absolute construction, which is 
known even in the earliest stage of the language, but becomes more fre- 
quent later. 

Transitional examples are : have tvd sura udite have madhydndine divdh 
(RV.), 'I call to thee at the arisen sun (when the sun has risen), I call at 
midtime of the day'; aparddhe krte 'pi ca na me kopah (MBh.), 'and even 
in case of an offense committed, there is no anger on my part'. 

The normal condition of the absolute construction is with a participle 
accompanying the noun : thus, stlrne barhisi samidhdne agndu (RV.), 'when 
the barhis is strewn and the fire kindled'; kale fubhe prdpte (MBh.), 'a pro- 
pitious time having arrived'; avasanndydrh rdtrdv astdcalacuddvalambini can- 


dramasi (H.), 'the night having drawn to a close, and the moon resting on 
the summit of the western mountain'. 

But the noun may be wanting, or may be replaced by an adverbial sub- 
stitute (as euam, tathd, iti): thus, asmdbhih samanujndte (MBh.), '[it] being 
fully assented to by us'; evam ukte kalind (MBh.), 'it being thus spoken by 
Kali'; tatha 'nusthite (H.), 'it being thus accomplished'. So likewise the 
participle may be wanting (a copula sati or the like being to be supplied): 
thus, dure bhaye, 'the cause of fear being remote'; while, on the other hand, 
the participle sati etc. is sometimes redundantly added to the other parti- 
ciple : thus, tatha krte sati, 'it being thus done'. 

C. The locative is frequently used adverbially or prepositionally : thus, 
arthe or krte, 'in the matter of, for the sake of. 

304. The pregnant construction by which the locative comes to express 
the goal or object of motion or action or feeling exercised is not uncommon 
from the earliest time. It is by no means to be sharply distinguished from 
the ordinary construction ; the two pass into one another, with a doubtful 
territory between. It occurs : 

a. Especially with verbs, as of arriving, sending, placing, communicating, 
bestowing, and many others, in situations where an accusative or a dative 
(or a genitive, 297a) might be looked for, and exchangeable with them: 
thus, sd id devesu gachati (RV.), 'that, truly, goes to (to be among) the 
gods' ; imam no yajndm amftesu dhehi (RV.), 'set this offering of ours among 
the immortals'; yd dsincdnti rdsam osadhisu (AV), 'who pour in the juice 
into the plants' (or, the juice that is in the plants); ma prayache "$vare dhanam 
(H.), 'do not offer wealth to a lord'; papdta medinydm (MBh.), 'he fell to (so 
as to be upon) the earth'; skandhe krtva (H.), 'putting on the shoulder'; 
samfrutya purvam asmdsu (MBh.), 'having before promised us 1 . 

b. Often also with nouns and adjectives in similar constructions (the 
instances not always easy to separate from those of the locative meaning 
'with reference to': above, 303 a): thus, dayd sarvabhutesu, 'compassion 
toward all creatures' ; anurdgam ndisadhe (MBh.), 'affection for the Nishadh- 
an' ; rdjd samyagvrttah sadd tvayi (MBh.), 'the king always behaved properly 
toward thee'. 

305. The prepositions construed with the locative stand to it only in 
the relation of adverbial elements strengthening and directing its meaning. 

In the Veda, such locative constructions are most frequent with d and 
ddhi: thus, mdrtyesv d, 'among mortals'; prthivydm ddhy osadhih, 'the 
plants upon the earth' ; tejo mdyi dhdrayd 'dhi (AV.), 'establish glory in 
me': less often, upa and dpi are used in the same way. In all ages of 
the language, antdr, 'within, among', is construed with the locative. 

306. Declensional forms are made by the addition of 
endings to the stem, or base of inflection. 

The stem itself, however, in many words and classes 

92 IV. DECLENSION. [306 

of words, is liable to variation, especially as assuming a 
stronger form in some cases and a weaker in others. 

And between stem and ending are sometimes inserted 
connecting elements (or what, in the recorded condition of 
the language, have the aspect of being such). 

Respecting all these points, the details of treatment, as exhibited by 
each class of words or by single words, will be given in the following chapters. 
Here, however, it is desirable also to give a brief general view of them. 

307. Endings: Singular. In the nominative, the 
usual masc. and fern, ending is * which, however, is want- 
ing in derivative a and -stems ; it is also euphonically lost 
(150) by consonant-stems. Neuters in general have no ending, 
but show in this case the bare stem ; a-stems alone add m (as 
in the accus. masc. I Among the pronouns, am is a frequent 
masc. and 1'em. nom. ending (and is found even in du. and pi.); 
and neuters show a form in d. 

In the accusative, m or am is the masc. and fern, end- 
ing am being added after a consonant and r and after i and 
M in the radical division, and m elsewhere after vowels. The 
neuter accusative is like the nominative. 

The instrumental ending for all genders alike is a. 
With final i and u- vowels, the a is variously combined, and in 
the older language it is sometimes lost by contraction with them. 
Stems in a make the case end in ena (sometimes eria in V.), 
and those in a make it end in aya; but instances occur, in the 
early language, of immediate addition of a to both a and a. 

The dative ending is in general e; and with it likewise 
the modes of combination of i and u final are various (and dis- 
appearance by contraction not unknown in the oldest language). 
The a-stems are quite irregular in this case, making it end in 
ay a excepted is the pronominal element snia, which combines 
(apparently] with e to smai. In the personal pronouns is found 
bhyam (or hyam}. 

A fuller ending ai (like gen.-abl. as and loc. am: see 
below) belongs to feminine stems only. It is taken (with inter- 
posed y] by the great class of those in derivative a: also by 
those in derivative i, and (as reckoned in the later language) 
in derivative u. And later it is allowed to be taken by femi- 
nine stems in radical i and M, and even by those in i and u : 
such have it in the earliest language in only rare and excep- 
tional instances. 

The ablative has a special ending, d (or #), only in 




a-stems, masc. and neut., the a being lengthened before it 
(except in the personal pronouns of 1st and 2d person; and 
these have the same ending in the pi., and even, in the old 
language, in the dual). Everywhere else, the ablative is ident- 
ical with the genitive. 

The genitive of -stems (and of one pronominal w-stem, 
amu] adds sya. Elsewhere, the usual abl.-gen. ending is as: 
but its irregularities of treatment in combination "with a stem- 
final are considerable. With i and w, it is either directly added 
(only in the old language), added with interposed n, or fused to es 
and os respectively. With r (or ar) it yields us (or ur : 169, end'. 

The fuller as is taken by feminine stems precisely as ai 
js taken in the dative : see above. 

The locative ending is i in consonant and r and a-stems 
(fusing with a to e in the latter). The i and w-stems (unless the 
final vowel is saved by an interposed n] make the case end in 
au; but the Veda has some relics or traces of the older forms 
(ay-i [?] and av-i] out of which this appears to have sprung. 
Vedic locatives from /-stems end also in a and i. The pronom- 
inal element sma makes the locative smin. Stems in an in the 
older language often lose the *, and use the bare stem as locative. 

The ending am is the locative correspondent to dat. ai and 
abl.-gen. as, and is taken under the same circumstances : see above. 

The vocative (unless by accent: 314) is distinguished 
from the nominative only in the singular, and not quite always 
there. In a-stems, it is the unaltered stem ; and so also in 
most consonant-stems : but neuters in an and in may drop 
the n; and the oldest language has sometimes a vocative in s 
from stems in nt and ns. Stems in r change this to ar. In 
masc. and fern, i and w-stems, the case ends respectively in e and 
o; in neuters, in the same or in i and u. Stems in a change 
a to e; derivative 1 and u are shortened ; radical stems in long 
vowels use the nominative form. 

308. Dual. The dual has except so far as the voca- 
tive is sometimes distinguished from nom. and accus. by a dif- 
ference of accent: 314 only three case-forms : one for nom., 
accus., and voc.; one for instr., dat., and abl.; and one for 
gen. and loc. (occasional confusion of the uses of the second and 
third is seen earlier). 

But the pronouns of 1st and 2d person in the older lan- 
guage distinguish five dual cases : see 492. 

The masc. and fern, ending for nom. -accu s. - voc. is in 
the later language usually au; but instead of this the Veda 
has prevailingly a. Stems in a make the case end in e. Stems in i 
and u, masc. and fern., lengthen those vowels; and derivative i in 

94 IV. DECLENSION. [308 

the Veda remains regularly unchanged, though later it adds au, 
The neuter ending is only *, with final a this combines to e. 

The universal ending for the instr. -dat. -abl. is bhyam, 
before which final a is made long. In the Veda, it is often to 
be read as two syllables, bhiam. 

The universal ending of gen.-loc. is os : before this, a 
and a alike become e (ai). 

309. Plural. In the nominative, the general masc. 
and fern, ending is as. The old language, however, often makes 
the case in asas instead of as from a-stems, and in a few examples 
also from a-stems. From derivative ?-stems, is instead of yas 
is the regular and usual Vedic form. Pronominal a-stems make 
the masc. nom. in e. 

The neuter ending (which is accusative also) is in general i ; 
and before this the final of a stem is apt to be strengthened, by 
prolongation of a vowel, or by insertion of a nasal, or by both. 
But in the Veda the hence resulting forms in ani, mi, uni are 
frequently abbreviated by loss of the m, and sometimes by 
further shortening of the preceding vowel. 

The accusative ending is also as in consonant-stems and 
in the radical division of i and w-stems (and in the old lan- 
guage even elsewhere). Stems in short vowels lengthen those 
vowels and add in the masculine n (for ns, of which abundant 
traces remain), and in the feminine s. In the neuter, this case 
is like the nominative. 

In the instrumental, the case-ending is everywhere bhis 
except in a-stems, where in the later language the case always 
ends in ais, but in the earlier either in ais or the more reg- 
ular ebhis (abhis in the two personal pronouns ; and the prono- 
minal stem a [501] makes ebhis only). 

The dative and ablative have in the plural the same 
form, with the ending bhyas (in Veda often bhias], before which 
only a is altered, becoming e. But the two personal pronouns 
distinguish the two cases, having for the ablative the singular ending 
(as above pointed out), and for the dative the peculiar bhyam (almost 
never in V. bhiam), which they extend also into the singular. 

Of the genitive, the universal ending is am; which 
(except optionally after radical I and M, and in a few scatter- 
ing Vedic instances) takes after final vowels an inserted conso- 
nant, s in the pronominal declension, n elsewhere ; before ri, a 
short vowel is lengthened ; before s, a becomes e. In the Veda, 
it is very frequently to be pronounced in two syllables, as a-am. 

The locative ending is su, without any exceptions, and 
the only change before it is that of a to e. 


The vocative, as in the dual, differs from the nomina- 
tive only by its accent. 

310. The normal scheme of endings, as recognized by 
the native grammarians (and conveniently to be assumed as 
the basis of special descriptions), is this : 




m. f. n. 

m. f. n. 

m. f. n. 


s m 

au i 

as i 



au i 

as i 





















It is taken in bulk by the consonantal stems, and by the 
radical division of ~i and w-stems ; by other vowel-stems, with 
more or less considerable variations and modifications. The 
endings which have almost or quite unbroken range, through 
stems of all classes, are bhyam and os of the dual, and b/iis, 
am, and su of the plural. 

311. Variation of Stem. By far the most import- 
ant matter under this head is the distinction made in large 
classes of words (chiefly those ending in consonants) between 
strong and weak stem-forms a distinction standing in 
evident connection with the phenomena of accent. In the 
nom. and accus. sing, and du. and the nom. pi. (the five 
cases whose endings are never accented: 316 a), the stem 
often has a stronger or fuller form than in the rest : thus, 
for example : ^IsTR^ raj an- am, {Ni4l rajan-au, {MMU raf- 
an-as } against ^T^T rajn-a and ^TsTPTCT raja-bkis ; or H^HM 
mahant-am and ^W^tudant-am against H<HI mahat-a and 
cT^TTT tudat-a. These five, therefore, are called the cases 
with strong stem, or, briefly, the strong cases; and the rest 
are called the cases with weak stem, or the weak cases. 
And the weak cases, again, are in some classes of words to be 
distinguished into cases of weakest stem, or weakest cas- 
es, and cases of middle stem, or middle cases : the former 

95 IV. DECLENSION. [311 

having endings beginning with a vowel (instr. to loc. sing., 
gen.-loc. du., gen. pi.); the latter, with a consonant (inst.- 
dat.-abl. du., instr., dat.-abl., and loc. pi.). 

The class of strong cases, as above denned, belongs 
only to masculine and feminine stems. In neuter inflection, 
the only strong cases are the nom.-acc. pi.; while, in those 
stems that make a distinction of weakest and middle form, 
the nom.-acc. du. belong to the weakest class, and the same 
cases sing, to the middle : thus, for example, compare SJHlRl 
pratyafte-i, nom.-acc. pi. neut., and 3rU ^i H^ pralyanc-as> 
nom. pi. masc.; yrfWl prafic-1, nom.-acc. du. neut., and 
SfcfNrTCT pratw-os, gen.-loc. du. ; HC^I pratyak, nom.-acc. 
sing, neut., and H rU N ^^pratyag- bhis, instr. pi. 

Even in words which exhibit no variation of stem, it is often con- 
venient to distinguish the same groups of cases by the names strong and 
weak and so on. 

312. Other variations concern chiefly the final vowel of a stem, and may 
be mainly left to be pointed out in detail below. Of consequence enough 
to mention here is only the puna-strengthening of a final i or u, which in 
the later language is always made before as of nom. pi. and e of dat. sing. 
in rnasc. and fern.; in the Veda, it does not always take place; nor is it 
forbidden in dat. sing. neut. also; and it is seen sometimes in loc. sing. 
Final r has ywraa-strengthening in loc. sing. 

313. Insertions between Stem and E nding. After vowel-stems, 
an added n often makes its appearance before an ending. This appendage 
is of least questionable origin in nom.-acc. pi. neut., where the interchange 
in the old language of the forms of a and i-stems with those of an and in- 
stems is pretty complete ; and the w-stems follow their analogy. Elsewhere, 
it is most widely and firmly established in the gen. pi., where in the great 
mass of cases, and from the earliest period, the ending is virtually nam after 
a vowel. In the i and M-stems of the later language, the instr. sing, of 
masc. and neut. is separated by its presence from the fern., and it is in the 
other weakest cases made a usual distinction of neuter forms from mas- 
culine; but the aspect of the matter in the Veda is very different; there 
the appearance of the n is everywhere sporadic ; the neuter shows no special 
inclination to take it, and it is not excluded even from the feminine. In the. 
ending ena from u-stems (later invariable, earlier predominating), its presence 
appears to have worked the most considerable transformation of original shape. 

The place of n before gen. pi. am is taken by s in pronominal a and a-stems. 
The y after a before the endings di, as, and dm is most probably an 
insertion, such as is made elsewhere (258). 


Accent in Declension. 

314. As a rule without exception, the vocative, if accented 
at all, is accented on the first syllable. 

And in the Veda (the case is a rare one), whenever a syllable written 
as one is to be pronounced as two by restoration of a semivowel to vowel 
form, the first element only has the vocative accent, and the syllable as 
written is circumflex (by 84 a): thus, dyaiis (i. e. dfihis] when dissyllabic, 
but dyatis when monosyllabic ; jykke when for jfake. 

But the vocative is accented only when it stands at the 
beginning of a sentence or, in verse, at the beginning also 
of a metrical division or pada; elsewhere it is accentless or 
enclitic : thus, dgne yam yajn&m paribhur dsi (RV.), 'O Agni ! 
whatever offering thou protectest 1 ; but upa tva 'gna e 'masi (RV.), 
'unto thee, Agni, we come'. 

A word qualifying a vocative usually an adjective, but not seldom also a 
noun in the genitive (very rarely in any other case) constitutes, so far as 
accent is concerned, a unity with it : thus, sdkhe vaso or vdso sakhe, 'excellent 
friend'; sUno sahasah or sdhasdh suno, 'oh son of might'; and suditi suno 
sahaso didihi (RV.), 'with excellent brightness, son of might, shine forth'. 

Two coordinate vocatives, whether noun or adjective, have usually the same 
accent ; but the Vedic texts furnish not a few irregular exceptions to this rule. 

For brevity, the vocative dual and plural will be given in the paradigms 
below along with the nominative, without taking the trouble to specify in 
each instance that, if the latter be accented elsewhere than on the first 
syllable, the accent of the vocative is different. 

315. As regards the other cases, rules for change of accent 
in declension have to do only with monosyllables and with stems 
of more than one syllable which are accented on the final ; for if a 
stem be accented on the penult, or any other syllable further 
back as in sdrpant, vari, bhagavant, sumdnas, sa/idsravaja 
the accent remains upon that syllable through the whole inflection 
(except in the vocative, as explained in the preceding paragraph). 

The only exceptions are a few numeral stems : see below, chap. VI. 

316. Stems accented on the final (including monosyllables) 
are subject to variation of accent in declension chiefly in virtue 
of the fact that some of the endings have, while others have 
not, or have in less degree, a tendency to draw the accent for- 
ward upon themselves. Thus : 

a. The endings of the nominative and accusative singular and dual and 
of the nominative plural have no tendency to take the accent away from the 
stem, and are therefore only accented when a final vowel of the stem and 
the vowel of the ending are blended together into a single vowel or diphthong. 
Thus, from dattd come dattm't (= dattd -f- au) and dnttas (= datta -\- as); 
but from nadt come nadydu (=nadi-\-au) and nadytt* (==nadt-}-aa). 
Whitney, Grammar. 7 

98 IV. DECLENSION. [316 

b. All the other endings sometimes take the accent; but those beginning 
with a vowel do so more readily than those beginning with a consonant. 
Thus, from nans come navti and naubhis; from mahant, however, come mahaid 
but mahddbhis. 

The general rules of accent, then, may be thus stated : 

317. In the declension of monosyllabic stems, the accent 
falls upon the ending in all the weak cases (without distinction 
of middle and weakest) : thus, nava, naubhyhm, navam, nausu ; 
vaci, vagbhis, vactim, vaksu. 

But some monosyllabic stems retain the accent throughout : thus, ptfbhis, 
gdvam, g6su. For such cases, see below, 350, 361 c, d, 375, 390, 427. 

318. Of polysyllables ending in consonants, only a few 
shift the accent to the ending, and that in the weakest (not the 
middle) cases. Such are : 

a. Present participles in ant or ill : thus, from tuddnt, tudatd and tudattis 
and tudatam; but tudddbhydm and tudatsu. 

b. A few adjectives having the form of such participles, as mahatd, 

c. Bases of which the accented final loses its syllabic character by syn- 
copation of the vowel : thus, majjnd, murdhne, ddmnd* (from majydn etc. : 423 . 

Other sporadic cases will be noticed under the different declensions. 
Case forms used adverbially sometimes show a changed accent: see 
chap. XVI. (lllOff.). 

319. Of polysyllabic stems ending in accented short vowels, 
the final of the stem retains the accent if it retains its syllabic 
identity : thus, dattena and datthya from dattd ; agnina and agndye 
from agni; and also dattebhyas, agnibhis, and so on. Otherwise, 
the accent is on the ending : and that, whether the final and 
the ending are combined into one, as in dattais, dhenau, agnln, 
dhenus, and so on ; or whether the final is changed into a semi- 
vowel before the ending : thus, dhenvu, pitrh. 

But am of the gen. pi. from stems in / and ti and r may, and in the 
older language always does, take the accent, though separated by n from the 
stem: thus, agnlnnm, dheniindm, pitfndm. In RV., even derivative i-stems 
show usually the same shift : thus, bahvindm. Of stems in a, only numerals 
(chap. VI.) follow this rule: thus, saptandm, dafdndm. 

320. Root-words in I and u as final members of compounds retain the 
accent throughout, not shifting it to any of the endings. And in the older 
language there are polysyllabic words in long final vowels which follow in 
this respect as in others the analogy of the root-declension (below, 355 ff.). 
Apart from these, the treatment of stems in derivative long vowels is, as 
regards accent, the same as of those in short vowels save that the tone 
is not thrown forward upon the ending in gen. plural. 

323] 99 



321. THE accordance in inflection of substantive and 
adjective stems is so complete that the two cannot be sep- 
arated in treatment from one another. 

They may be classified, for convenience of description, 
as follows: 

I. Stems in f a; 

II. Stems in ^ i and 3 u; 

III. Stems in ETT a, *> and "37 u : namely, A. radical- 
stems (and a few others inflected like them); B. derivative 
stems ; 

IV. Stems in ft r (or q[ or); 

V. Stems in consonants. 

There is nothing absolute in this classification and arrangement ; it is 
merely believed to be open to as few objections as any other. No general 
agreement has been reached among scholars as to the number and order 
of Sanskrit declensions. The stems in a are here treated first because of 
the great predominance of the class. 

322. The division-line between substantive and adjective, 
always an uncertain one in early Indo-European language, is 
even more wavering in Sanskrit than elsewhere. There are, 
however, in all the declensions as divided above unless we 
except the stems in r or ar - - words which are distinctly ad- 
jectives ; and, in general, they are inflected precisely like noun- 
stems of the same final : only, among consonant-stems, there are 
certain sub-classes of adjective bases with peculiarities of in- 
flection to which there is among nouns nothing corresponding. 
But there are also two considerable classes of adjective-com- 
pounds, requiring special notice ; namely : 

323. Compound adjectives having as final member a bare 
verbal root, with the value of a present participle : thus, su-dfc, 
'well-looking'; pra-budh, 'foreknowing ; a-druh, 'not hating'; veda- 
vid, 'Veda-knowing'; vrtm-han, 'Vritra-slaying'; upastha-sdd, 
'sitting in the lap'. Every root is liable to be used in this 



way, and such compounds are not infrequent in all ages of the 
language : see chapter on Compounds, below (XVIII.). 

This class is originally and essentially only a special class of compound 
adjectives, since in the earliest Veda the simple as well as the compounded 
root was sometimes used adjectively. But the compounded root was from the 
beginning much more often so used, and the later the more exclusively, 
so that practically the class is a separate and important one. 

324. Compound adjectives having a noun as final member, 
but obtaining an adjective sense secondarily, by having the idea 
of 'possession' added, and being inflected as adjectives in the 
three genders. Thus, prajakama, 'desire of progeny', becomes 
an adjective meaning 'desirous (i. e. having desire) of progeny'; 
sabharya (sa-\-b/iarya), 'having one's wife along'; and so on. 

In a few cases, also, the final noun is syntactically object of the pre- 
ceding member: thus, atimatra, 'immoderate' (at i matram, 'beyond measure'); 
yavayaddvesas, 'driving away enemies'. 

325. Hence, under each declension, we have to notice how 
a root or a noun-stem of that declension is inflected when final 
member of an adjective compound. 

As to accent, it needs only to be remarked that a mono- 
syllabic word ending a compound loses the peculiarity of mo- 
nosyllabic accentuation, and does not throw the tone foward upon 
the ending. 

Declension I. 

Sterns (masculine and neuter) in a. 

326. This declension contains the majority of all the 
declined stems of the language. 

Its endings deviate more widely than any others from 
the normal. 

327. Endings: Singular. The nom. masc. has the normal end- 
ing .<t. 

The ace. (masc. and neut.) adds m (not am); and this form has the 
office also of nom, neuter. 

The instr. changes a to ena uniformly in the later language; and even 
in the oldest Vedic this is the predominant ending (in RV., eight ninths of 
all cases). Its final is in Vedic verse not infrequently made long (ena), where 
favored by the metre. But the normal ending a thus, yajnd, suhdva, 
mahitvd (for yajnena etc.) is also not rare in the Veda. 

The dat. has aya (as if by adding aya to a), alike in all ages of the 

The abl. has t (or more probably d: it is impossible from the evidence 


of the Sanskrit to tell which is the original form of the ending), before which 
a is made long: this ending is found in no other noun declension, but only 
in the personal pronouns (of all numbers). 

The gen. has sya added to the final a; and this ending is also limited 
to a-stems (with the single exception of the pronoun amusya : chap. VII.). Its 
final a is in only three cases made long in the Veda; and its y is vocalized 
(asia) almost as rarely. 

The loc. ends in e (as if by combining the normal ending i with the 
final of the stem), without exception. 

The voc. is the bare stem. 

328. Dual. The dual endings in general are the normal ones. 

The nom., ace., arid voc. masc. end in the later language always in au. 
In the Veda, however, the usual ending is simple a (in RV., in seven 
eighths of the occurrences). The same cases in the neut. end in e, which 
appears to be the result of fusion of the stem-final with the normal ending i. 

The instr., dat., and abl. have bhyam (in only one or two Vedic in- 
stances resolved into 6/u'am), with the stem-final lengthened to a before it. 

The gen. and loc. have a y inserted after the stem-final before os (or 
as if the a had been changed to e}. In one or two (doubtful) Vedic in- 
stances (as also in the pronominal forms mos and yos), os is substituted for 
the final a. 

329. Plural. The nom. masc. has in the later language the normal 
ending as combined with the final a to as. But in the Veda the ending 
asas instead is frequent (one third of the occurrences in RV., but only one 
twenty-fifth in the peculiar parts of AV.). 

The ace. masc. ends in an (for earlier ans, of which abundant traces 
are left in the Veda, arid, under the disguise of apparent euphonic com- 
bination, even in the later language: see above, 208 ff.). 

The nom. and ace. neut. have in the later language" always the ending 
ani (like the an-stems: see 421; or else with n as in the gen. pi. before 
normal i). But in the Veda this ending alternates with simple a (which in 
RV. is to ani as three to two, in point of frequency; in AV., as three 
to four). 

The instr. ends later always in ais ; but in the Veda is found abund- 
antly the more normal form ebhis (in RV., nearly as frequent as aj's; in 
AV., only one fifth as frequent). 

The dat. and abl. have bhyas as ending, with e instead of the final a 
before it (as in the Vedic instr. ebhis, the loc. pi., the gen. loc. du. [?], and 
the instr. sing.). The resolution into ebhias is not infrequent in the Veda. 

The gen. ends in anora, the final a being lengthened and having n 
inserted before the normal ending. The a of the ending is not seldom (in 
less than half the instances) to be read as two syllables, aam: opinions are 
divided as to whether the resolution is historical or metrical only. A very 
small number (half-a-dozen) of examples of simple am as ending instead of 
anam occur in RV. 




The loc. ends in em that is to say, with the normal ending, before 
which the stem-final is changed to e (with consequent change of s to s: 180). 

Of accent in this declension, nothing requires to be said ; 
the syllable accented in the stem retains its own accent through- 

330. Examples of declension. As examples of 
the inflection of a-stems may be taken 37R kama, m., 'love'; 
~%3 devd, m. 7 'god'; EfTHT asya, n., 'mouth'. 

Singular: s" 








N.A. V. 

I. D. Ab. 







of-) |J-| 



Plural : 













kamau devau 

^iT^^UH ^TPTTT 1 ? 

kamabhyam devabhyam 





*%( H4 t-tj 








334 DECLENSION I., a-STEMs. 103 


kaman devfin asyani 

sRTifc^ ~|t^ SCTHT 

kamais devais asyms 

D. Ab. 

kamebhyas devebhyas asyebhyus 


, r _ , 

kamanam devanam asanam 

L. 37m 

kamesu devesu asyesu 

Examples of the peculiar Vedic forms are: 

Sing. : instr. ravdthena, yajnd (such genitive forms as dfvasia are purely 

Du. : nom. etc. masc. devd; gen.-loc. pastybs (stem pastyh}. 

PI.: nom.-voc. masc^ devdsas; neut. yugd; instr. devebhis; gen. cara- 
(/tam, devanaam. 

331. Among nouns, there are no irregularities in this de- 
clension. For irregular numeral bases in a (or an], see the 
next chapter. For the irregularities of pronominal stems in a, 
which are more or less fully shared also by a few adjectives of 
pronominal kindred, see the chapter on Pronouns. 


332. Original adjectives in a are an exceedingly large 
class, probably the majority of all adjectives. There is, however, 
no such thing as a feminine stem in a; for the feminine, the a 
is changed to a - - or often, though far less often, to i ; and its 
declension is then like that of sena or devi (365). An example 
of the complete declension of an adjective a-stem in the three 
genders will be given below (371). 

333. There are no verbal roots ending in a. But a is some- 
times substituted for the final a of a root (and, more rarely, 
for final an or am), and it is then inflected like an ordinary 
adjective in a (see below, 354). 

334. A noun ending in a, when occurring as final mem- 
ber of an adjective compound, is inflected like an original ad- 
jective in a, making its feminine likewise in a or . 

On the other hand, a feminine noun ending in derivative 
a shortens its final to a to form a masculine and neuter base. 


Declension II. 

Stems (of all genders) in * and u. 

335. The steins in ^ i and 3 u are inflected in so close 
accordance with one another that they cannot be divided 
into two separate declensions. They are of all the three 
genders, and tolerably numerous those in ^ t more nume- 
rous than those in 3 u, especially in the feminine (there are 
more neuters in 3 u than in ^ i) . 

The endings of this declension also differ frequently and 
widely from the normal, and the Vedic irregularities are nu- 

336. Endings: Singular. The nom. masc. and fern, adds to the 
stem the normal ending s. The nom. and ace. neut. is the bare stem, 
without ending. In the Veda, the final u of a few neuters is lengthened 
(248 b): thus, iwfi, puri. 

The ace. masc. and fern, adds m to the stem. Vedic forms in iam and 
warn, and, with n, inam and wnam, are excessively rare, and doubtful. 

The instr. fern, in the later language takes the normal ending a simply, 
while the masc. and neut. insert n before it, making ina and una. But in 
the Veda, forms in ya and va (or ia and ua] are not infrequent in masc. 
and neut. also ; while ina is found, very rarely, as a fern, ending. Moreover, 
fern, ya is often (in two thirds of the occurrences) contracted to I; and this 
is even sometimes shortened to i. An adverbial instr. in uyd from half-a- 
dozen stems in u occurs. 

The dat. masc. and fern, gunates the flnal of the stem before the end- 
ing e, making .aye and ave. These are the prevailing endings in the Veda 
likewise; but the more normal ye and ve also occur; and the fern, has in 
this case, as in the instr., sometimes the contracted form I. In the later 
language, the neuter is required in this, as in all the other "weakest" cases, 
to insert n before the normal ending : but in the Veda such forms are only 
sporadic; and the neut. dat. has also the forms we, ave, aye, like the other 

The abl. and gen. masc. and fern, have regularly, both earlier and later, 
the ending s with gunated vowel before it: thus, a, os; and in the Veda, 
the neut. forms the cases in the same way; although wnas, required later, 
is also not infrequent (mas does not occur). But the normal forms yas (or 
ias) and vas (or uas) are also frequent in both masc. and neut. As masc. 
ending, unas occurs twice in RV. 

The loc. masc. and fern, has for regular ending in the later language 
5ti, replacing both finals, i and u. And this is in the Veda also the most 
frequent ending; but, beside it, the i-stems form (about half as often in 
RV.) their loc. in a: thus, agna; and this is found once even in the neut. 


The RV. has a number of examples of iriasc. and neut. locatives in avi 
(the normal ending and the u gunated before it) from w-stems; and certain 
doubtful traces of a corresponding ayi from i-steins. Half-a-dozen locatives 
in i (regarded by the Vedic grammarians as prayrhya or uncombinable : 138 d) 
are made from /-stems. The later language requires the neuter locatives to 
be made mi and uni ; but the former never occurs in the oldest texts, and 
the latter only very rarely. 

The later grammar allows thedat., abl.-gen., and loc. fern, to be formed 
at will with the fuller fern, terminations of long-vowel stems, namely ai, as, 
am. Such forms are quite rare in the older language even from i-stems 
(less than 40 occurrences altogether in RV. ; three times as many in AV.); 
and from w-stems they are almost unknown (live in RV. and AV.). 

The-voc. gunates the final of the stem, in masc. and fern., alike in 
the earlier and in the later language. In the neut., it is later allowed to 
be of the same form or the unaltered stern : and this was probably the usage 
in the older time also ; not instances enough are quotable to determine the 
question (AV. has u once, and VS. o once). 

337. Dual. The later and earlier language agree in making the nom.- 
acc.-voc. masc. and fern, by lengthening the final of the stem. The same 
cases in the neuter (according to the rule given above) end later in ini and 
uni; but these endings are nearly unknown in the Veda (as, indeed, the 
cases are of only rare occurrence): AV. has inl twice (RV. perhaps once); 
VS. has uni once; RV. has ul from one M-stem, and I, once shortened to i, 
from one or two i-stems. 

The unvarying ending of instr.-dat.-abl., in all genders, is bhyam added 
to the unchanged stem. 

The gen. -loc. of all ages adds os to the stem in masc. and fern. ; in 
neut., the later language interposes, as elsewhere in the weakest cases, a n; 
probably in the earlier Vedic the form would be like that of the other gen- 
ders; but the only occurrence noted is one unos in AV. 

338. Plural. The nom.-voc. masc. and fern, adds the normal end- 
ing as to the gunated stem-final, making ayas and avas. The exceptions 
in the Veda are very few: one word (ari) has ias in both genders, and a 
few feminines have Is (like i-stems); a very few w-stems have uas. The 
neut. nom.-acc. ends later in Ini and uni (like ani from a: 329); but the 
Veda has I and i (about equally frequent) much oftener than mi; and u 
and (more usually) w, more than half as often as uni. 

The accus. masc. ends in In and tin, for older Ins and uns, of which 
plain traces remain in the Veda in nearly half the instances of occurrence, 
and even not infrequently in the later language, in the guise of phonetic 
combination (208 ff.). The accus. fem. ends in is and us. But both masc. 
and fem. forms in ias and uas are found sparingly in the Veda. 

The in st. of all genders adds bhis to the stem. 

The dat.-abl. of all genders adds bhyas (in V., almost never bhias] to the stem. 

The gen. of all genders is made alike in inam and unam (of which the 




a is not seldom, in the Veda, to be resolved into aam}. Stems with accented 
final in the later language may, and in the earlier always do, throw forward 
the accent upon the ending. 

The loc. of all genders adds su (as su: 180) to the stem-final. 

The accent is in accordance with the general rules already 
laid down, and there are no irregularities calling for special 

339. Examples of declension. As models of 
/-stems may be taken 5|ftl agni, m., 'fire'; irf^ gdti, ., 
'gait'; snff vdri, n., 'water'. 

Singular: / 

N. jyiUH 


agnim gdtim 

agnina gdtya wanna 

Ab. G. 


gdtaye,gdtyai varine 










gdtau, gdtyam 
















Plural : 







i. rfiiftH TTWHH snfrPm 

, s "^ f . . "^ 

agnibhis gdtibhis varibhis 

D.Ab. tlfijHJ^ rrfrRiq^ onf^HJ^ 

agnibhyas gdtibhyas varibhyas 

G. tiiJl-nH^ JIrffaF[ ^{ImiH^ 

agninhm gdtlnam varmam 

L. srftig JlfcFJ ^T3 

agnisu gdtisu varisu 

340. In order to mark more plainly the absence in Vedic language of 
some of the forms which are common later, all the forms of Vedic occurrence 
are added below, and in the order of their frequency. 

Singular. Norn, agnts etc., as above. 

Ace.: masc. agnfin, yayiam, tmn/nam(?); fern, and neut. as above. 

Instr. : masc. agnma, rayyd and urmid; fern, dcitti, utid, anuvrktf, 
dhasfna; neut. wanting. 

Dat. : masc. agndye ; fern, tujaye, uti, prutyai ; neut. f ucaye. 

Gen.-abl. : masc. agues; fern, ddites, hetyds and 6/mmias; neut. bhures. 

Loc. : masc. agnau, agnd, ajdyiC?)-, fern, dgatau, udita, dhdnasatayiC?), 
vedl, bhumyam; neut. apratd, saptdra$mau. 

Voc. : as above (neut. wanting). 

Dual: Nom.-acc.-voc.: masc. harl; fem.yuvatt; neut. tici, md/,/tdrmt(?). 

Instr.-dat.-abl. : as above. 

Gen.-loc. : masc. hdrios; fern, yuvatyos and jamios; neut. wanting. 

Plural. Nom.: masc. agndyas; fern, matdyas, bhumis; neut. {;wct, 
bhuri, bhurlni. 

Accus. : masc. agriin; fern, ksitis, ftica7/a(?). 

Instr., dat.-abl., and loc., as above. 

Gen. : masc. kavinam, rslnaam etc. 

341. As models of w-stems may be taken ST5T ?dtru, 
m., 'enemy'; R dhenu, f., 'cow'; JTJ mddhu, n., 'honey'. 


N. ST^H ^H TO 

O N. O *x O 

fdtrus dhenus mddhu 

A. 5T^J7 ^T^T TO 

O "s O *S O 

cdtrum dhenum mddhu 


f o 

truna dhenva mddhuna 





Ab. G. 



Dual : 




Plural : 




D. Ab. 




L - 

dhendve , dhenvai 
dhenos , dhenvas 



cdtrubhyam dhenubhyam 



ex *> 





















o % 


ex *x 


342. The forms of Vedic occurrence are given here for the w-stems in 
the same manner as for the i-stems above. 

Singular. Norn.: masc. and fern, as above; neut. wrw, ww. 

Accus. : masc. ketum, dbhiruam, sucetunam^} ; fern, dhenum. 

Instr. : masc. ketiina, papva and krdtud; fern, ddhenua and panua, aput/a; 
neut. mddhuna, mddhva. 

Dat. : masc. fecfave, pipve; fern, pdrave, isvai; neut. urduc, mddhune. 

Abl.-gen. : masc. manyos, pit-was, saramas; fern, sfndhos, isvds; ueut. 
mddhvas and rnddhuas, inddhos, mddhunas. 


Loo.: masc. purau, sunavi fern, sfndhiiu, rdjjvarn; neut. sanita, .si/mir/. 
adno, sflnuni. 

Voc. : as above. 

Dual. Nom.-acc.-voc. : masc. and fern, as above; neut. wrvf, jdnunl. 

Instr.-dat.-abl. : as above. 

Gen.-loc. : as above (but -vos or -uos). 

Plural. Nom. : masc. rbhtivas, mddhuas and nu'idhvas; fern, dhendvas, 
futfikratvas ; neut. puruni, pnrn, puru. 

Accus. : masc. rfwn, pafvds ; fern, fsus, mddhvas. 

Instr., dat.-abl., and loc., as above; also gen. (but with the resolution 
-unanm in part). 

343. Irregular declension. There are no irregular 
?/-stems, and only a very few /-stems. 

a. S&khij m., 'friend', has for the five strong cases a pecu- 
liarly strengthened base (vriddhied), namely sakhay, which in the 
nom. sing, is reduced to sdkha (without ending), and in the 
other cases takes the normal endings. The instr. and dat. sing, 
have the normal endings simply, without inserted n or guna; 
the abl.-gen. sing, adds us; and the loc. sing, adds au : the 
rest is like agni. Thus : 

Sing, safcfta, sdkhayam, stikhya, sdkhye, sdkhyus, sdkhyau, sdkhe.; Du. 
adkhayau, sakhibhyam, sdkhyos ; PI. sakhayas, stiktnn, etc. etc. 

The Veda has usually sdkhaya du., and often resolves the y to ?, in 
sfikhia, sakhius, etc. The compounds are usually declined like the simple 
word, unless sakha be substituted. 

b. Pad, m., is declined regularly in composition, and when it 
has the meaning 'lord, master' ; when uncompounded and mean- 
ing 'husband', it is inflected like sdkhi in the instr., abl.-gen., 
and loc. sing., forming patya, petty e, patyus, pdtyau. There are 
occasional instances of confusion of the two classes of forms. 

c. Jam', f., 'wife', has the gen. sing, j&nyus in the Veda. 

d. Ari, 'eager, greedy, hostile', has in the Veda aryds in pi. nom. and 
accus., masc. and fern. 

e. W, 'bird', has in RV. the nom. ves (beside vfa). 

f. The stems dksi, 'eye', dsthi, 'bone', dddhi, 'curds', and sdkthi, 'thigh', 
are defective, their forms exchanging with and complementing forms from 
bases in an (aksdn etc.): see the bases in an, below (431). 

g. The stem pathf, 'road', is used to make up part of the inflection of 
pdnthan : see below, 433. 

h. Kr6stu, m., 'jackal', lacks the strong cases, for which the corresponding 
forms of krostr are substituted. 


344. Original adjective stems in i are few; those in u are 
much more numerous (many derivative verb-stems forming a 


participial adjective in u}. Their inflection is like that of nouns, 
and has been included in the rules given above ; the stem is in 
general the same in all the three genders. In those weak cases, 
however namely, the dat., abl.-gen., and loc. sing., and 
the gen. -loc. dual in which neuter nouns differ from mas- 
culines in the later language by an inserted n (we have seen 
above that this difference does not exist in the Veda), the neuter 
adjective is allowed to take either form. 

But adjectives in u preceded by one consonant sometimes 
form a derivative feminine stem by adding i: thus, bahvi, urvi, 
prthvl,, vibhvi, and so on. More rarely, the u is prolonged to u 
to make a feminine-stem, which is then inflected like vadhu 
(below, 365). Some adjectives form their feminine in two of 
these ways, or even in all the three : thus, blbhatsu and blbhatsu ; 
tanu, tank, and tanvi. 

345. Roots ending in t or u (or. r : 380) regularly add a 
t when used as root-words or as root-finals of compounds ; and 
hence there are no adjectives of the root-class in this declension. 

Yet, in the Veda, a few words ending in a short radical u are declined 
as if this were sufflxal: thus, dsmrtadhru, sustu; and the AV. has prtanajf 
(once). Roots in u sometimes also shorten u to u : thus, prdbhu, vibhu, etc. 
(354); go (361) becomes gu in composition; and re perhaps becomes ri 
(362); while roots in a sometimes apparently weaken a to i (in -dhi from ydha). 

346. Compound adjectives having nouns of this declension 
as final member are inflected like original adjectives of the same 

Declension III. 

Stems in long vowels: a, 1, u. 

347. The stems ending in long vowels fall into two 
well-marked classes or divisions: A. monosyllabic stems - 
mostly bare roots and their compounds, with a compar- 
atively small number of others inflected like them; B. de- 
rivative feminine stems in Eft a and ^ , with a small num- 
ber in 3T u which in the later language have come to be 
inflected like them. The latter division is by far the larger 
and more important, since most feminine adjectives, and 
considerable classes of feminine nouns, ending in 5TF a or 
^ , belong to it. 




A. Root-words, and those inflected like them. 

348. The inflection of these stems is by the normal 
endings throughout, or in the manner of consonant-stems 
(with ^f am, not ^ m,' in the accus. sing.); peculiarities 
like those of the other vowel-declensions are wanting. The 
simple words are, as nouns, with few exceptions feminine; 
as adjectives (rarely), and in adjective compounds, they are 
alike in masculine and feminine forms. They may, for con- 
venience of description, be divided into the following sub- 
classes : 

1. Root-words, or monosyllables having the aspect of such. 
Those in a are so rare that it is hardly possible to make up a 
whole scheme of forms in actual use ; those in i and u are more 
numerous, but still very few. 

2. Compounds having such words, or other roots with long 
final vowels, as last member. 

3. Polysyllabic words, of various origin and character, in- 
cluding in the Veda many which later are transferred to other 

4. As an appendix to this class we may most conveniently 
describe the half-dozen bases, mostly of irregular inflection, 
ending in diphthongs. 

349. Monosyllabic stems. Before the endings begin- 
ning with vowels, final i is changed to iy and u to uv ; while 
final a is dropped altogether, except in the strong cases and in 
the ace. pi., which is like the nominative (according to the 
grammarians, a is lost here also : no instances of the occurrence 
of such a form appear to be quotable). Stems in i and u are 
in the later language allowed to take optionally the fuller end- 
ings at, as, am in the singular (dat., abl.-gen., loc.); but no 
such forms are ever met with in the Veda (except Wiiyai [?], 
RV., once). Before am of gen. pi., n may or may not be in- 
serted ; in the Veda it is regularly inserted, with a single ex- 
ception (dhiyhm, once). The vocative is like the nominative in 
the singular as well as the other numbers ; but instances of its 
occurrence in uncompounded stems are not found in the Veda, 
and must be extremely rare everywhere. The earlier Vedic dual 
ending is a instead of au. 

350. To the i and w-stems, the rules for monosyllabic 
accent apply : the accent is thrown forward upon the endings 
in all the weak cases except the accus. pi., which is like the 




nom. But the a-stems appear (the instances are extremely few) 
to keep the accent upon the stem throughout. 

351. Examples of declension. As models of mo- 
nosyllabic inflection we may take sfT/#, f., 'progeny'; sft 
dM, f.. 'thought'; and ?\ bhfy f., 'earth'. 

The first of these is rather arbitrarily extended from the four cases 
which actually occur: of the loc. sing, and gen. etc. du., no Vedic examples 
of a-stems are found. 

Singular : 





I>. it 


Ab.o. sra^ 


L. |% 



N. A. V. 

I. D. Ab. 


f f 




Plural : 




fas, Jas 





o -\ 


/ - "^^T^ 

dhls bhus 



dhiya bhuvh 

Rnr, fim H^, ^ 

c?Azy<?, dhiyai bhuve, bhuvai 

^ ,**. o -\ o -s 

dhiyas, dhiyas bhuvas, bhuvas 

dhiyi, dhiyam bhuvi, bhuvam 


c\ *\ f 






1 FC(TT 

C\ "^ 



, ~x x 

jabhis dhibhis 

D.Ab. sfFTO tfr-TTR 

f "** "\ 

jabhyas dhibhyds 

G. STHT^ isTFT?) 

I _-\ >I -s . , _ _, 

janam, Jam dhiyam, dhmam bhmam, bhunam 

L. arm uta w 

o c\o 




352. Monosyllabic stems in composition. When 
the nouns above described occur as final member of a compound, 
or when any root in a or i or u is found in a like position, 
the inflection of an -stem is as above. But 1 and w-stems 
follow a divided usage : the final vowel before a vowel-ending 
is either converted into a short vowel and semivowel (iy or uv, 
as above) or into a semivowel simply y or v). The accent is 
nowhere thrown forward upon the endings ; and therefore, when 
i and u become y and v, the resulting syllable is circumflex. 
Thus : 






Singular : 
N. V. 

Ab. G. 


N. A. V. -dhtyau 
I. D. Ab. 
G. L. -dhfyo 

N. A. V. 
D. Ab. 






















-bhuvos -bhvbs 

-bhtivas -bhvhs 



. , , _ 


, ^ 

-dhfyas -dhyhs 


\-dhiyam , 


L. -dhim 

As to the admissibility of the fuller endings ai, as, and am in the 
singular (feminine), grammatical authorities are somewhat at variance; but 
they are never found in the Veda, and have been omitted from the above 
scheme as probably unreal. 

If two consomants precede the final I or u, the dissyllabic forms, with 
iy and itu, are regularly written ; after one consonant, the usage is varying. 
The grammarians prescribe iy and uv when the monosyllabic stem has more 
Whitney, Grammar. 8 


the character of a noun, and y and v when it is more purely a verbal root 
with participial value. No such distinction, however, is to be seen in the 
Veda where, moreover, the difference of the two forms is only graphic, 
since the ya arid va forms and the rest are always to be read as dissyllabic : 
ia or la and ua or ua, and so on. 

353. A few further Vedic irregularities or peculiarities may be briefly 

Of the a-stems, the forms in as, am, a (du.) are sometimes to be read 
as dissyllables, aas, aam, aa. The dative of the stem used as infinitive is 
in at (as if <!-}-): thus, prakhyaf, pratimaf, paradat. 

Irregular transfer of the accent to the ending in compounds is seen in 
a case or two: thus, avadyabhiyd (RV.), ddhid (AV.). 

354. But compounds of the class above described are not 
infrequently transferred to other modes of inflection : the a 
shortened to a for a masculine stem, or declined like a stem 
of the derivative a-class (below, 365) as feminine ; the i and u 
shortened to * and u, and inflected as of the second declension. 

Thus, compound stems in -ga, -ja, -da, -stha, -bhu, and others, are 
found even in the Veda, and become frequent later; and sporadic cases from 
yet others occur : for example, yrtapdn, vayodhafs and ratnadhebhis, dhanasais 
(all RV.); and, from I and u compounds, vesa^rfs (TS.), dhrayas (RV.), 
ganafrfbhis (RV.), rtanfbhyas (RV.) and senanfbhyas (VS.) and gramanibhis 
(TB.), supuna (AV.), citibhrdve (TS.). Still more numerous are the feminines 
in a which have lost their root-declension : examples are prajd (of which 
the further compounds in part have root-forms), svadhd, fraddhd, pratimd, 
and others. 

355. Polysyllabic Stems. Stems of this division of 
more than one syllable are very rare indeed in the later lang- 
uage, and by no means common in the earlier. The Rig-Veda, 
however, presents a not inconsiderable body of them ; and as 
the class nearly dies out later, by the disuse of its stems or 
their transfer to other modes of declension, it may be best 
described on a Vedic basis. 

a. Of stems in a, masculines, half-a-dozen occur in the Veda: pdntha, 
mdntha, and rbhuksa are otherwise viewed by the later grammar: see below, 
433 4; ufdna (nom. pr.) has the anomalous nom. sing, wpana (and loc. as 
well as dat. ufdne); mahd, 'great 1 , is found only in accus. sing, and abund- 
antly in composition: dta, 'frame', has only dtasu not derivable from dta. 

b. Of stems in i, over seventy are found in the Veda, nearly all femi- 
nines, and all accented on the final. Half of the feminines are formed from 
masculines with change of accent: thus, kalyani (m. kalydna), purusi (m. 
jn/rusa); others show no change of accent: thus, yarn! (m. 7/ama); others still 
have no corresponding masculines : thus, nadi, laksmi, surmi. The masculines 
are about ten in number: for example, rathi, pravi, stari, a AT, apathi. 

356] DECLENSION III. A, RADICAL a-. 1-. AND e^-STEMS. 115 

c. Of stems in u. the number is smaller: these, too, are nearly all 
feminines. and all accented on the final. The majority of them are the 
feminine adjectives in U to masculines in <i or u (above, 344): thus. caranyU, 
carisnu, jighatsu, madhu. A few are nouns in w, with change of accent: 
thus, agru (dgru), prdaku (prdaku), pva^rU (fvtifura); or without change, as 
nrtu. And a few have no corresponding masculines: thus, tanft, radftd, 
cam'U. The masculines are only two or three; namely, prafU, krkadapTi, 
ma/csw(?); and their forms are of the utmost rarity. 

356. The mode of declension of these words may be il- 
lustrated by the following examples : ratM, m., 'charioteer'; nadi, 
f., 'stream'; tann, f., 'body'. 

No one of the selected examples occurs in all the forms : forms for which 
no example at all is quotable are put in brackets. No loc. sing, from any 
I-stem occurs, to determine what the form would be. The stem nadi is se- 
lected as example partly in order to emphasize the difference between the 
earlier language and the later in regard to the words of this division : nadi 
is later the model of derivative inflection. 
Singular : 

N. rathis nadis tanUs 

A. rathiam nadiam tanuam 

I. rathfa nadia tanua 

D. raihie nadte tanue 

Ab. G. rath fas nadfas tani'ias 

L . tanui 

V. rdthi (?) nadi tanu 


N. A. V. rathm nadia tantia 

I. D. Ab. [rathibhyam] nadibhyam [tanubhyam] 

G. L. [rath-Cos] nadios tantios 

Plural : 

N. A. rathias nadfas tantias 

I. [rathibhis] nadibhis tanubhis 

D.Ab. [rathibhyas] nadibhyas tanUbhyas 

G. rathinam nadlnam tanunam 

L. [rathisu] nadisu tanusu. 

The cases nadiam, fant/am, etc. are written above according to 
their true phonetic form, almost invariably belonging to them in the Veda: 
in the written text, of course, the stem-final is made a semivowel, and the 
resulting syllable is circumflexed : thus, nadyhm, tanvam, etc. ; only, as 
usual, after two consonants the resolved forms iy and uv are written instead; 
and also where the combination yv would otherwise result: thus, cakrfya, 
[agruvai,] and mitrayuvas. The RV. really reads starybm etc. twice, and 
tanvhs etc. four times ; and such contractions are more often made in the 
AV. The ending a of the nom.-acc.-voc. du. is the equivalent of the later 
cm. The nom. sing, in s from z-stems is found in the older language about 
sixty times, from over thirty stems. 



357. Irregularities of form, properly so called, are very few in this 
division: eamtZ as loc. sing, (instead of camv"i} occurs a few times; and 
there is another doubtful case or two of the same kind ; the final U is regarded 
as pragrhya or uncombinable (138j; tanui is lengthened to tanv\ in a passage 
or two; -yuvas is once or twice abbreviated to -yUs. 

358. The process of transfer to the other form of I and u-declension 
(below, 363 ff.), which has nearly extinguished this category of words in the 
later language, has its beginnings in the Veda ; but in RV. they are excessive- 
ly scanty: namely, dutidm, loc. sing., once, and puaprucfm, do., once, and 
dravitnud, instr. sing., with two or three other doubtful cases. In the 
Atharvan, wo find the ace. sing, kuh&m, tanUm, vadhUm; the instr. sing. 
palalid and one or two others; the dat. sing, vadhvdf, fvafruai, ayrtivai; 
the abl.-gen. sing, punarbhuvas, prdakuds, fuaf ruds , and the loc. sing. 
tanuam (with anomalous accent). The accusatives plural in Is and us are 
nowhere met with. 

359. Adjective compounds from these words are very few; those which 
occur are declined like the simple stems: thus, hiranyavcifis and sahdsrastarii, 
dtaptatanus and sdrvatanus, all nom. sing, masculine. 

Stems ending in diphthongs. 

360. There are certain monosyllabic stems ending in diph- 
thongs, which are too few and too diverse in inflection to make 
a declension of, and which may be most appropriately disposed 
of here, in connection with the stems in i and u, with which 
they have most affinity. They are : 

stems in an : nau and glau ; 

stems in ai: rat; 

stems in o: go and dyo (or dyu, div}. 

361. a. The stem nau, f., 'ship', is entirely regular, taking 
the normal endings throughout, and following the rules for 
monosyllabic accentuation (317) except that the accus. pi. 
is said (it does not appear to occur in accented texts) to be 
like the nom. Thus: naus, navam, nava, nave, navds, navi; 
navau, naubhyam, navos ; novas, navas, naubhis, naub/iyds, navam, 

The stem glati, m., 'ball', is apparently inflected in the same way; but 
few of its forms have been met with in use. 

b. The stem rai, f. (or m.), 'wealth', might be better described 
as ra with a union-consonant y (258) interposed before vowel 
endings, and is regularly inflected as such, with normal endings 
and monosyllabic accent. Thus : ras, rayam, ray a, raye, ray as, 
rayi ; rayau, rabhyhm, rayos; rayas, ray as, rabhis, rabhyds, rayam, 
rasu. But in the Veda the accus. pi. is either ray as or ray as; 
for accus. sing, and pi. are also used the briefer forms ram 




(which alone is of Vedic occurrence) and ras; and the gen. sing, 
is sometimes anomalously accented rayas. 

c. The stem go, m. or f., 'bull' or 'cow', is much more irreg- 
ular. In the strong cases, except accus. sing., it is strengthened 
to gau, forming (like nau) gaus, gavau, gavas. In accus. sing, 
and pi. it has (like rat) the brief forms ghm and gas. The 
abl.-gen. sing, is gos (as if from gu). The rest is regularly 
made from go, with the normal endings, but with accent always 
remaining irregularly upon the stem : thus, gdva, gave, gdvi ; 
gdvos ; gdvam; gobhyam, gobhis, gobhyas, gom. In the Veda, 
another form of the gen. pi. is gonam; the nom. etc. du. is 
(as in all other such cases) also gava; and gam, gos, and ghs 
are not infrequently to be pronounced as dissyllables. 

d. The stem dyo, f. (but in V. usually m.), 'sky, day', is yet 
more anomalous, having beside it a simpler stem dyu, which 
becomes div before a vowel-ending. The native grammarians 
treat the two as independent words, but it is more convenient 
to put them together. The stem dyo is inflected precisely like 
go, as above described. The complete declension is as follows 
(with forms not actually met with in use bracketed) : 

[dwau] dyavau 

[dyubhyam dyobhyam] 


















dyts I 



dydvi f 



dfvas dyavas 
divds, dyun [dyds] 
dyiibhis [dytibhis] 



y language. 

Both, dfvas 

The dat. sing, dydve is not found in the early language, 
and divas occur as accus. pi. in V. As nom. etc. du., dyava is, as usual, 
the regular Vedic form: once occurs dydvi (du.), as if a neuter form; and 
dyatis is found once used as ablative. The cases dyaus, dyam, and dyun 
(once) are read in V. sometimes as dissyllables ; and the first as accented 
vocative then becomes dyaus (i. e. dfaus : see 314). 

e. Adjective compounds having diphthongal stems as final member 
are not numerous. For go we have gu in such a position in dgu, sugii, and 
a few others ; and, correspondingly, rat seems to be reduced to ri in brhddraye 
and rdhddrayas (RV.). In revdnt (unless this is for rayivant), rat becomes re. 
In a few compounds, dyu or dyo is anomalously treated as first member: 
thus, dyausam$ita (AV.), dyaurda (K.), dyaurlokd (B.). 

B. Derivative stems in a. I, u. 
362. To this division belong all the a and ^-sterns 


which have not been specified above as belonging to the 
other or root-word division ; and also, in the later language, 
most of the I and w-stems of the other division, by transfer 
to a more predominant mode of inflection. Thus : 

a. The great mass of derivative feminine a-stems, sub- 
stantive and adjective. 

The inflection of these stems has maintained itself with little change 
through the whole history of the language, being almost precisely the same 
in the Vedas as later. 

b. The great mass of derivative feminine e-stems. 

This class is without exception in the later language. In the earlier, 
it suffers the exception pointed out above (355 b] : that feminines made with 
change of accent follow this mode of declension only when the accent is not 
on the T: thus, tdvisi, pdrumi, pdliknl, rtfhini. 

The i-stems of this division in general are regarded as made by con- 
traction of an earlier ending in ya. Their inflection has become in the later 
language somewhat mixed with that of the other division, and so far different 
from the Vedic inflection : see below, 364, end. 

Very few derivative stems in I are recognized by the grammarians as 
declined like the root-division; the Vedic words of that class are, if retained 
in use, transferred to this mode of inflection. 

A very small number of masculine z-stems (half-a-dozen) are in the 
Veda declined as of the derivative division: they are a few rare proper 
names, mdtali etc. ; and r&stri and sin (only one case each). 

c. The w-stems are few in number, and are transfers from 
the other division, assimilated in inflection to the great class of 
derivative i-stems (except that they retain the ending s of the 
nom. sing.). 

363. Endings. The points of distinction between this and the other 
division are as follows : 

In nom. sing, the usual s-ending is wanting: except in the u-stems 
and a very few t-stems namely, laksmi, tarl, tantrl which have pre- 
served the ending of the other division. 

The accus. sing, and pi. add simply m and s respectively. 

The dat., abl.-gen., and loc. sing, take always the fuller endings ai, 
as, am; and these are separated from the final of the a-stems by an inter- 
posed y. 

Before the endings a of instr. sing, and os of gen. -loc. du., the final 
of a-stems is treated as if changed to e; but in the Veda, the instr. ending 
a very often (in nearly half the occurrences) blends with the final to a. 
The ya of i-stems is in a few Vedic examples contracted to i, and even 
to i. A loc. sing, in I occurs a few times. 


In all the weakest cases above mentioned, the accent of an I or u-stem 
having acute final is thrown forward upon the ending. In the remaining 
case of the same class, the gen. pi., a n is always interposed between stem 
and ending, and the accent remains upon the former (in RV., however, it 
is usually thrown forward upon the ending, as in i and w-stems). 

In voc. sing., final a becomes e; final t and u are shortened. 

In nom.-acc.-voc. du. and nom. pi. appears in I (and u) -stems a marked 
difference between the earlier and later language, the latter borrowing the 
forms of the other division. The du. ending du is unknown in RV., and 
very rare in AV. ; the Vedic ending is i (a corresponding dual of ii-stems 
does not occur). The regular later pi. ending as has only a doubtful ex- 
ample or two in RV., and a very small number in AV. ; the case there (and 
it is one of very frequent occurrence) adds s simply; and though yas-forms 
occur in the Brahmanas, along with zs-forms, both are used indifferently as 
nom. and accus. Of a-stems, the du. nom. etc. ends in e, both earlier and 
later; in pi., of course, s-forms are indistinguishable from ow-forms. The 
RV. has a few examples of dsas for as. 

The remaining cases call for no remark. 

364. Examples of declension. As models of the 
inflection of derivative stems ending in long vowels, we 
may take ?RT sena, f. ; 'army'; ^TT kanya, f., 'girl'; 
devi, f., 'goddess'; ofSJ vadhu, f., 'woman'. 

Singular : 
















kanyh devi 

kanyhm devim 

SR^^I ^cETT 

kanyaya devya 

=fI*^JT7T 5^?J 

kanyhyai devyai 

oh^lMIH rfctllH 

kanyhyas devytis 

kanyayam devyam 














N.A. V. 



devyau } 




kanyhbhyam dev^bhyam 
kanyayos devyos 

Plural : 





kanyhs devyas 

kanytts devis 

cfl^tlhHH ^cjlMH 

kanyhbhis devibhis 

D. Ab. 





c\ -x _ 







ex , 



senabhyas kanyabhyas devibhyas vadhubhyas o 

^l^l^m eft^tl MIH rfcjHlH eJM'SIH 

senanam kanyhnam devinam vadlmnam 

kanyhsu devisu vadhusu 

In the Yeda, vadhu is a stern belonging to the other division (like tanu, 
above, 356). 

365. Examples of Vedic forms are : 

1. a-stems: instr. sing, mamsd (this simpler form is especially common 
from stems in td and ia); nom. pi. vafdsas (about twenty examples : Lanman, 
p. 362); accus. pi. aramgamdsas (a case or two). Half the 6fa/as-cases are 
to be read as bhias; the am of gen. pi. is a few times to be resolved into 
aam; and the a and am of nom. and accus. sing, are, very rarely, to be 
treated in the same manner. 

2. t-stems: instr. sing, pdrm, pdmi; loc. gauri ; nom. etc. du. devi; 
nom. pi. devis; gen. pi. bahvlndm. The final of the stem is to be read as 
a vowel (not y) frequently, but not in the majority of instances : thus, devid, 
devias, deviam, rddasios. 

The sporadic instances of transfer between this division and the pre- 
ceding have been already sufficiently noticed. 

3. In the language of the Brahmanas, the abl.-gen. sing, ending as is 
almost unknown, and instead of it is used the dat. ending di. The assump- 
tion of the same substitution is suggested, but not required, in a few RV. 


passages; and it is necessary once in AY. (iv. 5.6): svdpantv asydi jndtdyah, 
'let her relatives sleep'. Brahmana examples are: tdsydi di$dh (TS.), 'from 
that direction'; striydi payah (AB.), 'woman's milk'; dhenvaf vd etdd retail 
(TB.), 'that, forsooth, is the seed of the cow'; jydyasl ydjydydi (AB.), 'superior 
to the ydjyd'. 

366. The noun strf, f., 'woman' (probably contracted from sutrl, 'gene- 
ratress'), follows a mixed declension : thus, stri , striyam or sfrfm, striyd, 
striyai, striyds, striydm, strf; strfydu, stribhydm, striyos ; strfyas, strfyas or 
stris, stribhfs, stnbhyds, strindm, strlsu (but the accusatives strim and stris 
are not found in the older language, and the voc. stri is not quotable). The 
accentuation is that of a root-word; the forms (conspicuously the nom. sing.) 
are those of the other division. 


367. a. The occurrence of original adjectives in long final 
vowels, and of compounds having as final member a stem of 
the first division, has been sufficiently treated above, so far as 
masculine and feminine forms are concerned. To form a neuter 
stem in composition, the rule of the later language is that the 
final long vowel be shortened ; and the stem so made is to be 
inflected like an adjective in i or u (339, 341). 

Such neuter forms are very rare, and in the older language almost 
unknown. Of neuters from z-stems have been noted in the Veda only 
hariprfyam, ace. sing, (a masc. form), and suadhfas, gen. sing, (same as 
masc. and fern.); from ii-stems, only a few examples, and from stem-forms 
which might be masc. and fern, also : thus, vibhti, subhti, etc. (nom. -ace. 
sing.: compare 354); supud and mayobhtivd, instr. sing.; and mayobhu, ace. 
pi. (compare purti: 342); from a-stems occur only half-a-dozen examples 
of a nom. sing, in as, like the masc. and fern. form. 

b. Compounds having nouns of the second division as 
final member are common only from derivatives in a ; and these 
shorten the final to a in both masculine and neuter : thus, from 
a, 'not', and praja, 'progeny', come the masc. and neut. stem 
apraja, fern, aprcy'a, 'childless'. Such compounds with nouns 
in i and u are said to be inflected in masc. and fern, like the 
simple words (only with in and un in ace. pi. masc.); but the 
examples given by the grammarians are fictitious. The stem 
stri is directed to be shortened to -stri for all genders. 

368. It is convenient to give a complete paradigm, 
for all genders, of an adjective-stem in [ a. We take for 
the purpose efF7 papa, 'evil', of which the feminine is usu- 
ally made in m a in the later language, but in ^ 1 in the 




Singular : 









papas papam 








papau pape 

I. D.Ab. 



Plural : 



D. Ab. 

papas papni 


papan paphni 









pap ay am 












H(L(>J |i-| 







_ f -x 


371] DECLENSION IV., /--STEMS. 123 


papanam papanam paplnam 

papesu papasu papisu 

Declension IV. 

Stems in r (or ar). 

369. This declension is a comparatively limited one, 
being almost entirely composed of derivative nouns formed 
with the suffix rT tr (or cT|" tar\ which makes masculine 
nomina agentis (used also participially), and a few nouns of 
relationship . 

But it includes also a few nouns of relationship not made 
with that suffix: namely devf, m., svdsr and ndnandr, f . ; and, 
besides these, nf, m., sir (in V.), m., nsf (in V.), f., savyasthr, 
m., and the feminine numerals tisr and catasr (for which, see 
chap. VI.). The feminines in tr are only matr, duhitr, and yhtr. 

The inflection of these stems is quite closely analogous with 
that of stems in i and u (second declension) ; its peculiarity, as 
compared with them, consists mainly in the treatment of the 
stem itself, which has a double form, fuller in the strong cases, 
briefer in the weak ones. 

370. Forms of the Stem. In the weak cases (excepting 
the loc. sing.) the stem-final is r, which in the weakest cases, 
or before a vowel-ending, is changed regularly to r (129). 
But as regards the strong cases, the stems of this declension 
fall into two classes : in one of them which is very much 
the larger, containing all the nomina agentis, and also the nouns 
of relationship naptr and svdsr, and the irregular words sir and 
savyasthr the r is vriddhied, or becomes ar; in the other, 
containing most of the nouns of relationship, with nf and usf, 
the r is gunated, or changed to ar. In both classes, the loc. 
sing, has ar as stem-final. 

371. Endings. These are in general the normal, but 
with the following exceptions : 

The nom. sing. (masc. and fern.) ends always in a (for original ars). 
The voc. sing, ends in ar. 

The accus. sing, adds am to the (strengthened) stem; the accus. pi. 
has (like i and w-stems) n as masc. ending and s as fern, ending, with the 
r lengthened before them. 


The abl.-gen. sing, changes r to wr (or us: 169, end). 

The gen. pi. (as in i and w-stems) inserts n before am, and lengthens 
the stem-final before it. But the r of nr may also remain short. 

The above are the rules of the later language. The older presents cer- 
tain deviations from them. Thus : 

The ending in nom. etc. du. is (as universally in the Veda) regularly a 
instead of du (only ten cm-forms in RV.). 

The i of loc. sing, is lengthened to I in a few words : thus, kartdri. 

In the gen. pi., the RV. has once svdsram, without inserted n; and 
nardm instead of nrndm is frequent. 

Other irregularities of nr are the sing. dat. narc, gen. ndras, and loc. 
ndri. The Veda writes always nrndm in gen. pi., but its r is in a majority 
of cases metrically long. 

The stem wsr, f., 'dawn', has the voc. sing, usar, the gen. sing, usrds ; 
and the accus. pi. also usrds, and loc. sing, usrdm (which is metrically 
trisyllabic: usram), as if in analogy with I and w-stems. Once occurs usri 
in loc. sing., but it is to be read as if the regular trisyllabic form, usdri 
(for the exchange of s and , see 181 a). 

From stf come only tdras (apparently) and stfbhis. 

In the gen. -loc. du., the r is almost always to be read as a separate 
syllable, r, before the ending os : thus, pitrtis, etc. On the contrary, ndnandari 
is once to be read ndnandri. 

For neuter forms, see below, 378. 

372. Accent. The accentuation follows closely the rules 
for i and ?<-stems : if on the final of the stem, it continues, as 
acute, on the corresponding syllable throughout, except in the 
gen. pi., where it may be (and in the Veda always is) thrown 
forward upon the ending ; where, in the weakest cases, f becomes 
r, the ending has the accent. The two monosyllabic stems, nr 
and stf, do not show the monosyllabic accent : thus (besides the 
forms already given above), nfbhis, 

373. Examples of declension. As models of this 
mode of inflection, we may take from the first class (with 
5TTJ" ar in the strong forms) the stems ^JrT datr, m., 'giver, 
and ^RT svdsr, f., 'sister'; from the second class (with ET^f 
ar in the strong forms) the stem farT^r, m., 'father'. 

Singular : 

^THT ^oFfT focTT 

dath svdsa pith 


dataram svdsaram pitdram 





Ab. G. 


N.A. V. 


Plural : 










, "s 










The feminine stem 




t "N, 















atr, 'mother', is inflected pre- 

cisely like RT ^r, excepting that its accusative plural is 


The peculiar Vedic forms have been sufficiently instanced above; the 
only ones of other than sporadic occurrence being the nom. etc. du. datdra, 
avdsara, pitdra, and the gen. pi. of nr, nardm. 

374. The stem krostf, m., 'jackal' (lit'ly 'howler'), substi- 
tutes in the middle cases the corresponding forms of krostu. 

375. Neuter forms. The grammarians prescribe a com- 
plete neuter declension also for bases in tr, precisely accordant 
with that of vhri or mddhu (above, 339, 341). Thus, for ex- 
ample : 

Sing. Du. Plur. 

N. A. dhatf dhatfni dhatfni 

I. dhatfna dhdtfbhyam dhdtfbhis 

G. dhatfnas dhatf nos dhdtfndm 

V. dhdtr, dhdtar dhdtrnl dhdtfni. 

The weakest cases, however (as of * and w-stems used ad- 
jectively : 344), are allowed also to be formed like the corre- 
sponding masculine cases : thus, dhatrh etc. 

No such neuter forms chance to occur in the Veda, but they begin to 
appear in the Brahmanas, under influence of the common tendency (compare 
Germ. Better, Retterin; Fr. menteur, menteuse) to give the nomen agentis a 
more adjective character, making it correspond in gender with the noun 
which it (appositively) qualifies. Thus, we have in TB. bhartf and janayitf, 
qualifying antdriksam; and bhartfni and janayitfni, qualifying ndksatrdni ; as, 
in M., grahitfni, qualifying indriydni. 

When a feminine noun is to be qualified in like manner, the usual 
feminine derivative in I is employed: thus, in TB., bhartryas and bhartrydii, 
janayitryas and janayitrydu, qualifying dpah and ahoratre ; and such in- 
stances are not uncommon. 

The RV. shows the same tendency very curiously once in the accus. 

pi. matfn, instead of matfs, in apposition with masculine nouns (RV. x. 35.2). 

Other neuter forms in RV. are sthattir, gen. sing., dhmatdri, loc. sing. ; 

and for the nom. sing., instead of -tr, a few more or less doubtful cases, 

sthatar, sthatur, dhartdri (Lanman, p. 422). 


376. a. There are no original adjectives of this declension : 
for the quasi-adjectival character of the nouns composing it, see 
above (378). The feminine stem is made by the suffix i: thus, 
datrk, dhatri. 

b. Roots ending in r (like those in i and u: 345) add 
a t to make a declinable stem, when occurring as final member 
of a compound : thus, 'karmakrt (]/ 'kr\ vajrabhft (ybfy), balihft 
(yfo). From some r-roots, also, are made stems in ir and ur : 
see below, 383 a, b. 


c. Nouns in r as finals of adjective compounds are in- 
flected in the same manner as when simple, in the masculine 
and feminine ; in the neuter, they would doubtless have the 
peculiar neuter endings in nom.-acc.-voc. of all numbers. 

Declension V. 

Stems ending in Consonants. 

377. All stems ending in consonants may properly be 
classed together, as forming a single comprehensive declen- 
sion; since, though some of them exhibit peculiarities of 
inflection, these have to do almost exclusively with the stem 
itself, and not with the declensional endings. 

378. In this declension, masculines and feminiiies of 
the same final are inflected precisely alike ; and neuters are 
peculiar (as usually in the other declensions) only in the 
nom.-acc.-voc. of all numbers. 

The majority of consonantal stems, however, are not 
inflected in the feminine, but form a special feminine deriv- 
ative stem in ^ * (never in 5TT a), by adding that ending to 
the weak form of the masculine. 

Exceptions are in general the stems of divisions A and B 
namely, the radical stems etc . , and those in as and is and us . 
For special cases, see below. 

379. Variations, as between stronger and weaker forms, 
are very general among consonantal stems : either of two 
degrees (strong and weak), or of three (strong, middle, and 
weakest): see above, 311. 

The peculiar neuter forms, according to the usual rule 
(311), are made in the plural from the strong stem, in sin- 
gular and dual from the weak or, when the gradation 
is threefold, in singular from the middle stem, in dual 
from the weakest. 

As in the case of stems ending in short vowels (asyatu, 
)&rmi, mddhuni, datfni, etc.), a nasal sometimes appears in the 


special neuter plural cases which is found nowhere else in in- 
flection. Thus, from the stems in as, is, us, the nom. etc. pi. 
in -ansi, -insi, -unsi are very common at every period. Accord- 
ing to the grammarians, the radical stems etc. (division A) are 
treated in the same way; but examples of such neuters are of 
excessive rarity in the older language ; no Vedic text offers one, 
and in the Brahmanas have been noted only -hunti (AB. vii. 2), 
-vrnti (PB. xvi. 2.7 et al.), and -bhanji (KB. xxvii. 7) : it may 
be questioned whether they are not late analogical formations. 

380. The endings are throughout those given above 
(310) as the "normal". 

By the general law as to finals (150), the s of the nom. sing, 
masc. and fern, is always lost; and irregularities of treatment 
of the final of the stem in this case are not infrequent. 

The gen. and abl. sing, are never distinguished in form 
from one another nor are, by ending, the nom. and accus. pi. : 
but these sometimes differ in stem-form, or in accent, or in both. 

381. Change in the place of the accent is limited to mon- 
osyllabic stems and the participles in dnt (accented on the final). 
For details, see below, under divisions A and E. 

A few of the compounds of the root anc or ac show an irregular shift 
of accent in the oldest language: see below, 410. 

382. For convenience and clearness of presentation, it 
will be well to separate from the general mass of conson- 
antal steins certain special classes which show kindred pe- 
culiarities of inflection, and may be best described together. 

B. Derivative stems in as, is, us; 

C. Derivative stems in an (an, man, van); 

D. Derivative stems in in (in, min, vin); 

E. Derivative stems in ant (ant, mant, vant); 
1?. Perfect active participles in vans; 

G. Comparatives in yas. 

There remain, then, to constitute division A, especially 
radical stems, or those identical in form with roots, to- 
gether with a comparatively small number of others which 
are inflected like these. 

They will be taken up in the order thus indicated. 


A. Root-stems, and those inflected like them. 
383. The stems of this division may be classified as 
follows : 

a. Root-stems, having in them no demonstrable element 
added to a root : thus, re, 'verse' ; gir, 'song' ; pad, 'foot' ; 
die, 'direction'; mdh (V.), 'great'. 

Such stems, however, are not always precisely identical in form with 
the root : thus, vac from yvac, srdj from ysrj, mus from ymus, vrif from 
yvrafc (?), t/5 from yvas 'shine'; and from roots in final r come stems 
in ir and ur : thus, gfr, a-ffr, stir; jur, ttir, dhtir, pur, mur, 3tur; and 
psur from ypsar. 

With these may be ranked the stems with reduplicated root, as ciktt, 
yaviytidh, vdnivan, sasydd. 

Words of this division in uncompounded use are tolerably frequent in 
the older language : thus, in RV. are found more than a hundred of them ; 
in AV., about sixty; but in the classical Sanskrit the power of using any 
root at will in this way is lost, and the examples are comparatively few. 
In all periods, however, the adjective use as final of a compound is very 
common (see below, 401). 

b. Stems made by the addition of t to a final short vowel 
of a root. 

No proper root-stem ends in a short vowel, although there are (above, 
354) examples of transfer of such to vowel-declensions ; but i or u or r 
adds a t to make a declinable form: thus, -jft, -crtit, -Urt. Roots in r > 
however, as has just been seen, also make stems in ir or ur. 

As regards the frequency and use of these words, the same is true as 
was stated above respecting root-stems. The Veda offers examples of nearly 
thirty such formations, a few of them (mtt, rtt, stut, hrtit, vrt, and dytit if 
this is taken from dyu) in independent use. Of roots in r, kr, dhr, dhvr, 
&Ar, vr, sr } spr ) hr, hvr add the t. The roots ga (or gam) and han also 
make -gdt and -hdt by addition of the t to an abbreviated form in a (thus r 
adhvagdt, dyugdt, navagdt, and samhdt}. 

As to the infinitive use of various cases of the root-noun in these two- 
forms, see chap. XIII. 

C. Monosyllabic (also apparently reduplicated) stems not cer- 
tainly connectible with any verbal root in the language, but 
having the aspect of root-stems, as containing no traceable suffix : 
thus, tvdc, 'skin' ; path, 'road' ; hfd, 'heart' ; dp, 'water' ; as, 
'mouth' ; kakubh and kakud, 'summit' . 

Thirty or forty such words are found in the older language, and some 
of them continue in later use, while others have been transferred to other 
modes of declension or have become extinct. 

d. Stems more or less clearly derivative, but made with 
suffixes of rare or even isolated occurrence. Thus : 

Whitney, Grammar. 9 


1. derivatives (V.) from prepositions with the suffix vat: arvavdt, avdt, 
udvdt, nivdt, paravdt, pravdt, samvdt; 

2. derivatives (V.) in tat (perhaps abbreviated from tati], in a few iso- 
lated forms : thus, upardtat, devdtat, vrkdtat, satydtat, sarvdtat ; 

3. other derivatives in t, preceded by various vowels : thus, vehdt, vahdt, 
sravdt, safcdt, vdghat; ndpat; tadft, divtt, yosit, rohtt, sartt, hartt; martit; 
ydkrt, fdkrt; and the numerals for '30, 40, 50', trihfdt etc. (475); 

4. stems in ad: thus, drsdd, dhrsdd, bhasdd, vandd, cardd; 

5. stems in j, preceded by various vowels: thus, trsndj, dhrsdj, sandj ; 
utfj, vanfj, bhurfj, nintj(?); dsrj; 

6. a few stems ending in a sibilant apparently formative : thus, jnds, 
-das, bhdsj mds, bhis ; 

1. a remnant of unclassifiable cases, such as vistdp, vtpaf, kdprth, 
, isfdh, prkstidh, raghdtf?). 

384. Gender. The root-stems are regularly feminine as 
nomen actionis, and masculine as nomen agentis (which is probably 
only a substantive use of their adjective value : below, 400). 
But the feminine noun, without changing its gender, is often 
also used concretely : e. g., druhi. (ydruh, e be inimical') means 
'harming, enmity', and also 'harmer, hater, enemy' thus 
bordering on the masculine value. And some of the feminines 
have a completely concrete meaning. Through the whole divis- 
ion, the masculines are much less numerous than the feminines, 
and the neuters rarest of all. 

The independent neuter stems are hrd (also -hard), ddm, vdr, svbr, 
mds 'flesh', ds 'mouth', bhds, dos, and the indeclinables fdm and y<5& : also 
the derivatives ydkrt, fdkrt, dsrj. 

385. Strong and weak stem-forms. The distinc- 
tion of these two classes of forms is usually made either 
by the presence or absence of a nasal, or by a difference 
in the quantity of the stem-vowel, as long or short; less 
often, by other methods. 

386. A nasal appears in the strong cases of the follow- 
ing words : 

a. Compounds having as final member the root ac or anc: see below, 
407 ff . ; b. The stem J/MJ, sometimes (V.): thus, nom. sing, ydn (for 
yunk), accus. yCinjam, du. ytinja (but also ytijam and yuja); c. The 
stem -drf, as final of a compound (V.); but only in the nom. sing, masc., 
and not always : thus, anyadrn, Urn, kidfn, sadrn and pratisadrn : but also 
Idrk, tadfk, svardfk, etc.,- d. For path and pums, which substitute more 
extended stems, and for dant, see below, 3946. 

387. The vowel a is lengthened in strong cases as follows : 


a. Of the roots vac, sac, sap, nabh, pas, in a few instances (V.), at the 
end of compounds ; b. Of the roots vah and sah, but irregularly : see 
below, 403 5 ; c. Of ap 'water' (see 393) ; also in its compound rltyap ; 
d. Of pad, 'foot': in the compounds of this word, in the later language, 
the same lengthening is made in the middle cases also ; and in RV. and AV. 
the nom. sing. neut. is both -pat and -pat, while -pddbhis and -patsu occur 
in the Brahmanas; e. Of nas, 'nose'; f. Sporadic cases (V.) are: 
yaj (?), voc. sing. ; pathds and -rapas, accus. pi. ; vdnivanas, nom. pi. The 
strengthened forms bhdj and raj are constant, through all classes of cases. 

388. Other modes of differentiation, by elision of a or 
contraction of the syllable containing it, appear in a few stems : 

a. In -han: see below, 402; b. In ksam (V.), along with prolong- 
ation of a: thus, ksdmd du., ksdmas pi.; ksamd instr. sing., ksdmi loc. 
sing., ksmds abl. sing. ; c. In dvar, contracted (V.) to dur in weak cases 
(but with some confusion of the two classes); d. In svar, which becomes 
(RY.) sur in weak cases : later it is indeclinable. 

389. The endings are as stated above (380). 

Respecting their combination with the final of the stem, as 
well as the treatment of the latter when it occurs at the end of 
the word, the rules of euphonic combination (chap. III.) are to 
be consulted ; they require much more constant and various 
application here than anywhere else in declension. 

Attention may be called to a few exceptional cases of combination (V.): 
madbhis and mddbhyds from mas 'month' ; the wholly anomalous padbMs (RV. 
and VS.: AV. has always padbhis} from pad; and sardt and sarddbhyas cor- 
responding to a nom. pi. sardghas (instead of sardhas : 222). Ddn is appar- 
ently for ddm, by 143, end. Agnfdh is abbreviated from ayni-fdh. 

According to the grammarians, neuter stems, unless they end in a nasal 
or a semivowel, take in nom. -ace. -voc. pi. a strengthening nasal before the 
final consonant. But no such cases from neuter noun-stems appear ever to 
have been met with in use; and as regards adjective stems ending in a root, 
see above, 379. 

390. Monosyllabic stems have the regular accent of such, 
throwing the tone forward upon the endings in the weak cases. 

But the accusative plural has its normal accentuation as a 
weak case, upon the ending ; in only a minority (hardly more 
than a third) of the stems : namely in datds, pathds, padds, 
nidds, apds, usds, jnasds, puthsds, masds, mahds ; and sometimes 
in vacds, srucds, hrutds, sridhds, ksapds, vipds, durds, isds, dvisds, 
druhds (beside v&cas etc.). 

Exceptional instances, in which a weak case has the tone on the stem, 
occur as follows : sddd, nddbhyas, tana (also tana) and tdne, rant and ransw, 
vdnsu, svdni, vfpas, ksdmi, surd and suras (but sure), dhhas, and vdnas and 
bfhas (in vdnaspdti, brhaspdti}. On the other hand, a strong case is accented 



















on the ending in mahds, uom. pi., and kasdm (AV. : perhaps a false reading). 
And presd, instr. sing., is accented as if pres were a simple stem, instead of 
pra-fs. Vimrdhdh is of doubtful character. For the sometimes anomalous 
accentuation of stems in ac or anc, see 410. 

391. Examples of inflection. As an example of 
normal monosyllabic inflection, we may take the stem 
cHrT vac, f., 'voice' (from y^ff vac, with constant prolong- 
ation); of inflection with strong and weak stem, q^~ pad, 
m., 'foot'; of polysyllabic inflection, H ^^marut, m., 'wind' 
or 'wind-god'; of a monosyllabic root-stem in composition, 
f^eJH trivrt, 'three-fold', in the neuter. Thus : 

Singular : 




Ab. G. 












i_ _ 

I. D. Ab. 





Plural : 

padau " 



marutau trivrti 

marudbhyam trivfdbhyam 
marutos trivftos 

r "^ 

vacds, vacas padds 







vagbhis padbhis marudbhis trivfdbhis 




vagbhyds padbhyds mariidbhyas trivrdbhyas 

padam marutam trivrtam 






392. The stems in ir and wr, and u and ws, lengthen their 
vowel (245 b) when their final is followed by another consonant, 
and also in the nom. sing, (where the following s is lost) : thus, 
from yzV, f., f song', ^Ir (^ijj, giram, girh etc. ; girau, g'irbhyUm, 
giros ; giras, girbhis, glrbhyds, giram, girsu (165) ; and so pur, 
puram, purbhis, pursu; and afis, acisam, acisa, aclrbhis, aclhsuj 
and so on. 

393. The stem dp, f., ( water', is inflected only in the 
plural, and with dissimilation of its final before bh to d (151 d) : 
thus, Upas, apds, adbhis, adbhyds, aphvn, apsu. 

But RV. has the sing, instr. apd and gen. apds. In AV. often, and 
in an instance or two in RV., the nom. and accus. pi. forms are confused 
in use, dpas being employed as accus., and (in an instance or two) apds as 

394. The stem pums, m., 'man', is very irregular, sub- 
stituting pumcihs in the strong cases, and losing its s (necessarily) 
before initial bh of a case-ending, and likewise (by analogy with 
this, or by an abbreviation akin with that noticed at 231) in 
the loc. plural. The vocative is (in accordance with that of the 
somewhat similarly inflected perfect participles : below, 462 a) 
pitman in the later language, but pumas in the earlier. Thus : 
puman, pumansam, pumsa, pumse, pumsds, pumsi, puman; pum- 
ansau, pumbhytim, pumsos ; pumansas, pumsds, pumbhis, pumbhyds, 
pumshm, pumsu. 

The accentuation of the weak forms, it will be noticed, is that of a true 
monosyllabic stem. The forms with &/i-endings nowhere occur in the older 
language, nor do they appear to have been cited from the later. As to the 
retention of s unlingualized in the weakest cases (whence necessarily follows 
that in the loc. pi.), see 183. 

395. The stem path, m., 'road', is defective in declension, 
forming only the weakest cases, while the strong are made from 
pdntha or pdnthan, and the middle from pathl : see under an- 
stems, below, 433. 


396. The stem ddnt, m., Hooth', is perhaps of participial 
origin, and has, like a participle, the forms ddnt and ddt, strong 
and weak : thus (V.), ddn, ddntam, data, etc. ; datds ace. pi. 
etc. But in the middle cases it has the monosyllabic and not 
the participial accent : thus, dadbhis, dadbhyds. In nom. pi. 
occurs also -datas instead of -dantas. By the grammarians, the 
strong cases of this word are required to be made from ddnta. 

397. A number of other words of this division are de- 
fective, making part of their inflection from stems of a differ- 
ent form. 

Thus, hfd, n., 'heart', mdhs or mds, n., 'meat', mds, m., 'month', 
nds, f., 'nose', nip, f., 'night' (not found in the older language), pft, f., 
'army', are said by the grammarians to lack the nom. of all numbers and 
the accus. sing, and du. (the neuters, of course, the ace. pi. also), making 
them respectively from hfdaya, mdnsd, mdsa, ndsikd, nifd, pftand. But the 
usage in the older language is not entirely in accordance with this require- 
ment : thus, we find mds, 'flesh', accus. sing.; mds, 'month', nom. sing.; 
and ndsa, 'nostrils', du. From pft occurs only the loc. pi. prtsti and (RV., 
once) the same case with double ending, prtstisu. 

398. On the other hand, certain stems of this division, 
allowed by the grammarians a full inflection, are used to fill up 
the deficiencies of tho'pe of another form. 

Thus, dsrj f n., 'blood', fdkrt, n., 'ordure', ydkrt, n., 'liver', do's, n. 
(also m.), 'fore-arm', have beside them defective stems in an: see below, 
432. Of none of them, however, is anything but the nom. -ace. sing, found 
in the older language, and other cases later are but very scantily represented. 

Of ds, n., 'mouth', and tid, 'water', only a case or two are found, in 
the older language, beside asdn and dsya, and uddn and tidaka (432). 

399. Some of the alternative stems mentioned above are instances of 
transition from the consonant to a vowel declension : thus, ddnta, masa. 
A number of other similar cases occur, sporadically in the older language, 
more commonly in the later. Such are pdda, -mdda, -ddfa, bhrdjd, vistdpa, 
dvdra and dura, pura, dhura, -drya, ndsa, nidd, ksipd, ksapd, dfd, and 
perhaps a few others. 

A few irregular stems will find a more proper place under the head of 


400. Original adjectives having the root-form are compar- 
atively rare even in the oldest language. 

About a dozen are quotable from the RV., for the most part only in a 
few scattering cases. But mah, 'great', is common in RV., though it dies 
out rapidly later. It makes a derivative feminine stem, mahi, which con- 
tinues in use, as meaning 'earth' etc. 




401. But compound adjectives, having a root as final 
member, with the value of a present participle, are abundant 
in every period of the language. 

Possessive adjective compounds, also, of the same form, 
are not very rare : examples are yatdsruc, 'with offered bowl' ; 
s&ryatvac, 'sun-skinned' ; cdtuspad, 'four-footed ; suhurd, 'kind- 
hearted, friendly'; rityap (i. e. riti-ap], 'having streaming waters' ; 
sahdsradvar, 'furnished with a thousand doors'. 

The inflection of such compounds is like that of the simple root-stems, 
masculine and feminine being throughout the same, and the neuter varying 
only in the nom.-acc.-voc. of all numbers. 

Only rarely is a derivative feminine stem in I formed : in the older 
language, only from the compounds with ac or one (407 ff.), those with han 
(402), and those with pad, as ekapadi, dvipddl. 

Irregularities of inflection appear in the following : 

402. The root han, 'slay', as final of a compound, is in- 
flected somewhat like a derivative noun in an (below, 420 ff.), 
becoming ha in the nom. sing., and losing its n in the middle 
cases and its a in the weakest cases (but only optionally in the 
loc. sing.). Further, when the vowel is lost, h in contract with 
following n reverts to its original gh. Thus : 



A. vrtrahdnam 

I. vrtraghnd 

D. vrtraghne 


Ab. ) 
> vrtragnnas 

L. vrtraghnt, -hdni 


V. vrtrahan vrtrahanau 




As to the change of n to n, see 193, 195. 

A feminine is made by adding I to, as usual, the stem-form shown in 
the weakest cases : thus, vrtraghni. 

An accus. pi. -hdnas (like the nom.) also occurs. Vrtrdhabhis (RV. , 
once) is the only middle case-form quotable from the older language. Trans- 
itions to the a-declension begin already in the Veda: thus, to -hd (RV., 
AV.), -ghnd (RV.), -hana. 

403. The root vah, 'carry', at the end of a compound is 
said by the grammarians to be lengthened to vah in both the 
strong and middle cases, and contracted in the weakest cases to 
uh, which with a preceding a-vowel becomes au (137 d): thus, 
from havyavdh, 'sacrifice-bearing' (epithet of Agni\ havyavat, 
havyav&ham, havyauha, etc.; havyavahau, havyavadbhyam, havyauhos; 
havyavahas, havyauhas, havyavadbhis, etc. And cvetavdh (not quot- 


able) is said to be further irregular in making the nom. sing, 
in vas and the vocative in vas or vas. 

In the earlier language, only strong forms of compounds with vah have 
been found to occur: namely, -vdt, -vdham, -vdhdu or -vdha, and -vdhas. 
TS. has the irregular nom. sing, pasthavdt. 

404. Of very irregular formation and inflection is one 
common compound of vah, namely anadvdh (anas -\-vah, 'burden- 
bearing' or 'cart-drawing' : i. e. 'ox'). Its stem-form in the 
strong cases is anadvah, in the weakest anadu/i, and in the 
middle anadud (perhaps by dissimilation from anadud}. Moreover, 
its nom. and voc. sing, are made in van and van (as if from 
a v ant-stem}. Thus : 

Singular. Dual. Plural. 

N. anadvdn ) 3 ~-i- anadvdhas 

, \ anadvahau 

A. anadvaham ) anaduhas 

I. anaduha | anadudbhis 

D. anaduhe {.anadudbhyam \ 


_, ' I anaduhas 

G. } t anaduham 

} anaduhos 
L. anaduhi | anadutsu 

V. dnadvan anadvahau dnadvahas 

Anadudbhyas (AV., once) is the only middle case-form quotable from 

the older language. 

The corresponding feminine stem (of very infrequent occurrence) is either 

anaduhi (QB.) or anadvahl (K.). 

405. The root sah, 'overcome', has in the Veda a double irregularity- 
its s is changeable to * even after an a-vowel as also in its single oc- 
currence as an independent adjective (RV., tvdrh sdt) while it sometimes 
remains unchanged after an i or w-vowel; and its a is either prolonged or 
remains unchanged, in both strong and weak cases. The quotable forms are : 
-sdt, -sdham or -sdham or -sdham, -sdha, -sdhe or -sdhe, -sdhas or -sdhas 
or -sdhas; -sdha (du.); -sahas or -sdhas. 

406. The compound avaydj (j/yaj, 'make offering'), 'a certain priest' or 
(BR.) 'a certain sacrifice', is said to form the nom. and voc. sing, avayas, 
and to make its middle cases from avayds. 

Its only quotable form is avayds, f. (RV. and AV., each once). If the 
stein is a derivative from ava-{-yyaj, 'conciliate', avayas is probably from 
ava -j- yyd, which has the same meaning. 

407. Compounds with anc or ac. The root ac or 
anc makes, in combination with prepositions and other words, 
a considerable class of familiarly used adjectives, of quite irreg- 
ular formation and inflection, in some of which it almost loses 
its character of root, and becomes an ending of derivation. 

A part of these adjectives have only two stem-forms : a 
strong in anc (yielding an, from anks, in nom. sing, masc.), 




and a weak in ac ; others distinguish from the middle in ac a 
weakest stem in c, before which the a is contracted with a pre- 
ceding i or u into i or u. 

The feminine is made by adding I to the stem-form used 
in the weakest cases, and is accented like them. 

408. As examples of inflection we may take prhnc, 'forward, 
east' ; pratydnc, 'backward, west' ; visvanc, 'going apart 1 . 
Singular : 

pratydn pratyak vfsvan vfsvak 

pratydncam pratyak vfsvancam vfsvak 

pratlcd vfsucd 

pratice vfsuce 

pratlcds visucas 

pratlci vCsuci 

N. V. 

prdn prdk 


prdncam prdk 



Ab. G. 



N. A. V. 

prdncau prdci 

I. D. Ab. 


G. L. 



N. V. 

prdncas prdnci 


pracas pranci 



D. Ab. 






pratydncdu pratlci 

vfsvancdu wfsuci 

pratydncas pratydnci visvancas vfsvanci 

pratlcds pratydnci visucas vfsvanci 

pratydgbhis vftvagbhis 

pratydgbhyas vfsvagbhyas 

praticdm vfsucdm 

pratydksu vfsvaksu 

The feminine stems are prhcl, pratici, visuci, respectively. 
No example of the middle forms excepting the nom. etc. sing. neut. 
(and this generally used as adverb) is found either in RV. or AV. In the 
same texts is lacking the nom. etc. pi. neut. in nci; but of this a number 
of examples occur in the Brahmanas : thus, prdnci, pratydnci, arvanci, 
samydnci, sadhryanci, anvanci. 

409. a. Like prune are inflected dpanc, dvanc, pdranc, 
arvhnc, adhartinc, and others of rare occurrence. 

b. Like pratydnc are inflected nydnc (i. e. mane), samydnc 
(sam -f- anc, with irregularly inserted e), and udanc (weakest stem 
udic : ud -j- anc, with i inserted in weakest cases only), with a 
few other rare stems. 

c. Like visvanc is inflected anvdnc, also three or four others 
of which only isolated forms occur. 

d. Still more irregular is tirydnc, of which the weakest 
stem is tirdcc (tirds -f- ac : the other stems are made from tir -\~ anc 
or ac, with the inserted i). 

410. The accentuation of these words is irregular, as regards both the 
stems themselves and their inflected forms. Sometimes the one element has 


the tone and sometimes the other, without any apparent reason for the dif- 
ference. If the compound is accented on the final syllable, the accent is 
shifted in RV. to the ending in the weakest cases provided their stem shows 
the contraction to i or it : thus, prdca, arvdca, adhardcas, but pratlcd, anucds, 
samici. But AV. and later texts usually keep the accent upon the stem: 
thus, praticl, samici, anUcl (RV. has praticlm once). The change of accent 
to the endings, and even in polysyllabic stems, is against all usual analogy. 

B. Derivative stems in as, is, us. 

411. The stems of this division are prevailingly neuter; 
but there are also a few masculines, and a single feminine. 

412. The stems in ^f as are quite numerous, and 
mostly made with the suffix 5fH as (a small number also 
with clTT tas and ^ nas, and some are obscure); the others 
are few, and almost all made with the suffixes ^js and 

V 413. Their inflection is almost entirely regular. But 
masculine and feminine stems in ^H as lengthen the vowel 
of the ending in nom. sing.; and the nom.-acc.-voc. pi. neut. 
make the same prolongation (of 5f a or ^ i or 3 u) before 
the inserted nasal (anusvara). 

414. Examples of declension. As examples we 
may take JHT mdnas, n., 'mind'; ^H^H dngiras, m., 'An- 

*X " v *S 

giras'; ^TJ^T havis, n., 'libation'. 

Singular : 

mdnas dngiras havis 

mdnas dngirasam havis 

mdnasa dngirasa havisa 

mdnase dngirase havise 

Ab.G. H^HH JbH^-^HH ^fc^^^ 

mdnasas dngirasas havisas 






I. D.Ab. 
G. L. 











angirobhyam havirbhyam 




Plural : 

















dngirahsu havihsu 

In like manner, tftHH cdksus, n., 'eye', forms 

r O "X ' v3 

cdksusa, TJ^I^IH cdksurbhyam, ri^fq cdksunsi, and so on. 

O "X 6\ 

415. Vedic irregularities, a. The masc. and fern. du. ending d 
instead of du is as usual elsewhere; b. The fern, usds, 'dawn', often 
prolongs its a in the other strong cases (besides nom. sing.): thus, usdsam, 
usdsd, usasas. In instr. pi. occurs (RV., once) usddbhis instead of usdbhis 
(only quotable example of a middle case). From tofds is once found (RV.) 
in like manner the du. to$tisd; c. JanHs has the nom. sing. masc. janUs, 
like an as-stern; d. From svdvas and svdtavas occur in RV. nom. sing, 
masc. in van; e. One or two apparently contracted forms thus, vedhdm 
for vedhdsam, and surddhds for surddhasas, nom. pi. are met with. 

416. The grammarians regard ufdnas, m., as regular stem-form of the 
proper name noticed above (355 a), but give it the irregular nom. u^dnd and 
the voc. ucanas or ufana or uyanan. Forms from the as-stem, even nom., 
are sometimes met with in the later literature. 


As to forms from as-stems to dhan or altar and udhan or udhar, see 
below, 430. 


417. A few neuter nouns in as with accent on the radical 
syllable have corresponding adjectives or appellatives in as, with 
accent on the ending: thus, for example, apas, 'work', apds, 
'active'; tdras, 'quickness', tar&s, 'quick'; ydcas, 'beauty', yacds, 
'beauteous'. A few other similar adjectives as tavds, 'mighty', 
vedhds, 'pious' are without corresponding nouns. 

Original adjectives in is do not occur. But in us are found 
as many adjectives as nouns (about ten of each class); and in 
several instances adjective and noun stand side by side, without 
difference of accent such as appears in the stems in as: e. g. 
tdpus, 'heat' and 'hot' ; vdpus, 'wonder' and 'wonderful' . 

V" 418. Adjective compounds having nouns of this division 
as final member are very common : thus, sumdnas, 'favorably 
minded' ; dirghayus, 'long-lived' ; cukrdgocis, 'having brilliant 
brightness'. The stem-form is the same for all genders, and 
each gender is inflected in the usual manner, the stems in as 
making their nom. sing. masc. and fern, in as (like dngiras, 
above). Thus, from sumdna$, the nom. and accus. are as follows : 

singular. dual. plural, 

m. f. n. m. f. n. ra. f. n. 

N. sumdnas -nas ) 

isumanasau -nasi sumanasas -nansi 

A. sumanasam -nas ) 

and the other cases (save the vocative) are alike in all genders. 

From dirghdyus, in like manner: 
N. dirghdyus 

A. dlrghdyusam -yus 
I. dlrghdyusa dlrghdyurbhyam dlrghdyurbhis 

etc. etc. etc. 

419. The stem one/ids, 'unrivalled' (defined as meaning 'time' in the 
later language), forms the nom. sing. masc. and fern, aneha. 

C. Derivative stems in an. 

420. The stems of this division are those made by the 
three suffixes EFT an, R man. and ^T van. together with a 
few of more questionable etymology which are inflected like 
them. They are masculine and neuter only. 

\ 421. The stem has a triple form. In the strong cases 
of the masculine, the vowel of the ending is prolonged to 

>. dirghdyusau -yusl dlrghdyusas -yunsi 


ETT <i; in the weakest cases it is in general struck out al- 
together; in the middle cases, or before a case-ending be- 
ginning with a consonant, the final ^ n is dropped. The 
^ n is also lost in the nom. sing, of both genders (leaving 
5TT as final in the masculine, f a in the neuter). 

The peculiar cases of the neuter follow the usual ana- 
logy (311): the nom.-acc.-voc. pi. have the lengthening 
to 5fT a, as strong cases; the nom.-acc.-voc. du., as weakest 
cases, have the loss of 5f a but this only optionally, not 

In the loc. sing, also, the a may be either rejected or re- 
tained (compare the corresponding usage with r-stems : 373). 
And after the m or v of man or van, when these are preceded 
by another consonant^ the a is always retained, to avoid a too 
great accumulation of consonants. 

422. The vocative sing, is in masculines the pure stem ; 
in neuters, either this or like the nominative. The rest of 
the inflection requires no description. 

423. As to accent, it needs only to be remarked that when, 
in the weakest cases, an acute d of the suffix is lost, the tone 
is thrown forward upon the ending. 

424. Examples of declension. As such may be 
taken ^TsR rajan, m. ? 'king'; ttirH^ atmdn, m., 'soul, self; 

S(H*S naman, n., 'name'. Thus: 

"^ ' /' 

Singular : 

N. ^TsTT tjirHi *rm 

raja atma nama 

rajanam atmanam nama 

rajna atmdna namna 

rajne atmdne namne 

raj'nas atmdnas namnas 




rajni, rajani atmdni 



N. A. V. 


I.D.Ab. {(sc-UM^ 


G. L. fiifr^ 


N. ^isiMH^ 

raj anas 



D. Ab. { |sl^6| H^ 











namni, namani 
naman, nama 

namnl, namani 

atmdbhyam namabhyam 








The weakest cases of murdhdn, m., ^head', would be ac- 
cented murdhna, murdhne, murd/mos, murdhnds (ace. pi.), murdhnam, 
etc. ; and so in all similar cases (loc. sing., murdhni or murdhdni]. 

425. Vedic Irregularities, a. Here, as elsewhere, the ending of 
the nom.-acc.-voc. du. masc. is usually a instead of aw. 

b. The briefer form (with ejected a) of the loc. sing., and of the neut. 
nom. etc. du., is almost unknown to the older language. RV. writes once 
$atadavni, but it is to be read fataddvani; and a few similar cases occur in 
AV. In the Brahmanas also, such forms as dhamani and samani are much 
more common than such as ahni and lomni. 

c. But throughout both Veda and Brahmana, an abbreviated form of the 
loc. sing., with the ending i omitted, or identical with the stem, is common 
(in RV., of considerably more frequent occurrence than the regular form): 
thus, murdhdn, kdrman, ddhvan, beside murdhdni etc. ' 


d. In the nom.-acc. pi. neut., also, an abbreviated form is common, 
ending in a or (twice as often) a, instead of ani : thus, brdhma and brdhma, 
beside brdhmani : compare the similar series of endings from a-stems, 329. 

e. From a few stems in man is made an abbreviated inst. sing., with 
loss of m as well as of a : thus, mahina, prathina, varina, dana, prena, &/mna, 
for mahimnd etc. And draghmd (RV., once) is perhaps (Grassmann) for 

f. Other of the weakest cases than the loc. sing, are sometimes found 
with the a of the suffix retained : thus, for example, bhtimana, ddmane, 
ydmanas, uksdnas (accus. pi.), etc. In the infinitive datives trdmane, 
vidmdne, ddvdne, etc. the a always remains. Still more numerous are 
the instances in which the a, omitted in the written form of the text, is, 
as the metre shows, to be restored in reading. 

g. The voc. sing, in vas, which is the usual Vedic form from stems 
in vant (below, 453 b), is found also from a few in van, perhaps by a 
transfer to the uarU-declension: thus, rtauas, evayavas, khidvas (?), prataritvas, 
mdtarifvas, vibhavas. 

For words of which the a is not made long in the strong cases, see the 
next paragraph, b. 

426. A few steins do not make the regular lengthening of 
a in the strong cases (except the nom. sing.). Thus : 

a. The names of divinities, pusdn and aryamdn: thus, pusa, pusdnam, 
puma, etc. 

b. In the Veda, uksdn, 'bull' (but RV. uksdnam once); yosan, 'maiden': 
vfsan, 'virile, bull' (but vfsanam and vfsanas are also met with); fmdn, 
abbreviation of dtmdn ; and two or three other scattering forms : anarvdnam, 
jemana. And in a number of additional instances, the Vedic metre seems 
to demand a where a is written. 

427. The stems qvdn, m., 'dog', and yuvan, 'young', have 
in the weakest cases the contracted form cun and yhn (with 
retention of the accent) ; in the strong and middle cases they 
are regular. Thus, cvti, cvftnam, cuna, cune, etc., cvdbhyam, 
cvabhis, etc. 

In dual, RV. has once yUna for yuvana. 

428. The stem maghdvan, 'generous' (later, almost exclusively 
a name of Indra), is contracted in the weakest cases to maghon: 
thus, maghdva, maghdvanam, maghona, maghone, etc. 

The RV. has once the weak form maghdnas in nom. pi. 

Parallel with this is found the stem maghdvant (division E); and from 
the latter alone in the older language are made the middle cases: thus, 
rnaghdvadbhis, maghdvatsu, etc. (not maghdvabhis etc.). 

429. A number of an-stems are more or less defective, 
making a part of their forms from other stems. Thus : 


430. a. The stem dhan, n., 'day', is in the later language 
used only in the strong and weakest cases, the middle (with 
the nom. sing., which usually follows their analogy) coming 
from dhar or ahas : namely, dhar nom. -ace. sing., dhobhyam, 
dhobhis, etc. (PB. has aharbhis}; but dhna etc., dhni or dhani, 
dhni or dhani, dhani (and, in V., aha). In composition, only 
ahar or ahas is used as preceding member ; as final member, 
ahar, ahas, ahan, or the derivatives aha, alma. 

In the oldest language, the middle cases dhabhis, dhabhyas, dhasu 
also occur. 

b. The stem Udhan, n., 'udder', exchanges in like manner, in the old 
language, with udhar and udhas, but has become later an as-stern only 
(except in the fern, udhnl of adjective compounds): thus, udhar or udhas 
(so RV. : AV. only the latter), udhnas, udhan or udhani, udhabhis, udhahsu. 

431. The neuter stems aksdn, 'eye', asthdn, 'bone', dadhdn, 
'curd', sakthdn, 'thigh', form in the later language only the 
weakest cases, aksna, asthne, dadhnds, sakthni or sakthdni, and 
so on ; the rest of the inflection is made from stems in i, dksi 
etc. : see above, 343 f. 

In the older language, other cases from the an-stems occur : thus, 
aksdni . and aksdbhis ; asthdni, asthdbhis, and asthdbhyas , sakthdni. 

432. The neuter stems asdn, 'blood', yakdn, 'liver', cakdn r 
'ordure', asdn, 'mouth', uddn, 'water', dosdn, 'fore-arm', yusdn, 
'broth', are required to make their nom.-acc.-voc. in all numbers 
from the parallel stems dsrj, ydkrt, qdkrt, asya, udaka (in older 
language udakd], dos, yusa, which are fully inflected. 

Earlier occurs also the dual dosdnl. 

433. The stem pdnthan, m., 'road', is reckoned in the 
later language as making the complete set of strong cases, with 
the irregularity that the nom.-voc. sing, adds a s. The corre- 
sponding middle cases are made from pathi, and the weakest 
from path. Thus : 

from pdnthan pdnthas, pdnthanam; pdnthanau; pdnthanas; 

from pathi pathibhyam; pathibhis, pathlbhyas, pathlsu ; 

from path patha, pathe, pathds, pathi; pathos; pathds 
(accus.), patham. 

In the oldest language (RV.), however, the strong stem is only pdnthd: 
thus, pdnthas, nom. sing. ; pdntham, ace. sing. ; pdnthas, nom. pi. ; and 
even in AV., pdnthanam and pdnthanas are rare compared with the others. 
From patht occur also the nom. pi. pathdyas and gen. pi. pathmdm. RV. 
has once pathds, ace. pi., with long a. 

434. The stems mdnthan, m., 'stirring-stick', and rbhuksdn, m., an 
epithet of Indra, are given by the grammarians the same inflection with 
pdnthan; but only a few cases have been found in use. In V. occur from 


the former the ace. sing, mdntham, and gen. pi. mathlndm (like the cor- 
responding cases from pdnthan) ; from the latter, the nom. sing, rbhuksds and 
voc. pi. rbhuksas, like the corresponding Vedic forms of pdnthan; but also 
the ace. sing, rbhuksdnam and nom. pi. rbhuksdnas, which are after quite 
another model. 


435. Original adjective stems in an are almost exclusively 
those in van, as ydjvan, 'sacrificing', sutvan, 'pressing the soma', 

jitvan, 'conquering'. The stem is masc. and neut. only (two 
or three sporadic cases of its use as fern, occur in RV.); the 
corresponding fern, stem is made in vari: thus, yajvari, jitvari. 

436. Adjective compounds having a noun in an as final 
member are inflected after the model of noun-stems ; and the 
masculine forms are said to be allowed in use also as feminine ; 
but usually a special feminine is made by adding I to the weakest 
form of the masculine stem : thus, durnamnl, somarajm. 

437. But nouns in an occuring as final members of compounds often 
lose the n, or substitute a stem in a for that in an: thus, -raja, -adhva, 
-aha; the corresponding feminine is in a. And feminines in a, replacing 
an. are allowed to be widely formed in the compounds of this division. 

The remaining divisions of the consonantal declension are 
made up of adjective stems only. 

D. Derivative stems (adjective) in in. 

438. The stems of this division are those formed with 
the suffixes ^f in, jrpr min, and f^R vin. They are mascu- 
line and neuter only; the corresponding feminine being 
made by adding ^ i. 

The stems in in are numerous, since almost any noun in a 
in the language may form a possessive derivative adjective with 
this suffix: thus, bdla, 'strength', balin, m. n., balirii, f., 'pos- 
sessing strength, strong'. Stems in vin, however, are very few, 
and those in mm still fewer. 

439. Their inflection is quite regular, except that they 
lose their final ^ n in the middle cases (before an initial 
consonant of the ending), and also in the nom. sing., where 
the masculine lengthens the < i by way of compensation. 
The voc. sing, is in the masculine the bare stem; in the 
neuter, either this or the nominative. 

Whitney, Grammar. 10 




In all these respects, it will be noticed, the w-declension 
agrees with the aw-declension ; it differs from the latter only in 
never losing the vowel of the ending. 

440. Example of inflection. As such may be 
taken srf^FT balm, 'strong'. Thus: 



m. n. 

m. n. 





balk bali 

(khlH GJM 
balinam bait 



balinau balini 

balinas bafini 













sjidHi <MHH! ^M^H^ 

b&lin b&lin, bali balinau balini balinas balini 
The derived feminine stem in ini is inflected, of course, 
like any other feminine in derivative i (365). 

441. There are no irregularities in the inflection of in- 
stems, in either the earlier language or the later except the 
usual* Vedic dual ending in a instead of ait. 

E. Derivative stems (adjective) in ant (or at}. 

442. These stems fall into two sub-divisions: 1. those 
made by the suffix IH ant (or ?IrT at\ being, with a very 
few exceptions, active participles, present and future ; 
2. those made by the possessive suffixes J7H niant and 
Sffi^vant (or T^rnat and cfrT vat}. They are masculine and 
neuter only; the corresponding feminine is made by ad- 
ding ? I. 


1. Participles in ant or at. 

443. The stem has in general a double form, a stronger 
and a weaker, ending respectively in 5JfT ant and EFT ^- 
The former is taken in the strong cases of the masculine, 
with, as usual, the nom.-acc.-voc. pi. neuter; the latter is 
taken by all the remaining cases. 

But, in accordance with the rule for the formation of the feminine stem 
(below, 449), the future participles, and the present participles of verbs of 
the tod-class or accented a-class (752), and of verbs of the ad-class or root- 
class ending in a, are by the grammarians allowed to make the nom.-acc.- 
voc. du. from either the stronger or the weaker stem ; and the present par- 
ticiples from all other present-stems ending in a are required to make the 
same from the strong stem. 

444. Those verbs, however, which in the 3d pi. pres. 
active lose ^ n of the usual ending f^f nti (550) lose it 
also in the present participle, and have no distinction of 
strong and weak stein. 

Such are the verbs forming their present-stem by reduplication without 
added a : namely, those of the reduplicating or /m-class (655) and the inten- 
sives (1012): thus, from yhu, present-stem juhu, participle-stem jtihvat ; inten- 
sive-stem johu, intensive participle-stem jtihvat. Further, the participles of 
roots apparently containing a contracted reduplication : namely, cdksat, ddfat, 
dasat, fffsat, sdfcat] and the aorist participle dhdksat. Vavrdhdnt (RV., once), 
which has the n notwithstanding its reduplication, comes, like the desiderative 
participles (1032), from a stem in a: compare vavrdhdnta, vavrdhdsva. 

Even these verbs are allowed by the grammarians to make the nom.- 
acc.-voc. pi. neut. in anti. 

445. The inflection of these stems is quite regular. 
The nom. sing. masc. comes to end in ER an by the regu- 
lar (150) loss of the two final consonants from the etymo- 
logical form Qff^ants. The vocative of each gender is like 
the nominative. 

446. Steins accented on the final syllable throw the accent 
forward upon the case-ending in the weakest cases (not in the 
middle also). 

In the dual neut. (as in the feminine stem) from such participles, the 
accent is dntl if the n is retained, ati if it is lost. 

447. Examples of declension. As such may serve 





J^ bhdvant, 'being'; 5^ addnt, 'eating'; sT^r^ juhvat, 
'sacrificing'. Thus : 








N. A.V. 
I. D.Ab. 


Singular : / 

^opT ^cfrT 51^*T Ef^rT 

bhdvan bhdvat addn addt juhvat juhvat 

bhdvantam bhdvat addntamaddt juhvatamjujivat 

bhdvata adata juhvata 

bhdvate adate 

bhdvatas adatds 

Hc(fri t^lci 

bhdvati adati 

bhdvan bhdvat 


bhdvantau bhdvanti addntau adati juhvatau juhvati 

bhdvadbhyam adddbhyam juhvadbhyam 


Plural : 

bhdvantas bhdvanti addntas addnti juhvatas juhvati 

bhdvatas bhdvanti adatds addnti juhvatas juhvati 
bhdvadbhis adddbhis juhvadbhis 

bhdvadbhyas adadbhyas juhvadbhyas 



bhdvatam adathm juhvatam 

L. H^Tfg ^f f^rH 

bhdvatsu addtsu juhvatsu 

The future participle bhavisydnt may form in nom. etc. dual 
neuter either bhavisydnti or bhavisyatl; tuddnt, either tuddntl or 
tudati; yhnt (}/ya), either yantl or yatk. And juhvat, in nom. 
etc. plural neuter, may make also juhvanti (beside juhvati, as 
given in the paradigm above). 

But these strong forms (as well as bhdvantl, du., and its like from 
present-stems in unaccented a) are quite contrary to general analogy, and of 
somewhat doubtful character. No Vedic example of them is found; nor have 
they been noticed anywhere in the older language. The cases concerned, 
indeed, would be everywhere of rare occurrence. 

448. The Vedic deviations from the model as above given are few. 
The dual ending du is only one sixth as common as a. Anomalous accent 
is seen in a case or two : acoddte and rathiraydtam. The only instance in 
V. of nom. etc. pi. neut. is sdnti, with lengthened a : compare the forms 
in -manti and -vanti, below, 454 c. 

x 449. The feminine participle-stem, as already stated, 
is made by adding ^ i to either the strong or the weak 
stem-form of the masc.-neut. The rules as to which of the 
two forms shall be taken are the same with those given 
above respecting the nom. etc. dual neuter; namely: 

a. Participles from tense-stems ending in unaccented a 
add i to the strong stem-form, or make their feminine in anfi. 

Such are the bhu or unaccented a-class and the dlv or j/a-class of pre- 
sent-stems (chap. IX.), and the desideratives and causatives (chap. XIV.): 
thus, from ybhu (stem bhdva), bhdvantl; from ydlv (stem divyd), divyantl; 
from bubhusa and bhdvdya (desid. and caus. of |/&/m), bubhusanti and 

Exceptions to this rule are rare. RV. has tdksati and jaratl; Bopp 
(Gr., 530) quotes a few cases from the Nala. The AV. jlvanti, with irregular 
accent, is doubtless to be regarded as a proper name. 

b. Participles from tense-stems in accented d may add the 
feminine-sign either to the strong or to the weak stem-form, 
or may make their feminines in dnti or in atl (with accent as 
nere noted). 

Such are the present-stems of the tud or accented d-class (chap. IX.), 
the s-futures (chap. XII.), and the denominatives (chap. XIV.): thus, from 
ytud (stem tudd], tuddntl or tudati ; from bhavisyd (fut. of ybhu}, bhavisydnti 
or bhavisyati; from devayd (denom. of dcvd), devaydntl or devayati. 


The forms in anil from this class are the prevailing ones. No future 
fern, participle in ati is quotable from the older language. From pres. -stems 
in d are found there rnjati and sincati (RV.), tudati and pinvati (AV.). 
From denominatives, devayati (RV.), durasyad and fatruyad (AV.). 

Verbs of the ad or root-class (chap. IX.) ending in a are given by the 
grammarians the same option as regards the feminine of the present parti- 
ciple : thus, from >/j/a, ydnti or ydd. The older language affords no example 
of the former, so far as noted. 

c. From other tense-stems than those already specified 
that is to say, from the remaining classes of present-stems and 
from the intensives the feminine is formed in ad (or, if the 
stem be otherwise accented than on the final, in afi) only. 

Thus, adad from yad; jfihvatl from yhu; yunjad from yyuj; wnvad 
from ysu; kurvad from ykr ; krmad from ykri; dedifati from dedif (intens. 
of ydif). 

Exceptions are occasionally met with in the later language, as dvisantl 
(BL), rudantl and kurvanti (N.). And AV. has ydntl once. 

450. A few words are participial in form and inflection, 
though not in meaning. Thus : 

a. brhdnt (often written vrhdnt in the later language), 'great' : 
it is inflected like a participle (with brhati and brhdnti in du. 
and pi. neut.). 

b. mah&nt, 'great' ; inflected like a participle, but with the 
irregularity that the a of the ending is lengthened in the strong 
forms: thus, mahun, mah&ntam; mahantau (neut. mahad); mah- 
Untas, mahanti: instr. mahata etc. 

c. pfsant, 'speckled', and (in Veda only) rfifani, 'shining'. 

d. jdgat, 'movable, lively' (in the later language, as neuter noun, 'world'), 
a reduplicated formation from ygam, 'go'; its nom. etc. neut. pi. is allowed 
by the grammarians to be only jdganti. 

e. rhdnt, 'small' (only once, in RV., rhate). 

All these form their feminine in ati only : thus, brhati, 
mahad, prsafi and rucatl (contrary to the rule for participles), 

For ddnt, 'tooth', which is perhaps of participial origin, see above, 396. 

451. The pronominal adjectives iyant and Myant are in- 
flected like adjectives in mant and vant, having (452) iyan and 
kiyan as nom. masc. sing., iyafi and My ati as nom. etc. du. 
neut. and as feminine stems, and iyanti and kiyanti as nom. 
etc. plur. neut. 

But the neut. pi. fyanti and the loc. sing.(?) Mydti are found in RV 

2. Possess! ves in mant and vant. 

452. The adjectives formed by these two suffixes are 




inflected precisely alike, and very nearly like the participles 
in fffl^ant. From the latter they differ only by lengthening 
the 51 a in the nom. sing. masc. 

The voc. sing, is in an, like that of the participle (in the 
later language, namely : for that of the oldest, see below, 454 b). 
The neut. nom. etc. are in the dual only aft (or dfi], and in 
the plural anti (or dnti). 

The feminine is always made from the weak stem : thus, 
mati, vatl (or mdti, vdti}. 

The accent, however, is never (as in the participle) thrown 
forward upon the case-ending or the feminine ending. 

453. To illustrate the inflection of such stems it will 
be sufficient to give a part of the forms of H^MHTI paqumdnt, 
'possessing cattle', and ^PToftT Ihdgavant, 'fortunate, blessed'. 

Singular : 











pacumdntam pacumdt 

bhagavantam bhdgavat 







pacuman pdcumat 


bhdgavan bhdgavat 

o o 

pacumdntau pacumdfi 


bhdgavantau bhdgavati 

Plural : 

pacumdntas pacumdnti 

pacumdtas pacumdnti 

bJidgavantas bhdgavanti 
bhdgavatas bhdgavanti 





454. Vedic Irregularities, a. In dual masc. nom. etc., a (for 
au) is the greatly prevailing ending. 

b. In voc. sing, masc., the ending in the oldest language (RV.) is 
almost always in as instead of an (as in the perfect participle : below, 462 a) : 
thus, adrivas, harivas, bhanumas, havismas. Such vocatives in RV. occur 
more than a hundred times, while not a single unquestionable instance of 
one in an is to be found. In the other Vedic texts, vocatives in as are 
extremely rare (but bhagavas and its contraction bhagos are met with, even 
in the later language); and in their reproduction of RV. passages the as is 
usually changed to an. 

It was pointed out above (425 g) that the RV. makes the voc. in as 
also apparently from a few an-stems. 

c. In RV., the nom. etc. pi. neut., in the only two instances that 
occur, ends in dnti instead of anti: thus, ghrtdvanti, papumdnti. No such 
forms have been noted elsewhere in the older language: the SV. reads anti 
in its version of the corresponding passages, and a few examples of the same 
ending are quotable from the Brahmanas : thus, tavanti, etdvanti, ydvanti, 
pravanti, rtumanti, yugmanti. Compare 448, 451. 

d. In a few (eight or ten) more or less doubtful cases, a confusion of 
strong and weak forms of stem is made : they are too purely sporadic to 
require reporting. The same is true of a case or two where a masculine 
form appears to be used with a feminine noun (see Lanman). 

455. The stem dryant, 'running, steed', has the nom. sing. 
drva, from drvan; and in the older language also the voc. arvan 
and accus. drvanam. 

456. Besides the participle bhdvant, there is another stem 
bhavant, frequently used in respectful address as substitute for 
the pronoun of the second person (but construed, of course, 
with a verb in the third person), which is formed with the suffix 
vant, and so declined, having in the nom. sing, bhdvan; and 
the contracted form bhos of its old-style vocative bhavas is a 
common exclamation of address : 'you, sir !' Its origin is variously 
explained ; it is most probably a contraction of bhdgavant. 

457. The pronominal adjectives tdvant, etdvant, ydvant, and the Vedic 
ivant, mdvant, tvdvant, etc., are inflected like ordinary derivatives from nouns. 

F. Perfect Participles in vans. 

458. The active participles of the perfect tense-system 
are quite peculiar as regards the modifications of their stem. 
In the strong cases, including the nom. -ace. -voc. pi. neut., 
the form of their suffix is SfftT v ans, which becomes, by regu- 
lar process (150), van in the nom. sing., and which is 




shortened to 3FT van in the voc. sing. In the weakest cas- 
es, the suffix is contracted into 3^T us. In the middle cas- 

x * 

es, including the nom.-acc.-voc. neut. sing., it is changed 
to ofrT vat. 

A union-vowel i, if present in the strong and middle cases, 
disappears in the weakest, before MS. 

459. The forms as thus described are masculine and 
neuter only; the corresponding feminine is made by adding 
| I to the weakest form of stem, ending thus in 3^ft usl. 

460. The accent is always upon the suffix, whatever be 
its form. 

r 461. Examples of inflection. To show the in- 
flection of these participles, we may take the stems fi^TH 
vidvahs, 'knowing' (which has irregular loss of the usual 
reduplication and of the perfect meaning) from }/T^" vid; 
and crfTSJoTfa tasthivans, 'having stood', from j/FSTT stha. 
Singular: r^ 

m. n. m. n. 




vidvan vidvdt 

RteiHH^ fen^ 

vidv&hsam vidvdt 

tasthivan tasthivdt 

tasthivUiisam tasthivdt 



Ab. G. 







vidvan vidvat 



tdsthivan tdsthivat 


vidvahsau vidusl 

tasthivansau tasthusi 


I. D. Ab. 

vidvddbhyam tasthivddbhyam 

G. L. 

vidusos tasthusos 

Plnral : 

N.v. ^ 

vidvtinsas vidvfthsi tasthivansas tasthivaiisi 

A. f^fJTO^ Mdift fl^H^ crfr^lftr 

vidusas vidvfrhsi tasthusas tasthivnhsi 


vidvadbhis tasthivddbhis 

D. Ab. 

vidvddbhyas tasthivddbhyas 

G. fa^MI^ rT^IH^ 

vidusam tasthusam 


vidvdtsu tasthivdtsu 

The feminine stems of these two participles are 
vidusl and ri w$ tasthusl. 


Other examples of the different stems are : 
from yAr cakrvtihs, cakrvdt, cakrus, cakrusz ; 
from |/n niriivfins, niriivdt, ninyus, ninyusi ; 
from |/MM babhuvUhs, babhuvdt, babhuvus, babhuvusl ; 
from |/ tan tenivahs, tenivdt, tenus, tenusi. 

462. a. In the oldest language (RV.), the vocative sing. masc. (like 
that of vant and mant-stems : above, 454 b) has the ending vas instead of 
van.' thus, cikitvas (changed to -van in a parallel passage of AV.), titirvas, 
didivas, nudhvas. 

b. Forms from the middle stem, in vat, are extremely rare earlier: 
only three (tatanvdt and vavrtvo'J, neut. sing., and jagrvddbhis, instr. pi.), 
are found in RV., and not one in AV. And in the Veda the weakest stem 
and not the middle one, as later, is made the basis of comparison: thus, 
vidtistara, mldhtistama. 

c. An example or two of the use of the weak stem-form for cases 
regularly made from the strong are found in RV. : they are cakrtisam, ace. 
sing., and dbibhyums, nom. pi. ; emuam, by its accent (unless an error), is 
rather from a derivative stem emusd : and QB. has prostisam. Similar in- 
stances, especially from vidvdhs, are now and then met with later (see BR., 
under vidvcihs). 




d. The AV. has once bhaktivdhsas , as if a participial form from a noun ; 
but K. and TB. give in the corresponding passage bhaktivdnas ; cakhvdhsam 
(RV., once) is of doubtful character; okivdhsa (RV., once) shows a reversion 
to guttural form of the final of YUC, elsewhere unknown. 

G. Comparatives in yas. 

463. The comparative adjectives of primary formation 
(below, 467) have a double form of stem for masculine and 
neuter : a stronger, ending in TXfaj/ahs (usually 

in the strong cases, and a weaker, in EJTT yets (or 
in the weak cases (there being no distinction of middle and 
weakest). The voc. sing. masc. ends in VF(^yan (but for 
the older language see below, 465 a). 

The feminine is made by adding ^ i to the weak masc.- 
neut. stem. 

464. As models of inflection, it will be sufficient to 

give a part of the forms of 3TOT $reyas, 'better', and of 
gdriyas, 'heavier'. Thus : 

Singular : 






















gdriyahsau gdrlyasi 

Plural : 






gdriyansas gdriyahsi 



creyasas creyansi gariyasas gdriyansi 



creyobhis gdriyobhis 

etc. etc. 

The feminine stems of these adjectives are 44 tin) greya 
and J|(ldHl gdriyasl. 

465. a. The Vedic voc. masc. (as in the two preceding divisions : 
454 b, 462 a) is in yas instead of yan: thus, ojiyas, jyayas (RV. : no ex- 
amples elsewhere have been noted). 

b. No example of a middle case occurs in RV. or AV. 

c. In the later language are found a few apparent examples of strong 
cases made from the weaker stem-form: thus, kaniyasam, ace. masc., kariiyasau 
du. They are perhaps rather to be viewed as transition-forms to an a- 


466. Derivative adjective stems having a comparative 
and superlative meaning or often also (and more origin- 
ally) a merely intensive value are made either directly 
from roots (by primary derivation), or from other derivative 
or compound stems (by secondary derivation). 

The subject of comparison belongs properly to the chapter of derivation; 
but it stands in such near relation to inflection that it is, in accordance with 
the usual custom in grammars, conveniently and properly enough treated 
briefly here. 

467. The suffixes of primary derivation are ^JH lyas 
for the comparative and ^? isfha for the superlative. The 
root before them is accented, and usually strengthened by 
gunating, if capable of it or, in some cases, by nasali- 
zation or prolongation. They are much more frequently and 
freely used in the oldest language than later; in the class- 
ical Sanskrit, only a limited number of such comparatives 
and superlatives are accepted in use ; and these attach 
themselves in meaning for the most part to other adjectives 
from the same root, which seem to be their corresponding 


positives ; but in part also they are artificially connected 
with other words, unrelated with them in derivation. 

Thus, from, yksip, 'hurl', come ksepiyas and ksepistha, which 
belong in meaning to ksiprd, 'quick' ; from yvr, 'encompass', 
come vdriyas and vdristha, which belong to uru, 'broad' ; while, 
for example, kdniyas and kdnistha are attached by the grammarians 
to yuvan, 'young', or dlpa, 'small' ; and vdrslyas and vdrsistha 
to vrddtid 'old'. 

468. From Veda and Brahmana together, rather more than 
a hundred instances of this primary formation in Iyas and istha 
(in many cases only one of the pair actually occurring) are to 
be quoted. About half of these (in RV., the decided majority) 
belong, in meaning as in form, to the bare root in its adjective 
value, as used especially at the end of compounds, but some- 
times also independently : thus, from y'tap, 'burn', comes tdpist/ia, 
'excessively burning' ; from 1/yaj, 'offer', come ydjiyas and 
ydjistha, 'better and best (or very well) sacrificing' ; from yyudh, 
'fight', comes yodhtyas, 'fighting better' ; in a few instances, 
the simple root is also found used as corresponding positive : 
thus, ju, 'hasty, rapid', with jdviyas and jdvistha. In a little 
class of instances (eight), the root has a preposition prefixed, 
which then takes the accent : thus, agamistha, 'especially coming 
hither' ; vicayistha, 'best clearing away' ; in a couple of cases 
(dcramistha, dparavapistha, dstheyas), the negative particle is pre- 
fixed ; - - in a single word (cdmbhavistha), an element of another 
kind. The words of this formation often take an accusative 
object: thus, ndbhas tdriyan (RV.), 'traversing rapidly the cloud'; 
vrtrdm hdnisthah (RV.), 'best slayer of Vritra'. 

But even in the oldest language appears not infrequently 
the same attachment in meaning to a derivative adjective which 
(as pointed out above) is usual in the later speech. Besides the 
examples that occur also later, others are met with like vdristha, 
'choicest' (vdra, 'choice'), bdrhistha, 'greatest' (brhdnt, 'great'), 
osistha, 'quickest' (6sam, 'quickly'), and so on. Probably by 
analogy with these, like formations are in a few cases made 
from the apparently radical syllables of words which have no 
otherwise traceable root in the language : thus, kradhiyas and 
kradhistha (K.) from krdhu, sthdviyas and sthdvistha from sthurd, 
cdciyas (RV.) from cdcvant, dmyas (AV.) and dnistha (TS.) from 
anil; and so on. And yet again, in a few exceptional cases, 
the suffixes iyas and istha are applied to stems which are them- 
selves palpably derivative : thus, acistha from acu (RV. : only 
case), tiksniyas (AV.) from tiksnd, brdhmiyas and brdhmistha (TS.) 
from brahman, dhdrmistha (TA.) from dhdrman, drddhistha (TB. : 


instead of ddrhistha] from drdhd, rdghiyas (TS.) from raghu. 
These are beginnings, not followed up later, of the extension 
of the formation to unlimited use. 

In ndviyas or navyas and navistha, from ndva, 'new', and 
in sdnyas from sana, 'old 1 (all EV.), we have also formations 
unconnected with verbal roots. 

469. The stems in istha are inflected like ordinary adject- 
ives in a, making their feminines in a; those in lyas have a 
peculiar declension, which has been described above (463 ff.). 

470. Of peculiarities and irregularities of formation, the 
following may be noticed. 

The suffix lyas has in a few instances the briefer form yas, generally 
as alternative with the other: thus, tdvlyas and tdvyas, ndviyas and n&vyas, 
vdsiyas and vdsyas, pdnlyas and pdnyas; and so from rabh and sah; sdnyas 
occurs alone. From bhu come bhuyas and bhuyistha, beside which RV. has 
also bhdviyas. 

Of roots in a, the final blends with the initial of the suffix to e: thus, 
stheyas, dhestha, yestha; but such forms are in the Veda generally to be 
resolved, as dhd'istha, ydutha. The root jya forms jytetha, but jydyas (like 

The two roots in z, prl and pn, form preyas and prestha and pret/as 
and frestha. 

From the root of rjti come, without strengthening, rjlyas and r/i^fca ; 
but in the older language also, more regularly, rdjiyas and rdjistha. 

471. The suffixes of secondary derivation are rlj tara 
and rFT tama. They are of almost unrestricted application, 
being added to adjectives of every form, simple and com- 
pound, ending in vowels or in consonants and this from 
the earliest period of the language until the latest. The 
accent of the primitive remains (with rare exceptions) un- 
changed ; and that form of stem is generally taken which 
appears before an initial consonant of a case-ending (weak 
or middle form). 

Examples (of older as well as later occurrence) are : from 
vowel-stems, priydtara, vdhnitama, rathitara and rathitama (RV.), 
carutara, potrtama; from consonant-stems, cdmtama, cdcvattama, 
tavdstara and tavdstama, tuvisfama, vdpustara, tapasvitara, bhdga- 
vattara, hiranyavafimattama ; from compounds, ratnadh&tama , 
abhibhutam, sukfttara, purbhittami, bhuridavattara, cucivratatama, 

474] COMPARISON. 159 

But in the Veda the final n of a stem is regularly retained: thus, 
madfntara and madfntama, vrsdntama ; and of a perfect participle the weakest 
stem is taken : thus, vidtistara, midhustama. A feminine final I is shortened : 
thus, devitama (RV.), tejasvinitama (K.). 

In the older language, the words of this formation are not much more 
frequent than those of the other: thus, in RV. the stems in tara and tama 
are to those in lyas and istha as three to two; in AV., only as six to five: 
but later the former win a great preponderance. 

472. These comparatives and superlatives are inflected like 
ordinary adjectives in a, forming their feminine in a. 

473. That (especially in the Veda) some stems which are 
nouns rather than adjectives form derivatives of comparison is 
natural enough, considering the uncertain nature of the division- 
line between substantive and adjective value. Thus, we have 
matftama, nrtama, maruttama, and others. 

The suffixes tara and tama also make forms of comparison 
from some of the pronominal roots, as ka, ya, i (see below, 
520) ; and from certain of the prepositions, as ut; and the ad- 
verbial accusative (older, neuter; later, feminine) of a compar- 
ative in tara from a preposition is used to make a corresponding 
comparative to the preposition itself (below, 11 lie). 

The Hindu grammarians even allow the suffixes of comparison in the 
adverbial accusative feminine, taram and tamam, to be appended to con- 
jugational forms: thus, pacati, 'he cooks', pacatitaram, 'he cooks better': 
but such are barbarous combinations, having no warrant in the earlier uses 
of the language. 

The suffixes of secondary comparison are occasionally added to those of 
primary, forming double comparatives and superlatives : thus, gariyastara, 

The use of tama as ordinal suffix is noted below (487); 
with this value, it is accented on the final, and makes its femin- 
ine in i: thus, catatamd, m. and n., catatamk, f., 'hundredth'. 

474. From a few words, mostly prepositions, degrees of 
comparison are made by the briefer suffixes ra and ma: thus, 
ddhara and adhamd, dpara and apamd, dvara and avamd, upara 
and upamd, dntara, dntama, paramd, madhyamd, caramd. And 
ma is also used to make ordinals (below, 487). 





475. The simple cardinal numerals for the first ten 
numbers (which are the foundation of the whole class), 
with their derivatives, the tens, and with some of the high- 
er members of the decimal series, are as follows: 











ddca ' 











70 Httfri 


























The accent sajrta and a*fd is that belonging to these words in all 
accentuated texts; according to the grammarians, they are sdpta and dsta in 
the later language. See below, 483. 

The series of decimal numbers may be carried still further ; 
but there are great differences among the different authorities 

476] ODD NUMBERS. 161 

with regard to their names ; and there is more or less of discord- 
ance even from ayuta on. 

Thus, in the TS., we find ayuta, niyuta, prayuta, drbuda, nyhrbuda, 
samudrd, mddhya, dnta, parardhd; K. reverses the order of niyuta and 
prayuta, and inserts badva after nyarbuda (reading nyarbudha): these are 
probably the oldest recorded series. 

In modern time, the only numbers in practical nse above 'thousand' are 
laksa ('lac' or lakh 1 ) and koti ('crore'); and an Indian sum is wont to be 
pointed thus: 123,45,67,890, to signify '123 crores, 45 lakhs, 67 thousands, 
eight hundred and ninety'. 

As to the stem-forms pancan etc., see below, 484. As to the form 
saks instead of sas, see above, 146 end. The stem dva appears in com- 
position and derivation also as dva and dvi; catiir in composition is accented 
cdtur. The older form of asta is asta : see below, 483. Forms in -fat and 
-fati for the tens are occasionally interchanged. 

The other numbers are expressed by the various composition 
and syntactical combination of those given above. Thus : 

476. The odd numbers between the even tens are made 
by prefixing the (accented) unit to the ten to which its value 
is to be added: but with various irregularities. Thus: 

eka in '11' becomes eka, but is elsewhere unchanged; 

dva becomes everywhere dva; but in '42' '72' and in '92' it is inter- 
changeable with dvi, and in '82' dvi alone is used; 

for tri is substituted its nom. pi. masc. trdyas; but tri itself is also 
allowed in '43' '73' and in '93', and in '83' tri alone is used; 

sas becomes so in '16', and makes the initial d of dafa lingual (199 b); 
elsewhere its final undergoes the regular conversion (226 b) to * or d or n; 
and in '96' the n of navati is assimilated to it (199b); 

asta becomes [asta (483) in '18' '38', and has either form in the 
succeeding combinations. Thus : 

n ekada^a 

31 ekatrihfat 

61 ekasasti 

81 ekafiti 


12 dvdda?a 

32 dvdtrihcat 


82 dvyafiti 

13 trdyodafa 

33 trdyastrihfat 


83 tryaflti 

14 cdturdafa 

34 cdtustrih$at 

64 cdtuhsasti 

84 cdturafiti 

15 pdncadaya 

35 pdncatrififat 

65 pancasasii 

85 pdncaflti 

16 so da fa 

36 sdttrihfat 

66 sdtsasti 

86 sddaflti 

n saptddapa 

37 saptdtrinfat 

67 saptdsasti 

87 sapid flti 

18 astddafa 

38 astdtrintat 

fi< '.* r" " 

88 astdflti 

19 ndvadafa 

39 ndvatrihfat 

69 ndvasasti 

89 ndvaflti. 

The numbers 

'21' '29' are made 

like those for '31' '39'; 

the numbers 

'41'_'49' } '51'_<59', '71' '79', and '91' '99' are made like those for 
'61' '69'. 

Whitney, Grammar. 11 

162 VI. NUMERALS. [476 

The forms made with dvd and trayas are more usual than those with 
dvi and tri, which are hardly to be quoted from the older literature (V. and 
Br.). The forms made with asta (instead of asta) are alone found in the 
older literature (483), and are usual in the later. 

477. The above are the normal expressions for the odd 
numbers. But equivalent substitutes for them are also variously 
made. Thus : 

a. By use of the adjectives una, 'deficient', and adhika, 'redundant', 
in composition with lesser numbers which are to be subtracted or added, and 
either independently qualifying or (more usually) in composition with larger 
numbers which are to be increased or diminished by the others : thus, tryuna- 
sasti, 'sixty deficient by three' (i. e. '57'); astadhikanavati, 'ninety increased 
by eight' (i. e. '98'); ekddhikam patam, 'a hundred increased by one' (i. e. 
'101'); panconam catam, '100 less 5' (i. e. '95'). For the nines, especially, 
such substitutes as ekonavihcati, '20 less 1', or '19', are not uncommon; 
and later the eka, T, is left off, and unavihcati etc. have the same value. 

b. A case-form of eka, 'one', is connected by na, 'not', with a larger 
number from which one is to be deducted: thus, ekayd nd trincdt (QB. 
PB. KB.), 'not thirty by one' (i. e. '29'); ekasmdn nd pancacdt (in ordinal), 
'49' (TS.); ekasyai (abl. fern.: 366.3) nd pancacdt, '49' (TS.); most often, 
ekan (i. e. ekdt, irregular abl. for ekasmdt) nd vincatf, '19'; ekdn nd catdm, 
*99'. This last form is admitted also in the later language : the others are 
found in the Brahmanas. 

c. Instances of multiplication by a prefixed number are occasionally met 
with: thus, trisaptd, 'thrice seven'; trinavd, 'thrice nine'; tridacd, 'thrice ten'. 

d. Of course, the numbers to be added together may be expressed by 
independent words, with connecting 'and' : thus, ndva ca navatfy ca, or n&va 
navatfm ca, 'ninety and nine'; dvau ca vihcatfc ca, 'two and twenty'. But 
the connective is also (at least, in the older language) not seldom omitted: 
thus, navatfr ndva, '99'; trincdtam trin, '33'; acltir astdti, '88'. 

478. The same methods are also variously used for forming 
the odd numbers above 100. Thus: 

a. The added number is prefixed to the other, and takes the accent: 
for example, ekacatam, '101'; astdcatam, '108'; trincdchatam, '130'; asta- 
vihcaticatam, <148 J ; cdtuhsahasram (RV. : unless the accent is wrong), '1004'. 

b. Or, the number to be added is compounded with adhika, 'redundant', 
and the compound is either made to qualify the other number or is further 
compounded with it: thus, pancddhikarh catam or pancddhikacatam, '105'. 

Of course, una, 'deficient' (as also other words equivalent to una or 
adhika}, may be used in the same way: thus, panconam catam, '95'. 

C. Syntactical combinations are made at convenience : for example, ddca 
catdm ca, '110'; catam ekam ca, 401'. 

479. Another usual method (beginning in the Brahmanas) 
of forming the odd numbers above 100 is to qualify the larger 


number by an adjective derived from the smaller, and identical 
with the briefer ordinal (below, 487) : thus, dvadacdm catdm, 
'112' (lit'ly, 'a hundred of a 12-sort, or characterised by 12 ; 
catuccatvarihcdm catdm, '124'; satsastdm catdm, '166'. 

480. To multiply one number by another, among the higher 
or the lower denominations, the simplest and least ambiguous 
method is to make of the multiplied number a dual or plural, 
qualified by the other as any ordinary noun would be ; and this 
method is a common one in all ages of the language. For ex- 
ample : pdnca pancacdtas, five fifties ('250'); ndva navatdyas, 'nine 
nineties' ('810'); acitibhis tisrbhis, 'with three eighties' ('240'); 
pdnca gatani, 'five hundreds'; trlni sahdsrani, 'three thousands'; 
sastini sahdsrani, '60,000'; daca ca sahasrany astau ca catani, 
'10,800': and, combined with addition, trim cathni trdyastrincatam 
ca, '333'; sahasre dve panconam catam eva ca, '2095'. 

By a peculiar and wholly illogical construction, such a combination as 
trini sastifatani, which ought to signify '480' (3x100 + 60), is frequently 
used in the Brahmanas to mean '360' (3x100 + 60); so also due catustrinfe 
fate, '234' (not '268'); and other like cases. 

481. But the two factors, multiplier and multiplied, are 
also, and in later usage more generally, combined into a com- 
pound (accented on the final); and this is then treated as an 
adjective, qualifying the numbered noun; or else its neuter or 
feminine (in i] singular is used substantively : thus, dacacatas, 
'1000'; satcataih padatibhih (MBh.), 'with 600 foot-soldiers'; 
trdyastrihqat tricatah satsahasrtih (AV.), '6333'; dvicatdm or dmcatk, 
'200'; astadacacatl, '1800'. 

In the usual absence of accentuation, there arises sometimes a question 
as to how a compound number shall be understood: whether asta?atam, for 
example, is astdfatam, '108', or astacatdm, '800', and the like. 

482. Inflection. The inflection of the cardinal nu- 
merals is in many respects irregular. Gender is distinguish- 
ed only by the first four. 

a. Eka, T, is declined after the manner of . a pronominal 
adjective (like sdrva, below, 524); its plural is used in the 
sense of 'some, certain ones'. Its dual does not occur. 

Occasional forms of the ordinary declension are met with : 
thus, eke (loc. sing.), ekat. 

In the late literature, eka is used in the sense of 'a certain', 
or even sometimes almost of 'a', as an indefinite article. Thus, 
eko vyaghrah (H.), 'a certain tiger'; ekasmin dine, 'on a certain 
day'; haste dandam ekamadaya (H.), 'taking a stick in his hand'. 


16 4 VL NUMERALS. [482 

b. Dva, '2', is dual only, and is entirely regular^: thus, 
N.A.V. dvau (V. dva), m., dve, f. n. ; I. D.Ab. dvubhyam; 
G. L. dvdyos. 

c. Tri, '3', is in masc. and neut. nearly regular, like an 
ordinary stem in *; but their genitive is as if from tray a (only 
in the later language : the regular trmhm occurs once in RV.). 
For the feminine it has the peculiar stem tisr, which is inflected 
in general like an r-stem ; but the nom. and accus. are alike, 
and show no strengthening of the r ;. and the r is not prolonged 
in the gen. (excepting in the Veda). Thus : 

m. n. f. 

N. trdyas trini tisrds 

A. trin trini tisrds 

1. tribhfs tisrbhis 

D. Ab. tribhyds tisfbhyas 

G. trayandm tisrndm 

L. tristi tisfsu. 

The Veda has the abbreviated neut. nom. and accus. tri. The accent- 
uation tisrbhfs, tisrbhyds, tisrnam, and tisrsti is said to be also allowed in 
the later language. 

The stem tisr occurs in composition in tisrdhanvd (Br.), 'a bow along 
with three arrows'. 

d. Catur, '4', has catv&r (the more original form) in the 
strong cases ; in the fern, it substitutes the stem catasr, apparently 
akin with tisf, and inflected like it (but with anomalous change 
of accent, like that in the higher numbers : see below, 483). 

m. n. f. 

N. catvdras catvdri cdtasras 

A. cattiras catvdri cdtasras 

+, I. cattirbhis catasrbhis 

D. Ab. cattirbhyas catasfbhyas 

G. caturndm catasrndm 

L. cattirsu catasr su. 

The use of n before am of the gen. masc. and neut. after a final con- 
sonant of the stem is (as in sas: below, 483) a striking irregularity. The 
more regular gen. fern, catasfnam also sometimes occurs. In the later 
language, the accentuation of the final syllable instead of the penult is allow- 
ed in inst., dat.-abl., and loc. 

483. The numbers from '5' to '19' have no distinction of 
gender, nor any generic character. They are inflected, somewhat 
irregularly, as plurals, save in the nom. -ace., where they have 
no proper plural form, but show the bare stem instead. Of sds 
(as of catur\ nam is the gen. ending, with mutual assimilation 
(198b) of stem-final and initial of the termination. Asta (as 

485] INFLECTION. 165 

accented in the older language) has an alternative fuller form, 
asta, which is almost exclusively used in the older literature 
(V. and Br.), both in inflection and in composition (but some 
compounds with asta are found as early as the AV.); its nom.- 
acc. is astd (usual later : found in RV. once, and in AV.), or 
asta (RV.), or astau (most usual in RV.; also in AV., Br., 
and later). 

The accent is in many respects peculiar. In all the accented texts, the 
stress of voice lies on the penult before the endings this, bhyas, and su, 
from the stems in a, whatever be the accent of the stem : thus, pancdbhis 
from pdnca, navdbhyas from ndva, dafdsu from ddfa, navadafdbhis from 
ndvadafa, ekddafdbhyas from ekddafa, dvddafdsu from dvadaya; according to 
the grammarians, either the penult or the final is accented in these forms 
in the later language. In the gen. pi., the accent is on the ending (as in 
that of i, u, and r-stems). The cases of sas, and those made from the stem- 
form asta, have the accent throughout upon the ending. 

Examples of the inflection of these words are as follows : 
N. A. pdnca sat astati astd 

I. pancdbhis sadbhis astabhfs astdbhis 

D.Ab. pancdbhyas sadbhyds astdbhyds astabhyas 

G. pancdndm sanndm astdndm 

L. pancdsu satsu astdsu astdsu. 

Sapid (in the later language sdpta, as dsta for astd) and ndva and ddfa, 
with the compounds of ddfa ('11' '19'), are declined like panca, and with 
the same shift of accent (or with alternative shift to the endings, as pointed 
out above). 

484. The Hindu grammarians give to the stems for '5' and '7' '19' a 
final n: thus, pancan, saptan, astan, navan, dacan, and ekddafan etc. This, 
however, has nothing to do with the demonstrably original final nasal of '7', 
'9', and '10' (compare septem, nouem, decem; seven, nine, ten)', it is only 
owing to the fact that, starting from such a stem-form, their inflection is 
made to assume a more regular aspect, the nom.-acc. having the form of a 
neut. sing, in an, and the instr., dat.-abl., and loc. that of a neut. or masc. 
pi. in an: compare ndma, ndmabhis, ndmabhyas, ndmasu the gen. alone 
being like that, rather, of a a-stem : compare dafdndm with {ndranam and 
ndmnam or atmdnam. No trace whatever of a final n is found anywhere in 
the language, in inflection or derivation or composition, from any of these 

485. The tens, vihqati and trihc&t etc., with their compounds, 
are declined regularly, as feminine stems of the same endings, 
and in all numbers. 

Catd and sahdsra are declined regularly, as neuter (or, rarely, 
in the later language, as masculine) stems of the same final, in 
all numbers. 


The like is true of the higher numbers which have, in- 
deed, no proper numeral character, but are ordinary nouns. 

486. Construction. As regards their construction with 
the nouns enumerated by them : 

a. The words for '!' to '19' are used adjectively, agreeing 
in case, and, if they distinguish gender, in gender also, with 
the nouns : thus, dacdbhir viraih, 'with ten heroes' ; ye deva divy 
ekadaca sthd (AV.), 'what eleven gods of you are in heaven'; 
pancdsu jdnesu, 'among the five tribes'; catasfbhir girbhth, 'with, 
four songs'. 

b. The numerals above '19' are construed usually as nouns, 
either taking the numbered noun as a dependent genitive, or 
standing in the singular in apposition with it : thus, catam dasih 
or catam dasmam, 'a hundred slaves' or 'a hundred of slaves'; 
miwatyn Mrib/iis, 'with twenty bays'; sastyam cardtsu, 'in 60 au- 
tumns'; Catena, pacaih, 'with a hundred fetters'; catam sahdsram 
ayutam nydrbudam jaghhna cakrb ddsyunam (AV.), 'the mighty 
[Indra] slew a hundred, a thousand, a myriad, a hundred mil- 
lion, of demons'. 

Occasionally they are put in the plural, as if used more 
adjectively: thus, pancacadbhir vanaih, 'with fifty arrows'. 

c. In the older language, the numerals for '5' and upward 
are sometimes used in the nom.-acc. form (or as if indeclinably) 
with other cases also : thus, pdnca krstisu, 'among the five races' ; 
saptd rsmam, 'of seven bards'; sahdsram rsibhih, 'with a thousand 
bards'; catam purbhih, 'with a hundred strongholds'. 

487. Ordinals. Of the classes of derivative words 
coming from the original or cardinal numerals, the ordin- 
als are by far the most important; and the mode of their 
formation may best be explained here. 

Some of the first ordinals are irregularly made : thus, 

eka, '!', forms no ordinal; instead is used prathamd (i. e. 
pra-tama, 'foremost'); adi is rare in the Brahmanas, and adya 
even in the Sutras; 

from dvd, '2', and tri, '3', come dvitiya and trtiya (second- 
arily, through dvita and abbreviated trita}', 

catur, '4', sds, '6', and saptd, '?', take the ending tha : 
thus, caturthd, sastM, saptdtha; but for 'fourth' are used also 
turiya and turya, and saptdtha belongs to the older language 
only: pancatha, for 'fifth', is excessively rare; 

the numerals for '5' and '7' usually, and for '8', '9', '10', 
add ma, forming pamama, saptamd, astamd, navamd, daqamd; 


for 41th' to '19th', the forms are ekadacd, dvadacd, and 
so on (the same with the cardinals, except change of accent) ; 

for the tens and intervening odd numbers from '20' onward, 
the ordinal has a double form one made by adding the full 
(superlative) ending tamd to the cardinal : thus, vihcatitamd, 
trincattamd, acititamd, etc. ; the other, shorter, in a, with abbre- 
viation of the cardinal: thus, vihqd, '20th'; trincd, '30th'; ca- 
tvarincd, '40th'; pancacd, '50th'; sastd, '60th'; saptatd, '70th'; 
acitd, '80th'; navatd, '90th'; and so likewise ekavihcd, '21st', 
catustrihcd, '34th'; astacatvarihcd, '48th'; dvapancacd, '52d'; 
ekasastd, '61st'; and ekannavihfd and unavincd and ekonavtncd, 
'19th'; and so on. Of these two forms, the latter and briefer 
is by far the more common, the other being not quotable from 
the Veda, and extremely rarely from the Brahmanas. From '50th' 
on, the briefer form is allowed by the grammarians only to the 
odd numbers, made up of tens and units ; but it is sometimes 
met with, even in the later language, from the simple ten. 

Of the higher numbers, gatd and sahdsra form catatamd and 
sahasratamd ; but their compounds have also the simpler form : 
thus, ekagata, '101st'. 

Of the ordinals, prathamd (and adya\ dvitlya, trtiya, and 
turiya (with turya] form their feminine in a; all the rest make 
it in i. 

488. The ordinals, as in other languages, have other than ordinal offices 
to fill ; and in Sanskrit especially they are general adjectives to the cardinals, 
with a considerable variety of meanings, as fractionals, as signifying 'composed 
of so many parts' or 'so-many-fold', or 'containing so many', or (as was seen 
above, 479) 'having so many added'. 

In a fractional sense, the grammarians direct that their accent be 
shifted to the first syllable : thus, dvtiiya, 'half ; trtlya, 'third part' ; cdturtha, 
'quarter', and so on. But in accented texts only trtlya, 'third', and ttiriya, 
'quarter', are found so treated; for 'half occurs only ardhd; and caturthd, 
pancamd, and so on, are accented as in their ordinal use. 

489. Other numeral derivatives thus, 

multiplicative adverbs, as dvis, tris, catus, 'twice', thrice'., 
'four times'; 

adverbs with the suffixes dha and cas : for example, ekadhh, 
'in oneway', catadhu, 'in a hundred ways'; ekacas, 'one by one', 
catacds, 'by hundreds'; 

collectives, as dvitaya, or dvayd, 'a pair', ddcataya or dacdt, 
'a decade' 
belong rather to the dictionary, or to the chapter of derivation. 

168 [490 



490. THE pronouns differ from the great mass of nouns 
and adjectives chiefly in that they come by derivation from 
another and a very limited set of roots, the so-called 'pro- 
nominal' or 'demonstrative' roots. But they have also many 
and marked peculiarities of inflection some of which, 
however, find analogies also in a few adjectives; and such 
adjectives will accordingly be described at the end of this 

Personal Pronouns. 

491. The pronouns of the first and second persons are 
the most irregular and peculiar of all, being made up of 
fragments coming from various roots and combinations of 
roots. They have no distinction of gender. 

Their inflection in the later language is as follows : 

Singular : 

1st pers. 2d pers. 

N. ^^ OT^ 

ahdm tvdm 

A. TF[^ Jqj ^T^ ^7 

mam, ma tvam, tva 


mdya tvdya 

D. *-|c^|4-|, If ^jV^JTf, (^ 

mdhyam, me tubJiyam, te 

Ab. 3^ ^7^ 

mat tvdt 





I. D. Ab. 



Plural : 








T rR, ^ 


mama, me 





asman, nas 

asmdbhyam, nas 
asmakam, nas 







o ^ 
yusman, vas 

o , -\ 

^JCrjT^JTI 1 , ^{ 
yiismdbhyam, vas 

yusmakam, vas 



The briefer second forms for accus., dat., and gen., in all 
numbers, are accentless ; and hence they are not allowed to 
stand at the beginning of a sentence, or elsewhere where any 
emphasis is laid. 

The ablative mat is accentless in one or two AV. passages (xi. 4.26; 
xii. 3. 46). 

492. Forms of the older language. All the forms 

170 VI1 - PRONOUNS. [492 

given above are found also in the older language ; which, how- 
ever, has also others that afterward disappear from use. 

Thus, the Veda (RV.) has a few times the instr. sing, tva (like manisd 
for manlsdya); further, the Ioc 4 sing, tve, the dat. pi. (less often loc,) asme, 
and the loc. pi. yusme : the final e of these forms is uncombinable (or 
pragrhya: 138 b). The datives in bhyam are in RV. not seldom to be read 
as if in bhya, with loss of the final nasal; asmdkam and yusmdkam suffer 
the same loss only in a rare instance or two. The usual resolutions of 
semivowel to vowel are made, and are especially frequent in the forms of 
the second person (tuam for tvdm, etc.). 

But the duals, above all, wear a very different aspect earlier. In Veda 
and Brahmana, the nominatives are avdm and yuvdm, and only the accusa- 
tives avdm and yuvdm (but in RV. the dual forms of 1st pers. chance not 
to occur, unless in vam[?J, once, for a-uam); the instr. in RV. is either 
yuvdbhyam (not elsewhere found) or yuvabhyam; an abl. yuvdt appears once 
in RV., and avdt twice in TS.; the gen. -loc. is in RV. (only) yuv6s instead 
of yuvdyos. Thus we have here a distinction (elsewhere unknown) of five 
different dual cases by endings, in part accordant with those of the other 
two numbers. 

493. Peculiar endings. The ending am, appearing in the nom. 
sing, and pi. (and Vedic du.) of these pronouns, will be found often, though 
only in sing., among the other pronouns. The bhyam (or hyam) of dat. sing. 
and pi. is met with only here; its relationship with the bhyam, bhyas, bhis 
of the ordinary declension is palpable. The t (or d) of the abl., though here 
preceded by a short vowel, is doubtless the same with that of the a-declension 
of nouns and adjectives. That the nom., dat., and abl. endings should be 
the same in sing, and pi. (and in part in the earlier du. also), only the stem 
to which they are added being different, is unparalleled elsewhere in the 
language. The element sma appearing in the plural forms will be found 
frequent in the inflection of the singular in other pronominal words : in fact, 
the compound stem asma which underlies the plural of aham seems to be 
the same that furnishes part of the singular forms of at/am (501), and its 
value of 'we' to be a specialisation of the meaning 'these persons'. The 
genitives singular, mama and tdva, have no analogies elsewhere; the deri- 
vation from them of the adjectives mamaka and tavaka (below, 516) suggests 
the possibility of their being themselves stereotyped stems. The gen. pi., 
asmakam and yusmdkam, are certainly of this character : namely, neuter sing, 
case-forms of the adjective stems asmdka and yusmaka 1 other cases of which 
are found in the Veda. 

494. Stem-forms. To the Hindu grammarians, the 
stems of the personal pronouns are mad and asmad, and tvad 
and yusmad, because these are forms used to a certain extent, 
and allowed to be indefinitely used, in derivation and compo- 
sition (like tad, kad, etc.: see below, under the other pronouns). 
Words are thus formed from them even in the older language 




namely, mdtkrta, mdtsakhi, asmdtsakhi, tvddyoni, mattds (AV.), 
tvdtpitr (TS.), yuvdddevatya (QB.); but much more numerous are 
those that show the proper stem in a, or with the a lengthened 
to^a: thus,^ mavant ; asmatra, asmadruh, etc.; tvadatta, tvanid, 
tvnvasu, tvtihata, etc.; yusmftdatta, yusmesita, etc.; yuvddhita, 
yuvndatta, yuvUriita, etc. And the later language also has a few 
words made in the same way, as madrc. 

The Vedas have certain more irregular combinations, with complete 
forms : thus, tvamkama, mampapyd, mamasatyd, asmehiti, ahampurvd, aham- 
uttard, aharhyu, ahamsana. 

From the stems of the grammarians come also the deriv- 
ative adjectives madiya, tvadkya, asmadiya, yusmadlya, having a 
possessive value : see below, 516. 

For sva and svaydm, see below, 513. 

Demonstrative Pronouns. 

495. The simplest demonstrative, rT ta, which answers 
also the purpose of a personal pronoun of the third person, 
may be taken as model of a mode of declension usual in 
so many pronouns and pronominal adjectives that it is fairly 
to be called the general pronominal declension. 

But this root has also the special irregularity that in the 
nom. sing. masc. and fern, it has sds (for whose peculiar euphonic 
treatment see 176 a) and sa, instead of ids and ta (compare Gr. 
o, 7], TO, and Goth, sa, so, thata}. Thus: 

Singular : 



















172 VII. PRONOUNS. [495 


tdsya tdsyas 


tdsmin tdsyam 


N.A.V. ftt ft ft 

tau te te 

I. D. Ab. 

thbhyam thbhyam 

G. L. HU^ 



N. ft ftlH 

te thni 

A. ftT^ ftTH 

thn thni 

D. Ab. 

tebhyas thbhyai 


L. ftj 

tesu thsu 

The Vedas show no other irregularities of inflection than those which 
belong to all stems in a and a : namely, tend sometimes ; usually td for tali, 
du. ; often id for tdni, pi. neut. ; usually tebhis for tats, instr. pi. ; and the 
ordinary resolutions. The RV. has one more case-form from the root sa, 
namely sdsmin (occurring nearly half as often as tdsmin). 

496. The peculiarities of the general pronominal declension, 
it will be noticed, are these : 

In the singular, the use of t (properly d) as ending of nom.-acc. neut. ; 
the combination with the root of another element sma in masc. and neut. 
dat., abl., and loc., and of sy in fern, dat., abl.-gen., and loc.; and the 
masc. and neut. loc. ending in, which is restricted to this declension (except 
in the anomalous yddffmin, RV., once). 

The dual is precisely that of noun-stems in a and d. 


In the plural, the irregularities are limited to U for ids in nom. masc., 
and the insertion of s instead of n before am of the gen., the stem-final 
being treated before it in the same manner as before su of the loc. 

497. The stem of this pronoun is by the grammarians given 
as tad; and from that form come, in fact, the derivative ad- 
jective tadiya, with tattvd, tadvat, tanmaya; and numerous com- 
pounds, such as tacchila, tajjna, tatkara, tadanantara, tanmatra, 
etc. These compounds are not rare even in the Veda : so tddanna, 
tadvid, tadvacd, etc. But derivatives from the true root ta are 
also many: especially adverbs, as tatas, tdtra, tatha, tada; the 
adjectives tavant and tdti; and the compound tadfc etc. 

498. Though the demonstrative root ta is prevailingly of 
the third person, it is also freely used, both in the earlier 
language and in the later, as qualifying the pronouns of the 
first and second person, giving emphasis to them : thus, so 'Mm, 
'this I', or 'I here'; sd tvdm, 'thou there'; te vayam, 'we here'; 
and so on. 

499. Two other demonstrative stems appear to contain ta 
as an element; and both, like the simple ta, substitute sa in 
the nom. sing. masc. and fern. 

a. The one, tya, is tolerably common (although only a 
third of its possible forms occur) in RV., but rare in AV., and 
almost unknown later: its nom. sing., in the three genders, is 
syds, sya, tydt, and it makes the accusatives tydm, tyam, tydt, 
and goes on through the remaining cases in the same manner 
as ta. It has in RV. the instr. fern, tya (for tydya]. 

b. The other is the usual demonstrative of nearer position, 
'this here', and is in frequent use through all periods of the 
language. It prefixes e to the simple root, forming the nomin- 
atives esds, esa, etdt and so on through the whole inflection. 

The stem tya has neither compounds nor derivatives. But 
from eta are formed both, in the same manner as from the 
simple ta, only much less numerous : thus, etadda (QB.), etad- 
artha, etc., from the so-called stem etat; and etadfc and etavant 
from eta. 

\ 500. There is a defective pronominal stem, ena, which is 
accentless, and hence used only in situations where no emphasis 
falls upon it. It does not occur elsewhere than in the accusa- 
tive of all numbers, the instr. sing., and the gen. -loc. dual: 

m. n. f. 

Sing. A. enam enat enam 

I. enena enaya 




Du. A. enau ene ene 

G. L. enayos enayos 

PI. A. enan enani enas 

The RV. has enos instead of enayos, and in one or two instances accents 
a form: thus, endm, endsC?). 

This stem forms neither derivatives nor compounds. 

501. Two other demonstrative declensions are so irreg- 
ularly made up that they have to be given in full. The 
one, %|t|H aydm etc., is used as a more indefinite demon- 
strative, 'this' or 'that'; the other, ERTT asau etc., signifies 
especially the remoter relation, 'yon' or 'yonder'. 

They are as follows : 

Singular : 
m. n. 



aydm iddm 
imam iddm 












^TT <JR 
imau ime 












amum adds 







'O *^ 



o _-s 









Plural : 



ime imani 






amun amuni 





D. Ab. 








f t ^ 

amibhyas amubhyas 

amisam amusam 



The same forms are used in the older language, without variation, except 
that ima occurs for imau and imdni. and amu for amuni; amuya when used 
adverbially is accented on the final, amuya; asau (with accent, of course, on 
the first, dsau) is used also as vocative. 

502. The former of these two pronouns, aydm etc., plainly shows itself 
to be pieced together from a number of defective stems. The majority of 
forms come from the root a, with which, as in the ordinary pronominal de- 
clension, sma (f. sy) is combined in the singular. All these forms from a 
have the peculiarity that in their substantive use they are either accented, 
as in the paradigm, or accentless (like ena and the second forms from dhdm 
and tvdm}. The remaining forms are always accented. From and come, with 
entire regularity, anena, andya, andyos. The strong cases in dual and plural, 
and in part in singular, come not less regularly from a stem imd. And 
aydm, iydm, iddm are evidently to be referred to a simple root i (iddm being 
apparently a double form: id, like tad etc., with ending am). 

The Veda has from the root a also the instrumental end and ayd (used 
in general adverbially), and the gen. loc. du. ay6s; from zma, imdsya occurs 
once in RV. The RV. has in a small number of instances the irregular 
accentuation dsmai, dsya, dbhis. 

In analogy with the other pronouns, iddm is by the gram- 
marians regarded as representative stem of this pronominal de- 
clension ; and it is actually found so treated in a very small 
number of compounds (idammdya and icldmrupa are of Brahmana 
age). A.S regards the actual stems, ana furnishes nothing further ; 
from ima comes only the adverb imdtha (RV., once); but a and 
* furnish a number of derivatives, mostly adverbial : thus, for 
example, dtas, dtra, dtha ; itds, id (Vedic particle), ida, ihd, itara, 
tm (Vedic particle), idfc, perhaps evd and evdm, and others. 

176 VII. PRONOUNS. [503 

503. The other pronoun, asati etc., has amft for its leading stem, which 
in the singular takes in combination, like the o-stems, the element sma 
(f. at/), and which shifts to ami in part of the masc. and neut. plural. In 
part, too, like an adjective w-stem, it lengthens its final in the feminine. 
The gen. sing, amtisya is the only example in the language of the ending 
sya added to any other than an a-stem. The nom. pi. ami is unique in 
form; its I is (like that of a dual) pragrhya, or exempt from combination 
with a following vowel (138 b). Asati and adds are also without analogies as 
regards their endings. 

The grammarians, as usual, treat adds as representative 
stem of the declension, and it is found in this character in an 
extremely small number of words, as adomula; adomdya is of 
Brahmana age. The QB. has also asaunaman. But most of 
the derivatives, as of the cases, come from amu: thus, amutas, 
amutra, amutha, amurhi, amuvdt, amuka. 

In the older language occurs the root tva (accentless), meaning 'one, 
many a one' ; it is oftenest found repeated, as 'one' and 'another'. It follows 
the ordinary pronominal declension. 

Fragments of another demonstrative root or two are met with : thus, 
dmas, 'he', occurs in a formula in AV. and in Brahmanas etc. ; av6s as 
gen.-loc. dual is found in RV. ; the particle u points to a root u. 

Interrogative Pronoun. 

504. The characteristic part of the interrogative pro- 
nominal root is R k; it has the three forms ^ ka, fa ki, 
3\ ku; but the whole declensional inflection is from 3\ ka. 

O i 

excepting the nom. -ace. sing, neut., which is from fa ki, 
and has the anomalous form faq kim (not elsewhere known 
in the language from a neuter /-stem). The nom. and 
accus. sing., then, are as follows: 

m. n. f. 

N - ^H faft SfiT 

Ms kim ka 

Mm kim 

and the rest of the declension is precisely like that of rT 
ta (above, 495). 

The Veda has its usual variations, Tea and kebhis for kdni and kais. It 
also has, along with kfm, the pronominally regular neuter kdd; and kdm 


(or kam) is a frequent particle. The masc. form kis, corresponding to kim, 
occurs as a stereotyped case in the combinations ndkis and mdkis. 

505. The grammarians treat kirn as representative stem of 
the interrogative pronoun ; and it is in fact so used in a not 
large number of words, of which a few kimmdya, kimkard, 
kimkamya, kimdevata, and the peculiar kimyti go back even 
to the Veda and Brahmana. In closer analogy with the other 
pronouns, the form kad, a couple of times in the Veda (katpayd. 
kddartha), and not infrequently later, is found as first member of 
compounds. Then, from the real roots ka, ki, ku are made 
many derivatives ; and from ki and ku, especially the latter, many 
compounds : thus, kdti, katha, kathdm, kada, katard, katamd, kdrhi; 
kiyant, kulfc; kutas, kutra, kuha, kvd, kucard, kukarman, ku- 
mantrin, etc. 

506. Various forms of this pronoun, as kad, kim, and ku 
(and, rarely, ko), at the beginning of compounds, have passed 
from an interrogative meaning, through an exclamatory, to the 
value of prefixes signifying an unusual quality either some- 
thing admirable, or, oftener, something contemptible. This use 
begins in the Veda, but becomes much more common in later time. 

507. The interrogative pronoun, as in other languages, 
turns readily in its independent use also to an exclamatory 
meaning. Moreover, it is by various added particles converted 
to an indefinite meaning : thus, by ca, cand, cit, dpi, va, either 
alone or with the relative ya (below, 511) prefixed: thus, kdc 
cand, 'any one'; nd ko 'pi, 'not any one'; yani ktini cit, 'what- 
soever'. Occasionally, the interrogative by itself acquires a simi- 
lar value. 

Relative Pronoun. 

508. The root of the relative pronoun is ET 2/#, which 
from the earliest period of the language has lost all trace 
of the demonstrative meaning originally (doubtless) belonging 
to it, and is used as relative only. 

509. It is inflected with entire regularity according to 

the usual pronominal declension: thus, 

Singular. Dual. Plural, 

m. n. f. m. n. f. m. n. f. 


yds yat ya 

yam ydt yam 

ye yani yas 

yau ye ye 

yan yam yas, 

Whitney, Grammar. 12 

178 VH- PRONOUNS. [509 

yena ydya 

D. ETFl UHT ynbhyam ^ 

ydsmai ydsyai yebhyas yabhyas 

etc. etc. etc. - etc. etc. 

The Veda shows its usual variations of these forms: yd for ydti and for 
ydni, and yebhis for j/afs ,- yos for I/CM/OS also occurs once ; yena, with pro- 
longed final, is in RV. twice as common as yena. Resolutions occur '.in 
ydbhias, and yesaam and ydsaam. 

510. The use of yat as representative stem begins very 
early : we have ydtkama in the Veda, and yatkarin, yaddevatyd 
in the Brahmana ; later it grows more general. From the proper 
root come also a considerable series of derivatives : ydtas, ydti, 
ydtra, ydtha, ydda, yddi, yavant, yatard, yatamd ; and the com- 
pound yadrc. 

511. The combination of ya with ka to make an indefinite 
pronoun has been noticed above (507). Its own repetition 
as ydd-yat gives it sometimes a like meaning, won through 
the distributive. 

512. One or two marked peculiarities in the Sanskrit use 
of the relative may be here briefly noticed : 

a. A very decided preference for putting the relative clause before that 
to which it relates: thus, yah sunvatdh sdkhd tdsma fndraya gayata (RV.), 
'who is the friend of the soma-presser, to that Indra sing ye' ; yam yajnam 
paribhur dsi sd {d devesu gachati (RV.), 'what offering thou protectest, that 
in truth goes to the gods'; ye trisaptdh pariydnti bald tesdm dadhdtu me 
(A.V.), 'what thrice seven go about, .their strength may he assign to me'; 
asdu y6 adhardd grhas tdtra santv ardyyah (AV.), 'what house is yonder in 
the depth, there let the witches be'; sahd ydn me dsti tena (TB.), 'along 
with that which is mine' ; hansdndm vacanam yat tu tan mam dahati (MBh.), 
'but what the words of the swans were, that burns me' ; sarvasya locanam 
castram yasya nd 'sty andha eva sah (H.), 'who does not possess learning, 
the eye of everything, blind indeed is he'. The other arrangement is com- 
paratively unusual. 

b. A frequent conversion of the subject or object of a verb by an added 
relative into a substantive clause : thus, me 'mdm prd "pat pdtiruseyo vadhti 
yah (AV.), 'may there not reach him a human deadly weapon' (lit'ly, 'what 
is such a weapon'); part no pdhi ydd dhdnam (AV.), 'protect of us what 
wealth [there is]'; apamdrgd 'pa mdrstu ksetriydm capdthaf ca ydh (AV.), 
'may the cleansing plant cleanse away the disease and the curse'; puskarena 
hrtarh rdjyam yac cd 'nyad vasu kimcana (MBh.), 'by Pushkara was taken 
away the kingdom and whatever other property [there was]'. 




Emphatic Pronoun. 

513. The isolated and uninflected pronominal word 
t<^H svayam (from the root sva] signifies 'self, own self. 
By its form it appears to be a nom. sing. 7 and it is often- 
est used as nominative, but along with words of all persons 
and numbers; and not seldom it represents other cases also. 

Svayam is also used as a stem in composition : thus, 
svayamja, svayambhu. But sva itself (usually adjective : below, 
516) has the same value in composition : and even its inflected 
forms are (in the older language very rarely) used as reflexive 

Nouns used pronominally. 

514. The noun atmdn, 'soul', is widely employed, in the 
singular, as reflexive pronoun of all three persons. 

The adjective bhavant, f. bhavafi, is used (as already pointed 
out : 456) in respectful address as substitute for the pronoun of 
the second person. Its construction with the verb is in accord- 
ance with its true character, as a word of the third person. 

Pronominal Derivatives. 

515. From pronominal roots and stems, as well as from 
the larger class of roots and from noun-stems, are formed 
by the ordinary suffixes of adjective derivation certain words 
and classes of words, which have thus the character of pro- 
nominal adjectives. 

Some of the more 'important of these may be briefly no- 
ticed here. 

516. Possessives. From the representative stems mad 
etc. are formed the adjectives madiya, asmadiya, tvadiya, yusmadlya, 
tadtiya, and yadkya, which are used in a possessive sense : 'relat- 
ing to me, mine', and so on. 

Other possessives are mamaM (also mdmaka, RV.) and tavakd, 
from the genitives mama and tava. 

An analogous derivative from the genitive amusya is amusyayanti (AV. 
etc.), 'descendant of such a one'. 

It was pointed out above (493) that the "genitives" asmdkam and yu- 
smdkam are really stereotyped cases of possessive adjectives. 


180 VII. PRONOUNS. [516 

Corresponding to svaydm '513) is the possessive sva, mean- 
ing 'own', as relating to all persons and numbers. The RV. 
has once the corresponding simple possessive of the second per- 
son, tvd, 'thy' . 

For the use of sva as reflexive pronoun, see above, 513, end. 

All these words form their feminines in a. 

Other derivatives of a like value have no claim to be mentioned here. 
But (excepting sva) the possessives are so rarely used as to make but a 
small figure in the language, which prefers generally to indicate the possessive 
relation by the genitive case of the pronoun itself. 

517. By the suffix vant are formed from the pronominal 
roots, with prolongation of their final vowels, the adjectives 
mavant, tv&vant, yusmnvant, yuvtivant, tavant, ettivant, ytivant, mean- 
ing 'of my sort, like me', etc. Of these, however, only the 
last three are in use in the later language, in the sense of 
'tantus and 'quantus. They are inflected like other adjective 
stems in vant, making their feminines in vatl (452 if.). 

Words of similar meaning from the roots t and Id are iyant 
and kiyant, inflected in the same manner: see above, 451. 

518. The pronominal roots show a like prolongation of 
vowel in combination with the root drc, 'see, look', and its 
derivatives drca and (quite rarely) dfksa: thus, madrc, madrca ; 
asmadrca: tvadrg etc.; yusmadfc etc.; tadfc etc.; etadfc, etadfca, 
etadfksa (VS.); yadfc etc.; idfc and Iddfc etc. They mean 'of my 
sort, like or resembling me', and the like, and the last five are 
not uncommon, with the sense of l talij and 'qualis 1 . The forms 
in drc are unvaried for gender ; those in drga (and drksa ?] have 
feminines in t. 

519. From to, ka, ya come tdti, 'so many', Mti, 'how 
many?' ydti, 'as many'. They have a quasi-numeral character, 
and are inflected (like the numerals pdnca etc.: above, 483) only 
in the plural, and with the bare stem as nom. and accus.: thus, 
N.A. tdti; I. etc. tatibhis, t&tibhyas, tdtmam, tdtisu. ' 

520. From ya (in V. and Br.) and ka come the compara- 
tives and superlatives yatard and yatamd, and katard and katamd ; 
and from i, the comparative itara. For their inflection, see 
below, 523. 

521. Derivatives with the suffix ka, sometimes conveying 
a diminutive or a contemptuous meaning, are made from certain 
of the pronominal roots and stems (and may, according to the 
grammarians, be made from them all): thus, from fa, takam, 




takdt, takas; from sa, saka : from ya, yakds, yaka, yake : from 
asau, asakau: from amu, amuka. 

For the numerous and frequently used adverbs formed from pronominal 
roots, see Adverbs (below, chapter XVI.). 

Adjectives declined pronominally. 

522. A number of adjectives some of them coming 
from pronominal roots, others more or less analogous with 
pronouns in use are inflected, in part or wholly, accord- 
ing to the pronominal declension (like cT ta. 495 y . Thus : 

523. The comparatives and superlatives from pronominal 
roots namely, katard and katamd, yatard and yatamd, and itara ; 
also anyd, 'other', and its comparative anyatard are declined 
like ta throughout. Their feminine stems are in a. 

But even from these words forms made according to the adjective de- 
clension are sporadically met with (e. g. itarayam, K.). 

524. Others words are so inflected except in the nom.- 
acc.-voc. sing, neut., where they have the ordinary adjective 
form am, instead of the pronominal at (ad). Such are sdrva, 
'all', vtcva, 'all, every', eka, 'one'. 

These, also, are not without exception, at least in the earlier language 
(e. g. vfyvaya, vfyvat, RV. ; eke loc. sing., AV.). 

525. Yet other words follow the same model usually, or 
in some of their significations, or optionally ; but in other senses, 
or without known rule, lapse into the adjective inflection. 

Such are the comparatives and superlatives from prepositional stems: 
ddhara and adhamd, dntara and dntama, dpara and apamd, dvara and avamd, 
uttara and uttamd, upara and upamd. Of these, pronominal forms are de- 
cidedly more numerous from the comparatives than from the superlatives. 

Further, the superlatives (without corresponding comparatives) paramd. 
caramd, madhyamd; and also anyatama (whose positive and comparative belong 
to the class first mentioned: 523). 

Further, the words para, 'distant, other' ; purva, 'prior, east' ; ddksina, 
'right, south'; ubhdya (f. ubhdyl or ubhayi), 'of both kinds or parties'; and 
the rare sama (accentless), 'any or every one 1 , simd, 'each, all', nema, 'the 
one, half ; and the possessive svd. 

526. Occasional forms of the pronominal declension are met with from 
numeral adjectives: e. g. prathamdsyas, trtiyasydm; and from other words 
having an indefinite numeral character: thus, dlpa, 'few': ardhd, 'half; 
kevala, 'all'; dvtiaya, 'of the two kinds' and others. 

182 [527- 



f 527. THE subject of conjugation or verbal inflection 
involves, as in the other languages of the family, the dis- 
tinctions of voice, tense, mode, number, and person. 

Then, besides the simpler or ordinary conjugation of 
a verbal root, there are certain more or less fully developed 
secondary or derivative conjugations. 

528. Voice. There are (as in Greek) two voices, active 
and middle, distinguished by a difference in the personal 
endings. This distinction is a pervading one: there is no 
active personal form which does not have its corresponding 
middle, and vice versa; and it is extended also in part to 
the participles (but not to the infinitive). 

529. An active form is called by the Hindu grammarians 
parasmai padavn, 'a word for another', and a middle form is 
called atmane padam, 'a word for one's self: the terms might be 
best paraphrased by 'transitive' and 'reflexive'. And the distinc- 
tion thus expressed is doubtless the original foundation of the 
difference of active and middle forms : in the recorded condition 
of the language, however, the antithesis of transitive and re- 
flexive meaning is in no small measure blurred, or even alto- 
gether effaced. 

530. Some verbs are conjugated in both voices, others 
in one only; sometimes a part of the tenses are inflected 
only in one voice, others only in the other or in both; of 
a verb usually inflected in one voice sporadic forms of the 
other occur; and sometimes the voice differs according as 
the verb is compounded with certain prepositions. 

531. The middle forms outside the present-system (for 
which there is a special passive inflection: see below. 768), 


and sometimes also within that system, are liahle to be used 
likewise in a passive sense. 

532. Tense. The tenses are as follows: 1. a present, 
with 2. an imperfect, closely related with it in form, having 
a prefixed augment; 3. a perfect, made with reduplication 
(to which in the Veda is added, 4. a so-called pluperfect, 
made from it with prefixed augment); 5. an aorist, of three 
different formations : a. simple; b. reduplicated; c. sibilant; 
6. a future, with 7. a conditional, an augment-tense, stand- 
ing to it in the relation of an imperfect to a present; and 
8. a second, a periphrastic, future (not found in the Veda). 

The tenses here distinguished (in accordance with prevailing 
usage) as imperfect, perfect, pluperfect, and aorist receive those 
names from their correspondence in mode of formation with tenses 
so called in other languages of the family, especially in Greek, 
and not at all from differences of time designated by them. In 
no period of the Sanskrit language is there any expression of 
imperfect or pluperfect time nor of perfect time, except in 
the older language, where the "aorist" has this value ; later, im- 
perfect, perfect, and aorist (of rare use) are so many undiscrim- 
inated past tenses or preterits : see below, under the different 

533. Mode. In respect to mode, the difference be- 
tween the classical Sanskrit and the older language of the 
Veda and, in a less degree, of the Brahmanas is 
especially great. 

In the Veda, the present tense has, besides its indicative 
inflection, a subjunctive, of considerable variety of formation, 
an optative, and an imperative (in 2d and 3d persons). The 
same three modes are found, though of much less frequent oc- 
currence, as belonging to the perfect; and they are made also 
from the aorists, being of especial frequency from the simple 
aorist. The future has no modes (an occasional case or two 
are purely exceptional). 

In the classical Sanskrit, the present adds to its indic- 
ative an optative and an imperative of which last, more- 
over, the first persons are a remnant of the old subjunc- 


tive. And the aorist has also an optative, of somewhat 
peculiar inflection, usually called the precative (or bene- 

534. The present, perfect, and future tenses have each 
of them, alike in the earlier and later language, a pair of 
participles, active and middle, sharing in the various pe- 
culiarities of the tense-formations ; and in the Veda are 
found such participles belonging also to the aorist. 

535. Tense-systems. The tenses, then, with their 
accompanying modes and participles, fall into certain well- 
marked groups or systems: 

I. The present-system, composed of the pres- 
ent tense with its modes, its participle, and its pret- 
erit which we have called the imperfect. 

II. The perfect-system, composed of the per- 
fect tense (with, in the Veda, its modes and its preterit, 
the so-called pluperfect) and its participle. 

III. The aorist-system, or systems, simple, 
reduplicated, and sibilant, composed of the aorist tense 
along with, in the later language, its "pnecative" opta- 
tive (but. in the Veda, with its various modes and its 

IV. The future-systems: a. the old or sibi- 
lant future, with its accompanying preterit, the condi- 
tional, and its participle; and b. the new periphrastic 

536. Number and Person. The verb has, of course, 
the same three numbers with the noun, namely singular, 
dual, and plural ; and in each number it has the three per- 
sons, first, second, and third. All of these are made in 
every tense and mode except that the first persons of 
the imperative numbers are supplied from the subjunctive. 


537. Verbal adjectives and nouns: Partici- 
ples. The participles belonging to the tense-systems have 
been already spoken of above (534). There is besides, com- 
ing directly from the root of the verb, a participle, prevail- 
ingly of past and passive (or sometimes neuter) meaning. 
Future passive participles, or gerundives, of several different 
formations, are also made. 

538. Infinitives. In the older language, a very con- 
siderable variety of derivative abstract nouns only in a 
few sporadic instances having anything to do with the tense- 
systems are used in an infinitive or quasi -infinitive 
sense ; most often in the dative case, but sometimes also in 
the accusative, in the genitive and ablative, and (very rarely) 
in the locative. In the classical Sanskrit, there remains a 
single infinitive, of accusative case-form, having nothing 
to do with the tense-systems. 

539. Gerund. A so-called gerund (or absolutive) - 
being, like the infinitive, a stereotyped case-form of a de- 
rivative noun is a part of the general verb-system in 
both the earlier and later language, being especially fre- 
quent in the latter. In the Veda it has a somewhat various 
form ; in the later language, it has only two forms, one for 
simple verbs, and the other for compound. Its value is that 
of an indeclinable active participle, of indeterminate but 
prevailingly past tense-character. 

A second gerund, an adverbially used accusative in form, 
is found, but only rarely, both earlier and later. 

540. Secondary conjugations. The secondary or 
derivative conjugations are as follows: a. the passive; b. the 
intensive; c. the desiderative; d. the causative. In these, 
a conjugation-stem, instead of the simple root, underlies 
the whole system of inflection. Yet there is clearly to be 
seen in them the character of a present-system, expanded 



into a more or less complete conjugation ; and the passive 
is so purely a present-system that it will be described in 
the chapter devoted to that part of the inflection of the 

Under the same general head belongs the subject of 
denominative conjugation, or the conversion of noun and 
adjective-stems into conjugation-stems. Further, that of 
compound conjugation, whether by the prefixion of prepo- 
sitions to roots or by the addition of auxiliary verbs to noun 
and adjective-stems. And finally, that of periphrastic con- 
jugation, or the looser combination of auxiliaries with verbal 
nouns and adjectives. 

541. The characteristic of a proper (finite or personal) 
verb-form is its personal ending. By this alone is deter- 
mined its character as regards number and person and 
in part also as regards mode and tense. But the distinc- 
tions of mode and tense are mainly made by the formation 
of tense and mode-stems, to which, instead of to the pure 
root, the personal endings are appended. 

In this chapter will be given a general account of the per- 
sonal endings, and also of the formation of mode-stems from 
tense-stems, and of those elements in the formation of tense- 
stems the augment and the reduplication which are found 
in more than one tense-system. Then, in the following chapters, 
each tense-system will be taken up by itself, and the methods 
of formation of its stems, both tense-stems and mode-stems, and 
their combination with the endings, will be described in detail. 

Personal Endings. 

542. The endings of verbal inflection are, as was pointed 
out above, different throughout in the active and middle voices. 
They are also, as in Greek, usually of two somewhat varying 
forms for the same person in the same voice : one fuller, called 
primary ; the other briefer, called secondary. There are also less 
pervading differences, depending upon other conditions. 

A condensed statement of all the varieties of ending for each person 
and number here follows. 


543. Singular: First person. The primary ending in 
the active is mi. The subjunctive, however (later imperative), 
has ni instead ; and in the oldest Veda this ni is sometimes 
wanting, and the person ends in a (as if the ni of am were 
dropped). The secondary ending is m; and to this m an a has 
come to be so persistently prefixed, appearing always where the 
tense-stem does not itself end in a (vam for varm or varam in 
RV., once, is an isolated anomaly), that it is necessary to reckon 
am as ending, alternate with m. But the perfect tense has 
neither mi nor m ; its ending is simply a (sometimes a : 248 c) : 
or, from -roots, au. 

The primary middle ending, according to the analogy of 
the other persons, would be regularly me. But no tense or 
mode, at any period of the language, shows any relic whatever 
of a m in this person : the primary ending, present as well as 
perfect, from a-stems and others alike, is e; and to it corre- 
sponds * as secondary ending, which blends with the final of an 
a-stem to e. The optative has, however, a instead of i; and 
in the subjunctive (later imperative) appears ai for *e. 

544. Second person. In the active, the primary ending 
is si, which is shortened to s as secondary : as to the loss of 
this s after a final radical consonant, see below, 555. But the 
perfect and the imperative desert here entirely the analogy of 
the other forms. The perfect ending is invariably tha (or tha, 
248 c). The imperative is far less regular. The fullest form 
of its ending is dhi; which, however, is more often reduced to 
hi; and in the great majority of verbs (including all a-stems, 
at every period of the language) no ending is present, but the 
bare stem stands as personal form. In a very small class of 
verbs (722) ana is the ending. The Veda has also an ending 
tat; and this is even used sporadically in other persons of the 
imperative (see below, 570 1). 

In the middle voice, the primary ending, both present and 
perfect, is se. The secondary stands in no apparent relation to 
this, being thas; and in the imperative is found only sva (or 
sva: 248 c), which in the Veda is not seldom to be read as ma. 
In the older language, se is sometimes strengthened to sai in the 

545. Third person. The active primary ending is ti; 
the secondary, t: as to the loss of the latter after a final radical 
consonant, see below, 555. But in the imperative appears in- 
stead the peculiar ending tu ; and in the perfect no characteristic 
consonant is present, and the third person has the same ending 
as the first. 


The primary middle ending is te, with ta as corresponding 
secondary. In the older language, te is often strengthened to 
tai in the subjunctive. In the perfect, the middle third person 
has, like the active, the same ending with the first, namely e 
simply; and in the older language, the third person present also 
often loses the distinctive part of its termination, and comes to 
coincide in form with the first. To this e perhaps corresponds, 
as secondary, the t of the aorist 3d pers. passive (842 ff.). The 
imperative has tarn (or, in the Veda, rarely am] for its ending. 

546. Dual: First person. Both in active and in middle, 
the dual first person is in all its varieties precisely like the 
corresponding plural, only with substitution of v for the m of the 
latter : thus, vas (no vasi has been found to occur), va, vahe, 
vahi, vahai. The person is, of course, of comparatively rare 
use, and from the Veda no form in pas, even, is quotable. 

547. Second and Third persons. In the active, the 
primary ending of the second person is tkas, and that of the 
third is tas; and this relation of th to t appears also in the 
perfect, and runs through the whole series of middle endings. 
The perfect endings are primary, but have u instead of a as 
vowel ; and an a has become so persistently prefixed that their 
forms have to be reckoned as athus and aim. The secondary 
endings exhibit no definable relation to the primary in these 
two persons ; they are tarn and tarn ; and they are used in the 
imperative as well. 

In the middle, a long a which, however, with the final 
a of a-stems becomes e has become prefixed to all dual 
endings of the second and third persons, so as to form an in- 
separable part of them. The primary endings, present and per- 
fect, are athe and cite; the secondary (and imperative) are atham 
and atam (or, with stem-final a, ethe etc.). 

The Rig -Veda has a very few forms in aithe and aite, apparently from 
ethe and ete with subjunctive strengthening (they are all detailed below : 
see 615, 701, 737, 752, 836, 1008, 1043). 

548. Plural: First person. The earliest form of the 
active ending is mast, which in the oldest language is more 
frequent than the briefer mas (in RV., as five to one; in AV., 
however, only as three to four). In the classical Sanskrit, mas 
is the exclusive primary ending; but the secondary abbreviated 
ma belongs also to the perfect and the subjunctive (imperative). 
In the Veda, ma often becomes ma (248 c), especially in the 

The primary middle ending is make. This is lightened in 
the secondary form to mahi; and, on the other hand, it is regu- 


larly (in the Veda, not invariably) strengthened to mahai in the 
subjunctive (imperative). 

549. Second person. The active primary ending is tha . 
The secondary, also imperative, ending is ta (in the Veda, ta only 
once in impv.j. But in the perfect any characteristic consonant is 
wanting, and the ending is simply a. In the Veda, the syllable 
na, of problematic origin, is not infrequently added to both 
forms of the ending, making ihana (rarely thana] and tana. The 
forms in which this occurs will be detailed below, under the 
different formations : the addition is very rarely made excepting 
to persons of the first general conjugation. 

The middle primary ending is dhve, which belongs to the 
perfect as well as the present. In the subjunctive of the older 
language it is sometimes strengthened to dhvai. The secondary 
(and imperative) ending is dhvam (in RV. , once dhva] ; and 
dhvat is once met with in the imperative (570). In the Veda, 
the v of all these endings is sometimes resolved into u, and the 
ending becomes dissyllabic. 

550. Third person. The full primary ending is anti in 
the active, with ante as corresponding middle. The middle sec- 
ondary ending is anta, to which should correspond an active 
ant; but of the t only altogether questionable traces are left, 
in the euphonic treatment of a final n (207) ; the ending is an. 
In the imperative, antu and antam take the place of anti and 
ante. The initial a of all these endings is like that of am in 
the 1st sing., disappearing after the final a of a tense- stem. 

Moreover, anti, antu, ante, antam, anta are all liable to be 
weakened by the loss of their nasal, becoming ati etc. In the 
active, this weakening takes place only after reduplicated stems 
(and after a few roots which are treated as if reduplicated : 
639 ff.); in the middle, it occurs after all tense-stems save those 
ending in a. 

Further, for the secondary active ending an there is a sub- 
stitute tts (or ur : 169, end), which is used in the same reduplicating 
verbs that change anti to ati etc., and which accordingly appears 
as a weaker correlative of an. The same us is also used uni- 
versally in the perfect, in the optative (not in the subjunctive), 
in those forms of the aorist whose stem does not end in a, and 
in the imperfect of roots ending in a, and a few others (621). 

The perfect middle has in all periods of the language the 
peculiar ending re, and the optative has the allied ran, in this 
person. In the Veda, a variety of other endings containing a r 
as distinctive consonant are met with : namely, re (and ire] and 
rate in the present ; rata in the optative (both of present and of 




aorist) ; rire in the perfect ; ranta, ran, and ram in aorists (and 
in an imperfect or two) ; ram and ratam in the imperative. The 
three rate, ratam, and rata are found even in the later language 
in one or two verbs (629). 

551. Below are given, for convenience, in tabular form, 
the schemes of endings as accepted in the classical or later 
language : namely, a. the regular primary endings, used in the 
present indicative and the future (and the subjunctive in part) ; 
and b. the regular secondary endings, used in the imperfect, 
the conditional, the aorist, the optative (and the subjunctive in 
part) ; and further, of special schemes, c. the perfect endings 
(chiefly primary, especially in the middle) ; and d. the imperative 
endings (chiefly secondary). To the so-called imperative endings 
of the first person is prefixed the a which is practically a part 
of them ; though really containing the mode-sign of the sub- 
junctive from which they are derived. 

552. Further, a part of the endings are marked with an 
accent, and a part are left unaccented. The latter are those 
which never, under any circumstances, receive the accent ; the 
former are accented in considerable classes of verbs, though by 
no means in all. It will be noticed that, in general, the un- 
accented endings are those of the singular active ; but the 2d 
sing, imperative has an accented ending ; and, on the other 
hand, the whole series of 1st persons imperative, active and 
middle, have unaccented endings (this being a characteristic of 
the subjunctive formation which they represent). 

553. The schemes of normal endings, then, are as follows : 

a. Primary Endings. 















dnti, dti 













b. Secondary Endings. 
ma f, d 

an, us 


c. Perfect Endings. 
md e 

d se 

Us t 


d. p. 

vdhe make 

&the dhve 

dte ante, ate 







dnta, dta, ran 





d. Imperative Endings. 

1 ani dva dma ai avahdi dmahdi 

2 dhi, 7if, tarn ta svd dthdm dhvdm 

3 tu tdm dntu, dtu tdm dtdm dntam, dtdm 

554. In general, the rule is followed that an accented ending, if dis- 
syllabic, is accented on its first syllable and the constant union-vowels 
are regarded, in this respect, as integral parts of the endings. But the 
3d pi. ending ate of the pres. indie, middle has in RV. the accent ate in 
a number of verbs (see 613, 685, 699, 718;; and an occasional instance 
is met with in other endings : thus, make (see 719, 735 . 

555. The secondary endings of the second and third persons singular, 
as consisting of an added consonant without vowel, should regularly (150) 
be lost whenever the root or stem to which they are to be added itself ends 
in a consonant. And this rule is in general followed; yet not without ex- 
ceptions. Thus : 

a. A root ending in a dental mute sometimes drops this final mute 
instead of the added s in the second person ; and, on the other hand, a root 
or stem ending in s sometimes drops this s instead of the added t in the 
third person in either case, establishing the ordinary relation of s and t 
in these persons, instead of s and s, or t and t. A similar loss of any 
other final consonant before the ending is exceedingly rare. For instances, 
see below, 692. 

b. Again, a union-vowel is sometimes introduced before the ending, 
either a or I: see below, 621, 631, 819, 880. 

In a few isolated cases in the older language, this I is changed to ai : 
see below, 904 b, 1068. 

556. The changes of form which roots and stems undergo 
in their combinations with these endings will be pointed out in 
detail below, under the various formations. Here may be simply 
mentioned in advance, as by far the most important among them, 
a distinction of stronger and weaker form of stem in large classes 
of verbs, standing in relation with the accent the stem being 
of stronger form when the accent falls upon it, or before an 
accentless ending, and of weaker form when the accent is on 
.the ending. 

Of the endings marked as accented in the scheme, the ta of 2d pi. is 
not infrequently in the Veda treated as unaccented, the tone resting on the 
stem, which is strengthened. Much less often, the tarn of 2d du. is treated 
in the same way : other endings, only sporadically. 

Subjunctive Mode. 

557. Of the subjunctive mode (as was pointed out above) 
only fragments are left in the later or classical language : namely, 
in the so-called first persons imperative, and in the use (58O) 


of the imperfect and aorist persons without augment after mn 
prohibitive. In the oldest period, however, it was a very fre- 
quent formation, being three or four times as common as the 
optative in the Rig- Veda, and nearly the same in the Atharvan ; 
but already in the Brahmanas it becomes comparatively rare. 
Its varieties of form are considerable, and sometimes perplexing. 

558. In its most normal and regular formation, a special 
mode-stem is made for the subjunctive by adding to the tense- 
stem an a which combines with a final a of the tense-stem 
to a. The accent rests upon the tense-stem, which accordingly 
has the strong form. Thus, from the strong present-stem dok 
(yduh) is made the subjunctive-stem doha; from juho (yhu), 

juhdva; from yunaj C^yuj), yundja ; from bh&va (ybhu), bhdva; 
from tudd (ytud), tudti; from ucyd (pass., yvac), ucyti; and so on. 

559. The stem thus formed is inflected in general as an 
a-stem would be inflected in the indicative, with constant accent, 
and a for a before the endings of the first person (733) but 
with the following peculiarities as to ending etc. : 

560. In the active, the 1st sing, has ni as ending: thus, ddhani, 
yundjdni, bhdvdni. But in the Rig- Veda sometimes a simply : thus, at/a, 

In 1st du., 1st pi., and 3d pi., the endings are always the secondary: 
thus, dohdva, dtihdma, d6han; bhdvdva, bhdvdma, bhdvdn. 

In 2d and 3d du. and 2d pi., the endings are always primary: thus, 
ddhathas, ddhatas, dohatha ; bhdvdthas, bhdvdtas, bhdvdtha. 

In 2d and 3d sing., the endings are either primary or secondary: thus,. 
dtihasi or dohas, d6hati or ddhat; bhdvasi or bhdvds, bhdvdti or bhdvat. 

Occasionally, forms with double mode-sign a (by assimilation to the 
more numerous subjunctives from tense-stems in a) are met with from non- 
a-stems : thus, dsdtha from as ; at/as, at/at, dydn from e 

561. In the middle, forms with secondary instead of primary endings 
are very rare, being found only in the 3d pi. (where they are more frequent 
than the primary), and in a case or two of the 3d sing. 

The striking peculiarity of subjunctive middle inflection is the frequent 
strengthening of e'to di in the endings. This is less general in the very 
earliest language than later. In 1st sing., di alone is found as ending, even 
in RV.; and in 1st du. also (of rare occurrence), only dvahdi is met with. 
In 1st pi., dmahdi prevails in RV. and AV. (dmahe is found a few times), 
and is alone known later. In 2d sing., sdi for se does not occur in RV., 
but is the only form in AV. and the Brahmanas. In 3d sing., tdi for te 
occurs once in RV., and is the predominant form in AV., and the only one 
later. In 2d pi., dhvdi for dhve is found in one word in RV., and a few 
times in the Brahmanas. In 3d pi., nidi for nte is the Brahmana form (of 
far from frequent occurrence); it occurs neither in RV. nor AV. No such 


dual endings as thai and tai, for the arid te, are anywhere found; but RV. 
has in a few words (nine: above, 527, end) aithe and aite, which appear 
to be a like subjunctive strengthening of ethe and ete (although found in one 
indicative form, krnvdite}. Before the ai-endings the penultimate vowel is 
regularly long a; but antdi instead of antdi is two or three times met with, 
and once (TS.) atdi for dtdi. 

562. The subjunctive endings, then, in combination with 
the subjunctive mode-sign, are as follows : 

active. middle. 

8. d. p. s. d. p. 

(dvahdi (dmahdi 

1 am ova ama ai 

[avahe [amahe 

(asi (ase (adhve 

2 i athas atha { aithe 

[as [asai [adhvai 

(ati (ate (ante, anta 

3 atas an aite <_ 
[at [atai [antdi 

And, in further combination with final a of a tense-stem, 
the initial a of all these endings becomes a : thus, for example, 
in 2d pers. : asi or as, athas, atha, ase, adhve. 

563. Besides this proper subjunctive, with mode-sign, in its triple 
form with primary, with strengthened primary, and with secondary end- 
ings there is in the older language another, without mode-sign and with 
secondary endings, or in all respects coinciding with the forms of an augment- 
tense (imperfect or aorist) save for the absence of the augment. Subjunc- 
tives of this character are frequent in RV., decidedly less common in later 
Vedic, and very little used in the Brahmanas except after md prohibitive 
(580) after which they stand also in the later language. 

These forms are sometimes called "imperfect subjunctive", but the 
appellation is an evident misnomer: "improper subjunctive" is preferable. 
Since (below, 587) the forms of augmented tenses are also freely used in 
an indicative sense without augment in the oldest Veda, the distinction of 
the two classes of use is often difficult to make. 

As to the uses of the subjunctive, see below, 572 ff. 

Optative Mode. 

564. As has been already pointed out, the optative is of 
comparatively rare occurrence in the language of the Vedas ; 
but it gains rapidly in frequency, and already in the Brahmanas 
greatly outnumbers the subjunctive, which still later it comes 
almost entirely to replace. 

Its mode of formation is the same in all periods of the 
language . 

Whitney, Grammar. 13 




565. The optative mode-sign is in the active voice a dif- 
ferent one, according as it is added to a tense-stem ending in a, 
or in some other final. In the latter case, it is ya, accented ; 
this ya is appended to the weaker form of the tense-stem, and 
takes the regular series of secondary endings, with, in 3d plur., 
MS instead of an, and loss of the a before it. After an a-stem, 
it is , unaccented ; this i blends with the final a to e (which 
then is accented or not according to the accent of the a) ; and 
the e is maintained unchanged before a vowel-ending (am, us), 
by means of an interposed euphonic y. 

In the middle voice, the mode-sign is i throughout, and 
takes the secondary endings, with a in 1st sing., and ran in 
3d pi. After an a-stem, the rules as to its combination to e, 
the accent of the latter, and its retention before a vowel-ending 
with interposition of a y, are the same as in the active. After 
any other final, the weaker form of stem is taken, and the ac- 
cent is on the ending (except in one class of verbs, where it 
falls upon the tense-stem : see 645) ; and the I (as when com- 
bined to e) takes an inserted y before a vowel-ending. 

It is, of course, impossible to tell from the form whether i or 1 is 
combined with the final of an a-stem to e; but no good reason appears to 
exist for assuming i, rather than the I which shows itself in the other class 
of stems in middle voice. 

566. The combined mode-sign and endings of the optative, 
then, are as follows, in their double form, for a-stems and for 
others : 

\ a. for non-a-stems. 
































b. combined with the final of a-stems. 
eva ema eya evahi 




etam eta ethas eyatham 

etam eyus eta eyatdm 

The ya is in the Veda not seldom resolved into ia. 

The Vedic 3d pi. middle forms in rata will be detailed below, under the 
various formations. 

567. Precative. Precative forms are such as have a 
sibilant inserted between the optative-sign and the ending. They 
are made almost only from the aorist stems, and, though allowed 
by the grammarians to be formed from every root the active 

570] OPTATIVE MODE. 195 

precative from the simple aorist, the middle from the sibilant 
aorist are practically of rare occurrence at every period of the 
language, and especially later. 

The inserted s runs in the active through the whole series of persons ; 
in the middle, it is allowed only in the 2d and 3d persons sing, and du., 
and the 2d pi., and is quotable from the older literature only for the 2d and 
3d sing. In the 2d sing, act., the precative form, by reason of the nec- 
essary loss of the added s, is not distinguishable from the simple optative; 
in the 3d sing, act., the same is the case in the later language, which 
(above, 555) saves the personal ending t instead of the precative-sign s; 
but the RV. usually, and the other Vedic texts to some extent, have the 
proper ending yds (for yast). 

The accent is as in the simple optative. 

568. The precative endings, then, accepted in the later 
language (including, in brackets, those which are identical with 
the simple optative), are as follows : 

active. middle, 

s. d. p. s. d. p. 

1 yasam yasva yasma [iy&] [wdhi] [Imdhi] 

2 [yas] yastam yasta isihas lyasthdm idhvdm 

3 [ydt] ydstdm ydsus istd iydstdm [Iran] 

As to the uses of the optative, see below, 572 ff. 

Imperative Mode. 

569. The imperative has no mode-sign; it is made by 
adding its own endings directly to the tense-stem, just as 
the other endings are added to form the indicative tenses. 

Hence, in 2d and 3d du. and 2d pi., its forms are indistinguishable 
from those of the augment-pretexit from the same stem with its augment 

The rules as to the use of the different endings especially in 2d sing., 
where the variety is considerable will be given below, in connection with 
the -various tense-systems. The ending tat, however, has so much that is 
peculiar in its use that it calls for a little explanation here. 

570. The Imperative in tat. This is nowhere a frequent form, 
though found from the earliest period of the language, and allowed to be 
made in the latest ; in the five leading Vedic texts it is formed from nearly 
fifty verbs, and has less than seventy occurrences. Its usual value is that 
of a second person singular; but it occurs as 1st sing, once, in AV. (dvyu- 
sdm jdgrtdd ahdm, 'let me watch till day-break'); as 3d sing., toward a dozen 
times (e. g. punar md "vi?atdd rayih, TS., 'let wealth come again to me'; 
aydrh tydsya rdjd murdhdnarh m pdtayatdt, $B., 'the king here shall make 



his head fly off'); and as 2d pi. several times in TS. (e. g. apah . . . devesu 
nah sukrto brutdt, 'ye waters, announce us to the gods as well-doers'), and 
many times in a Brahmana passage (repeated in K. xvi. 21. TB. iii. 6. 1, and 
AB. ii. 6, 7) in which, moreover, two authorities (K. and AB.) have once 
varayadhvdt for vdrayatdt: no other occurrence of dhvat has been noted. 

571. As regards its meaning, this form has been shown 
(Delbriick) to have prevailingly in the Brahmanas, and traceably 
but much less distinctly in the Vedic texts, a specific tense-value 
added to its mode-value as signifying, namely, an injunction 
to be carried out at a later time than the present : it is (like 
the Latin forms in to and tote] a posterior or future imperative. 
Examples are : ihai J vd ma tfsthantam abhyehi J ti bruhi tdm Iti na dgatdm 
pratiprdbrutat (QB.), 'say to her "come to me as I stand just here", and 
[afterward] announce her to us as having come'; ydd urdhvds tfsthd drdvine 
'hd dhattdt (RV.), 'when thou shalt stand upright, [then] bestow riches here' 
(and similarly in many cases); utkulam udvah.6 bhavo 'dtihya prdti dhdvatdt 
(AV.), 'be a carrier up the ascent; after having carried up, run back again'; 
vdnaspdtir ddhi tvd sthdsyati tdsya vittdt (TS.), 'the tree will ascend thee, 
take [then] note of it'. 

According to the grammarians, the form in tat may be used when bene- 
diction is intended. 

Uses of the Modes. 

572. Of the three modes, the imperative is the one 
most distinct and limited in office, and most unchanged in 
use throughout the whole history of the language. It signi- 
fies a command or injunction an attempt at the exercise 
of the speaker's will upon some one or something outside 
of himself. 

This, however (in Sanskrit as in other languages), is by 
no means always of the same force ; the command shades off 
into a demand, an exhortation, an entreaty, an expression of 
earnest desire. The imperative also sometimes signifies an as- 
sumption or concession ; and occasionally, by pregnant construc- 
tion, it becomes the expression of something conditional or 
contingent; but it does not acquire any regular use in depend- 
ent-clause-making . 

573. The optative appears to have as its primary office 
the expression of wish or desire; in the oldest language, 
its prevailing use in independent clauses is that to which 
the name "optative" properly belongs. 




The so-called precative forms (567, are restricted to this 
use, but are not otherwise distinguished from the simple optatives. 

'But the expression of desire, on the one hand, passes nat- 
urally over into that of request or entreaty, so that the optative 
becomes a softened imperative ; and, on the other hand, it comes 
to signify what is generally desirable or proper, what should or 
ought to be, and so becomes the mode of prescription ; or, yet 
again, it is weakened into signifying what may or can be, what 
is likely or usual, and so becomes at last a softened statement 
of what is. 

Further, the optative in dependent clauses, with relative 
pronouns and conjunctions, becomes a regular means of ex- 
pression of the conditional and contingent, in a wide and in- 
creasing variety of uses. 

574. The subjunctive, as has been pointed out, be- 
comes nearly extinct at an early period in the history of the 
language; there are left of it in classical usage only two 
relics ; the use of its first persons in an "imperative" sense, 
or to signify a necessity or obligation resting on the speak- 
er, or a peremptory intention on his part ; and the use of 
its other persons, with the negative particle m md, in a 
prohibitive or negative imperative sense. 

And the general value of the subjunctive from the begin- 
ning was what these relics would seem to indicate : its funda- 
mental meaning is perhaps that of requisition, less peremptory 
than the imperative, more so than the optative. But this meaning 
is liable to the same modifications and transitions with that of 
the optative ; and subjunctive and optative run closely parallel 
with one another in the oldest language in their use in in- 
dependent clauses, and are hardly distinguishable in dependent. 
And instead of their being (as in Greek) both maintained in 
use, and endowed with nicer and more distinctive values, the 
subjunctive gradually disappears, and the optative assumes alone 
the offices formerly shared by both. 

575. The difference, then, between imperative and sub- 
junctive and optative, in their fundamental and most char- 
acteristic uses, is one of degree : command, requisition, 
wish : and no sharp line of division exists between them ; 
they are more or less exchangeable with one another, and 
combinable in coordinate clauses. 


Thus, in AV., we have in impv.: catdm jlva carddah, 'do 
thou live a hundred autumns'; ubhau tau jivatam jarddasti, 'let 
them both live to attain old age'; -- in subj., adyd jivani, 'let 
me live this day' ; catdm jivati carddah, 'he shall live a hundred 
autumns'; - -- in opt., jlvema cardddm catani, 'may we live hund- 
reds of autumns'; sdrvam hyur jivyasam (prec.j, 'I would fain 
live out my whole term of life'. Here the modes would be 
interchangeable with a hardly perceptible change of meaning. 

Examples, again, of different modes in coordinate construc- 
tion are : iydm agne nari pdtim videsta . . . suvana putran mdhisl 
bhavati gatva pdtim subhdga m rajatu (AV.), 'may this woman, 
O Agni ! find a spouse ; giving birth to sons she shall become 
a chieftainesg ; having attained a spouse let her rule in happiness' ; 
gopayd nah svastdye prabudhe nah punar dadah (TS.), 'watch over 
us for our welfare, grant unto us to wake again'; syhn nah 
sunuh . . . sa te sumatir bhutv asme (RV.), 'may there be to us 
a son ; let that favor of thine be ours' . It is not very seldom 
the case that versions of the same passage in different texts show 
different modes as various readings. 

There is, in fact, nothing in the earliest employment of 
these modes to prove that they might not all be specialized uses 
of forms originally equivalent having, for instance, a general 
future meaning. 

576. As examples of the less characteristic use of sub- 
junctive and optative in the older language, in independent 
clauses, may be quoted the following : a gha ta gachan uttara 
yughni (RV.), 'those later ages will doubtless come'; ydd . . . nd 
mara iti mdnyase (RV.), 'if thou thinkest "I shall not die'"; nd 
ta nacanti nd dabhati tdskarah (RV.), 'they do not become lost ; 
no thief can harm them'; kdsmai devhya havisa vidhema (RV.), 
'to what god shall we offer oblation?' agniria rayim acnavat . . . 
dive-dive (RV.), 'by Agni one may gain wealth every day'; utai 
'nam brahmdne dadyat tdtha syona civa syat (AV.), 'one should 
give her, however, to a Brahman ; in that case she will be 
propitious and favorable'; dhar-ahar dadyat (QB.), 'one should 
give every day'. 

577. The uses of the optative in the later language are of 
the utmost variety, covering the whole field occupied jointly by 
the two modes in earlier time. A few examples from a single 
text (MBh.) will be enough to illustrate them : uchistam nai 'va 
bhunjlyam na kuryam padadhavanam, 'I will not eat of the rem- 
nant of the sacrifice, I will not perform the foot-lavation' ; jnatln 
vrajet, 'let her go to her relations'; nai 'vam sa karhicit kuryat, 
'she should not act thus at any time'; katham vidyam nalam 
nrpam, 'how can I know king Nala?' utsarge samcayah syat tu 

581] USES OF THE MODES. 199 

vindeta 'pi sukharii kvacit, 'but in case of her abandonment there 
may be a chance ; she may also find happiness somewhere' ; 
hitham vaso wkarteyam na ca budhyeta me priya, 'how can I cut 
off the garment and my beloved not wake?' 

578. The later use of the first persons subjunctive as so-, 
called imperative involves no change of construction from former 
time, but only restriction to a single kind of use : thus, divyava, 
k let us two play'; kirn karavcini te, 'what shall I do for thee?' 

579. The subjunctive with mi is in the oldest language 
almost the sole form of prohibitive expression, and is very 
common. The kind of subjunctive employed is that which cor- 
responds to the augmentless forms of a past tense ; and in the 
great majority of cases (five sixths inRV., nine tenths in AV.) 
it is the augmentless aorist that is chosen. Thus : prd pata me 
'hd raiisthah (AV.), 'fly away; do not stay here'; dvisdnc ca md- 
hyam rddhyatu ma ca 'Mm dvisate radham (AV.), 'both let my 
foe be subject to me, and let me not be subject to my foe'; 
urv dcyam dbhayam jyotir indra ma no dirgha abhi nacan tamisrah 
(RV.), 'I would win broad fearless light, O Indra; let not the 
long darknesses come upon us'; ma na ayuh prd mosih (RV.), 
'do not steal away our life'; ma libher (impf.) na marisyasi (RV.), 
'do not fear; thou wilt not die'; ma smai 'tint sdkhm kuruthah 
(AV.), 'do not make friends of them'. 

Only one optative (bhujema) is used prohibitively with md in RV., and 
only once (in a probably corrupted passage) an imperative; neither con- 
struction is found in AV.; and the cases in the later language are rare. 

580. This very definite and peculiar construction, of an 
augmentless past tense with ma, has preserved itself in use, and 
is occasionally met with in the later language: thus, samaqvasihi 
ma cucah, 'be comforted, do not grieve'; ma bhaih, 'do not fear' 
(both MBh.). 

But the use of the optative with nd, 'not', in a prohibitive 
sense appears even (very rarely) in the Veda, and becomes later 
the prevalent construction ; thus, nd risyema kada cand (RV.), 
'may we suffer no harm at any time'; nd ca 'tisrjen nd juhuyat 
(AV.), 'and if he do not grant permission, let him not sacrifice'; 
tad u tdtha nd kuryat (QB.), 'but he must not do that so'; na 
diva cayita (QGS.), ; let him not sleep by day'; na tvam vidyur 
jandh (MBh.), 'let not people know thee'. This in the later 
language is the correlative of the prescriptive optative, and both 
are extremely common ; so that in a text of prescriptive char- 
acter the optative forms may come to outnumber the indicative 
and imperative together (as is the case, for example, in Manu). 

581. In all dependent constructions, it is still harder even 


in the oldest language to establish a distinction between sub- 
junctive and optative : a method of use of either is scarcely 
to be found to which the other does not furnish a practical 
equivalent -- and then, in the later language, such uses are 
represented by the optative alone. A few examples will be 
sufficient to illustrate this : 

a. After relative pronouns and conjunctions in general : 
ya vyuchur yac ca nundm vyuchhn (RV.), ' which have shone forth 
[hitherto], and which shall hereafter shine forth'; yo l to jayata 
asmakam sd eko 'sat (TS.), 'whoever shall be born of her, let 
him be one of us'; yo vai thn vidyat pratydksam sd brahma vedita 
syat (AV.), 'whoever shall know them face to face, he may pass 
for a knowing priest'; putranam. . . jatUnam jandyac ca yhn (AV.), 
'of sons born and whom thou mayest bear'; ydsya . . . dtithir 
grhhn agdchet (AV.), 'to whosesoever house he may come as 
guest'; yatamdtha kamdyeta tdtha kuryat (QB.), 'in whatever way 
he may choose, so may he do it'; ydrhi hota ydjamanasya nhma 
grhmyht tdrhi bruyat (TS.j, 'when the sacrificing priest shall name 
the name of the offerer, then he may speak'; svarupam yada 
drastum ichethah (MBh.), 'when thou shalt desire to see thine 
own form'. 

b. In more distinctly conditional constructions : ydjama 
dev&n yddi cakndvama (RV.), 'we will offer to the gods if we 
shall be able'; ydd agne syhm ahdm tvdm tvdm va gha sya ahdm 
syus te satya iha "cisah (RV.), 'if I were thou, Agni, or if thou 
wert I, thy wishes should be realized on the spot'; yo dyam 
atisdrpat pardsfan nd sd mucyatai vdriinasya rajnah (AV.), 'though 
one steal far away beyond the sky, he shall not escape king 
Varuna'; ydd dnacvan upavdset ksodhukah syad ydd acniyad rudrb 
'sya pacttn abhi many eta (TS.), 'if he should continue without 
eating, he would starve : if he should eat, Rudra would attack 
his cattle'; prarthayed yadi mam kaccid dandyah sa me puman 
bhavet (MBh.), 'if any man soever should desire me, he should 
suffer punishment'. These and the like constructions, with the 
optative, are very common in the Brahmanas and later. 

c. In final clauses : ydtha 'ham catruho 'sani (AV.), 'that 
I may be a slayer of my enemies'; grnanti ydtha pibatho dndhah 
(RV.), 'that being praised with song ye may drink the draught'; 
urau ydthu tdva cdrman mddema (RV.), 'in order that we rejoice 
in thy wide protection'; iipa jamta ydthe 'yam pimar agdcJiet (QB.j, 
'contrive that she come back again'; krpam bury ad yatha mayi 
(MBh.), 'so that he may take pity on me'. This is in the Veda 
one of the most frequent uses of the subjunctive ; and in its 
correlative negative form, with ned, 'in order that not' or 'lest 1 , 
it continues not rare in the Brahmanas. 




The indicative is also used in final clauses after yatha: thus, yatha 
'yarn nafyati tathd vidheyam (H.), 'it must be so managed that he perish' 
(and thus usually in H.). 

With the conditional use of subjunctive and optative is further to be 
compared that of the so-called "conditional" tense : see below, chap. XII. 

582. No distinction of meaning has been established be- 
tween the modes of the present-system and those (in the older 
language) of the perfect and aorist-sy stems. 


583. Participles, active and middle, are made from all 
the tense-stems except the periphrastic future, and, in 
the later language, the aorist (and aorist participles are rare 
from the beginning). 

584. The general participial endings are ST^T ant (weak 
form 5fcT at; fern. Trft antl or STrft atl: see above, 449) for 
the active, and 5TFT ana (fern. t\\*\\ ana) for the middle. But: 

a. After a tense-stem ending in a, the active participial 
suffix is virtually nt, one of the two a's being lost in the com- 
bination of stem-final and suffix. 

b. After a tense-stem ending in a, the middle participial 
suffix is mana instead of ana. 

c. The perfect has in the active the peculiar suffix vans 
(weakest form us, middle form vat; fern, usi: see, for the in- 
flection of this participle, above, 458 ff.). 

For details, as to form of stem etc., and for special ex- 
ceptions, see the following chapters. 


585. The augment is a short a, prefixed to a tense-stem 
- and, if the latter begin with a vowel, combining with 

that vowel irregularly into the heavier or vrddhi diphthong 
(136 a). It is always (without any exception) the accented 
element in the verbal form of which it makes a part. 

In the Veda, the augment is in a few forms long a : thus, emap, auar, 
avrni, avrnak, avidhyat, ayunak, ayukta, arinak, araik. 

586. The augment is a sign of past time. And an aug- 
ment-preterit is made from each of the tense-stems from which 


the system of conjugation is derived: namely, the imperfect, 
from the present-stem ; the pluperfect (in the Veda only), from 
the perfect-stem ; the conditional, from the future-stem ; while 
in the aorist such a preterit stands without any corresponding 
present indicative. 

587. In the older language (mainly in the Veda; the usage 
is a rare one in the Brahmana) the augment is often lost, and 
the augmentless forms have the same value as if they were 
complete ; or, rather more often, they are used as subjunctives 
(above, 563). 

The accentuation of the augmentless forms is throughout accordant with 
that of the corresponding unaugmented tense that is to say, where such 
a tense exists (which is not the case with the varieties of sibilant aorist). 


588. The derivation of conjugational and declensional 
stems from roots by reduplication, either alone or along 
with other formative elements, has been already spoken of 
(259), and the formations in which reduplication appears 
have been specified: they are, in primary verb-inflection, 
the present (of a certain class of verbs), the perfect (of 
nearly all), and the aorist (of a large number); and the in- 
tensive and desiderative secondary conjugations contain in 
their stems the same element. 

589. The general principle of reduplication is the pre- 
fixion to a root of a part of itself repeated if it begin 
with consonants, the initial consonant and the vowel ; if it 
begin with a vowel, that vowel, either alone or with a follow- 
ing consonant. The varieties of detail, however, are very 
considerable. Thus, especially as regards the vowel, which 
in present and perfect and desiderative is regularly shorter 
and lighter in the reduplication than in the root-syllable, 
in aorist is longer, and in intensive is strengthened. The 
differences as regards an initial consonant are less, and 
chiefly confined to the intensive; for the others, certain 
general rules may be here stated, all further details being 




left to be given in connection with the account of the sep- 
arate formations. 

590. The consonant of the reduplicating syllable is in 
general the first consonant of the root : thus, W^" paprach 
from ysC prach; wS\W\J*$li8 fr m VfiPT^w; ^^bubudh 
from ysiTJ. But: 

' o ~x 

a. A non-aspirate is substituted in reduplication for an 
aspirate: thus, ^TJT dadha from |/ETT; f^R foMr from y>T 6/*r. 

b. A palatal is substituted for a guttural or for ^ h: 
thus, rJSfi ca^r from y^\ kr ; nrfl^ cikhid from 

SQW jagrc&h from ]/ER grabh ; ^"^ jahr from 

The occasional reversion, on the other hand, of a palatal in the radical 
syllable to guttural form has been noticed above (216.9). 

c. Of two initial consonants, the second, if it be a 
non-nasal mute preceded by a sibilant, is repeated instead 
of the first: thus, cTFSCT tastha from y F2JT stha; rj^r^ caskand 
form i/FfrJ skand ; 3Fp pasprdh from yFT^-T sprdh: -- but 
TFF sasmr from y F7. 

c. c 

Accent of the Verb. 

591. The statements which have been made above, and 
those which will be made below, as to the accent of verbal 
forms, apply to those cases in which the verb is actually accented. 

But, according to the grammarians, and according to the 
invariable practice in accentuated texts, the verb is in the great 
majority of its occurrences unaccented or toneless. 

That is to say, of course, the verb in its proper forms, its personal or 
so-called finite forms. The verbal nouns and adjectives, or the infinitives 
and participles, are subject to precisely the same laws of accent as other nouns 
and adjectives. 

592. The general rule, covering most of the cases, is this : 
The verb in an independent clause is unaccented, unless it stand 
at the beginning of the clause or also, in metrical text, at 
the beginning of a pada. 

For the accent of the verb, as well as for that of the vocative case 
(above, 314), the beginning of a pada counts as that of a sentence, what- 
ever be the logical connection of the pada with what precedes it. 

Examples of the unaccented verb are : agnim ide purohitam, 'I praise 


Agni, the house-priest'; sd id devesu gachati, 'that, truly, goes to the god?'; 
ague supdyano bhava, '0 Agni, be easy of access'; iddm indra prnuhi somapa, 
'hear this, Indra, sorna-drinker'; ndmas te rudra krrtmas, 'homage to thee, 
Rudra, we offer'; ydjamdnasya pa?un pahi, 'protect the cattle of the sacrificer'. 
Hence, there are two principal situations in which the verb 
retains its accent : 

593. First, the verb is accented when it stands at the 
beginning of a clause or, in verse, of a pada. 

Examples of the verb accented at the head of the sentence are, in prose, 
cundhadhvam ddfvydya kdrmane, 'be pure for the divine ceremony'; apnotl 
J mdm lokdm, 'he wins this world'; in verse, where the head of the sentence 
is also that of the pada, syame 'd indrasya cdrmani, 'may we be in Indra's 
protection'; darcdya ma ydtudhandn, 'show me the sorcerers'; gdmad vdjebhir 
a sd nahj 'may he come with good things to us'; in verse, where the head 
of the clause is within the pada, tesam pahi crudhi hdvam, 'drink of them, 
hear our call'; sdstu mdtd sdstu pitd sdstu $vd sdstu vifpdtih, 'let the mother 
sleep, let the father sleep, let the dog sleep, let the master sleep'; vfyvakar- 
man ndmas te pdhy asmdn, 'Vicvakarman, homage to thee; protect us!' 
yuvam. .. rdjna uce duhitd prche vdm nara, 'the king's daughter said to you 
"I pray you, ye men'"; vaydm te vdya indra viddhi su nah prd bharamahe, 
'we offer thee, Indra, strengthening; take note of us 1 . 

Examples of the verb accented at the head of the pada when this is not 
the head of the sentence are: dthd te dntamdndm vidydma sumatindm, 'so 
may we enjoy thy most intimate favors'; dhdtd 'syd agrtivdi pdtim dddhdtu 
pratikamyam, 'Dhatar bestow upon this girl a husband according to her wish'; 
ydtudhdnasya somapa jaht prajdm, 'slay, Soma-drinker, the progeny of the 

594. Certain special cases under this head are as follows: 

a. As a vocative forms no syntactical part of the sentence to which it 
is attached, but is only an external appendage to it, a verb following an 
initial vocative, or more than one, is accented, as if it were itself initial in 
the clause or pada: thus, d?rutkarna fractal hdvam, *0 thou of listening ears, 
hear our call!' site vdndamahe tva, 1 Q Sita, we reverence thee'; vfyve deva 
vdsavo rdksate J mdm, 'all ye gods, ye Vasus, protect this man'; utd "gay 
cakrtisam deva deva jivdyathd punah, 'likewise him, gods, who has com- 
mitted crime, ye gods, ye make to live again'. 

b. If more than one verb follow a word or words syntactically connected 
with them all, only the first loses its accent, the others being treated as if 
they were initial verbs in separate clauses, with the same adjuncts under- 
stood: thus, tardnir fj jayati kseti ptisyati, 'successful he conquers, rules, 
thrives'; amitrdn . . . pdrdca indra prd mrnd jahi ca, 'our foes, Indra, drive 
far away and slay'; asmdbhyam jesi yotsi ca, 'for us conquer and fight'; 
dgnisomd havisah prdsthitasya vitlrh hdryatarh vrsana jusetham, '0 Agni and 
Soma, of the oblation set forth partake, enjoy, ye mighty ones, take plea- 


c. In like manner (but much less often^, an adjunct, as subject or 
object, standing between two verbs and logically belonging to both, is reckoned 
to the first alone, and the second has the initial accent: thus, jaht prajdrh 
ndyasva ca, 'slay the progeny, and bring [it] hither'; crnotu nah subhdgd 
bodhatu tmdnd, 'may the blessed one hear us, [and may she] kindly regard [u>] 1 . 

d. As to cases in which a single verb standing between two adjuncts 
has the initial accent perhaps as being in the division of the sentence reckoned 
to the second rather than the first, see below, 597. 

595. Second, the verb is accented, whatever its position, 
in a dependent clause. 

a. The dependency of a clause is in the very great majority of cases 
conditioned by the relative pronoun t/a, or one of its derivatives or compounds. 
Thus: yarn yajndm paribhur dsi, 'what offering thou protectest'; 6 te yanti 
ye pacyan, 'they are coming who shall behold her hereafter'; solid 
ydn me dsti Una, 'along with that which is mine'; ydtra nah purve pitdrah 
pareytih, 'whither our fathers of old departed'; adya muriya yddi ydtudhano 
dsmi, 'let me die on the spot, if I am a sorcerer'; ydthd J hdny anupurvdm 
bhdvanti, 'as days follow one another in order'; ydvad iddm bhtivanarh vfcvam 
dsti, 'how great this whole creation is'; ydtkdmds te juhumds tan no astu, 
'what desiring we sacrifice to thee, let that become ours'; yatamds tftrpsdt, 
'whichever one desires to enjoy'. 

The presence of a relative word in the sentence does not, of course, 
accent the verb, unless this is really the predicate of a dependent clause : thus, 
dpa tye tdydvo yathd yanti, 'they make off like thieves (as thieves do}'; ydt 
sthd jdgac ca rejate, 'whatever [is] movable and immovable trembles'; yathd- 
kdmarh ni padyate, ; he lies down at his pleasure'. 

b. The particle ca when it means 'if, and ced (ca + id), 'it', give an 
accent to the verb : thus, brahmd ced dhdstam dgrahlt, 'if a Brahman has 
grasped her hand'; tvdrh ca soma no vdco jwdturh nd mardmahe, 'if thou, 
Soma, wiliest us to live, we shall not die'; d ca gdchdn mitrdm end da- 
dhdma, 'if he will come here, we will make friends with him'. 

c. There are a very few passages in which the logical dependence of a 
clause containing no subordinating word appears to give the verb its 
accent : thus, sdm dcvaparndf cdranti no ndro 'smdkam indra rathino jayantu, 
'when our men, horse-winged, come into conflict, let the chariot-fighters of 
our side, Indra, win the victory'. Rarely, too, an imperative so following 
another imperative that its action may seem a consequence of the latter's is 
accented; thus, tuyam d galii kdnvesu su sdcd ptba, 'come hither quickly; 
drink along with the Kanvas' (i. e. in order to drink). 

d. A few other particles give the verb an accent, in virtue of a slight 
subordinating force belonging to them : thus, especially hi (with its negation 
nahf), which in its fullest value means 'for', but shades off from that into 
a mere asseverative sense ; the verb or verbs connected with it are always 
accented: thus, m te muncantdm vimtico hi sdnti, 'let them release him, for 
they are releasers'; ydc cid dhf. . . andcastd iva smdsi, 'if we, forsooth, are 


as it were unrenowned'; also ned (na-\-id), meaning 'lest, that not': 
thus, net tva tdpati sUro artfsa, 'that the sun may not burn thee with his 
beam'; virdjam ned vichindddni J ti, 'saying to himself, "lest I cut off the 
viraf- and the interrogative kuvfd, 'whether?' thus, ukthebhih kuvid 
agdmat, 'will he come hither for our praises?' 

596. But further, the verb of a prior clause is not infre- 
quently accented in antithetical construction. 

Sometimes, the relation of the two clauses is readily capable of being 
regarded as that of protasis and apodosis ; but often, also, such a relation is 
very indistinct; and the cases of antithesis shade off into those of ordinary 
coordination, the line between them appearing to be rather arbitrarily drawn. 

In the majority of cases, the antithesis is made distincter by the pres- 
ence in the two clauses of correlative words, especially anya am/a, eka 
efca, va va, ca ea: thus, prd-pra J nye ydnti pdry anyd asate, 'some go 
on and on, others sit about' (as if it were 'while some go' etc.); tid va 
sincadhvam upa va prnadhvam, 'either pour out, or fill up'; sdrh ce 'dhydsva 
'gne prd ca vardhaye 'mam, 'both do thou thyself become kindled, Agni, 
and do thou increase this person'. But it is also made without such help : 
thus, pro, 'jatah prajd jandyati part prdjata grhnati, 'the unborn progeny he 
generates, the born he embraces'; dpa yusmdd dkramm nd J smdn updvartate, 
'[though] she has gone away from you, she does not come to us'; nd 'ndhb 
'dhvarytir bhdvati nd yajndrh rdksansi ghnanti, 'the priest does not become 
blind, the demons do not destroy the sacrifice'. 

597. "Where the verb would be the same in the two antithetical clauses, 
it is not infrequently omitted in the second: thus, beside complete expres- 
sions like urvi ca 'si vdsvl ca 'si, 'both thou art broad and thou art good', 
occur, much oftener, incomplete ones like agnfr amusmih lokd dsid yamb 
'smfn, 'Agni was in yonder world, Yama [was] in this'; asthnd 'nydh prajdh 
pratitfsthanti mansena J nydh, 'by bone some creatures stand firm, by flesh 
others'; dvipac ca sdrvarh no rdksa cdtuspad ydc ca nah svdm, 'both protect 
everything of ours that is biped, and also whatever that is quadruped 
belongs to us'. 

Examples from the Brahmanas like the first of those here given (with 
the second verb expressed), and like the third (in composition with a pre- 
position), show that this explanation of the verbal accent is preferable to the 
one formerly given namely, that the verb is to be regarded as understood 
in the first clause and initial in the second. 

598. In a very small number of more or less doubtful 
cases, the verb appears to be accented for emphasis. 

Thus, before cand, 'in any wise'; in connection with the asseverative 
particles fd, dha, ktla, angd, evd, but sporadically; and so on. The detail 
and examination of the cases is not worth while here*. 

* The specialities and irregularities of the EV. as regards verbal accent are discussed 
by A. Mayr in Sitzungsb. d. Wiener Akad. for 1871; of the AV., by W. D. Whitney in 
J. A. 0. 8., vol. v. (and Kuhn's Beitrage, vol. i.) ; of the TS., by A. Weber in Ind. Stud., 
-vol. xiii. 

601] 207 



599. THE present-system, or system of forms coming 
from the present-stem, is composed (as was pointed out 
above) of a present indicative tense, along with a subjunc- 
tive (mostly lost in the classical language), an optative, an 
imperative, and a participle, and also a past tense, an 
augment-preterit, to which we give (by analogy with the 
Greek) the name of imperfect. 

These forms generally go in Sanskrit grammar by the name of "special 
tenses", while the other tense-systems are styled "general tenses" as if 
the former were made from a special tense-stem or modified root, while the 
latter came, all alike, from the root itself. There is no reason why such a 
distinction and nomenclature should be retained; since, on the one hand, 
the "special tenses" come in one set of verbs directly from the root, and, 
on the other hand, the other tense-systems are mostly made from stems 
and, in the case of the aorist, from stems having a variety of form comparable 
with that of present-stems. 

600. Practically, the present-system is the most pro- 
minent and important part of the whole conjugation, since, 
from the earliest period of the language, its forms are very 
much more frequent than those of all the other systems 

Thus, in the Veda, the occurrences of personal forms of this system are 
to those of all others about as three to one ; in the Aitareya Brahmana, as 
five to one ; in the Hitopadega, as six to one ; in the Qakuntala, as eight to 
one ; in Manu, as thirty to one. 

601. And, as there is also great variety in the manner 
in which different roots form their present-stem, this, as 
being their most conspicuous difference, is made the basis 
of their principal classification ; and a verb is said to be of 
this or of that conjugation, or class, according to the way 
in which its present-stem is made. 

208 IX - PRESENT-SYSTEM. [602 

602. In a small minority of verbs, the present-stem is 
identical with the root. Then there are besides (excluding 
the passive and causative) eight more or less different ways 
of forming a present-stem from the root, each way being 
followed by a larger or smaller number of verbs. These 
are the "classes" or "conjugation-classes", as laid down by 
the native Hindu grammarians. They are arranged by the 
latter in a certain wholly artificial and unsystematic order 
(the ground of which has never been pointed out) ; and they 
are wont to be designated in European works according to 
this order, or else, after Hindu example, by the root stand- 
ing at the head of each class in the Hindu lists. A differ- 
ent arrangement and nomenclature will be followed here, 
namely as below the classes being divided (as is usual 
in European grammars) into two more general classes or 
conjugations, distinguished from one another by wider differ- 
ences than those which separate the special classes. 

603. The classes of the FIRST CONJUGATION are as 
follows : 

I. The root-class (second class, or ad-class, of 
the Hindu grammarians); its present-stem is coincident 
with the root itself: thus, ^ ad, 'eat'; ^ , 'go'; EfT 
ya, 'go'; fef dvis, 'hate'; 3^ duh, 'milk'. 

II. The reduplicating class (third or hu- 
class); the root is reduplicated to form the present-stem : 
thus, sp^" juliu from y^ hu, 'sacrifice'; ^7 dada from 
I/^T, 'give'; \3p\ bibhr from j/H, 'bear'. 

III. The nasal class (seventh or rudk-class) ; a 
nasal, extended to the syllable ^ na in strong forms, is 
inserted before the final consonant of the root: thus, 
~^{jundh (or "^TTTCJ runadh) from i/"^U rudh; TFj(^yunj 
(or W^yunaj] from >/?JsT yuj. 

IV. a. The ^w-class (fifth or sw-class); the syl- 


lable *7 nu is added to the root: thus, R sunu from 
yH; 3T3 apnu from yJETR ap. 

b. A very small number 'only half-a-dozen) of 
roots ending already in ^ n, and also one very common 
and quite irregularly inflected root not so ending (off 
kr, 'make'), add 3 u alone to form the present-stem. This 
is the eighth or tow-class of the Hindu grammarians ; it 
may be best ranked by us as a sub-class, the u- class: 
thus, cH tanu from i/rR tan. 

O -V 

V. The na- class (ninth or ^n-class); the syllable 
JTT na (or, in weak forms, ^ m\ is added to the root : 
thus, stjlmi farina (or jtiluD krinl] from i/sfft" &n, 'buy'; 
FrPTT stabhna (or FrPft stabhnl) from i/FcR sta&A, 'estab- 

604. These classes have in common, as their most fund- 
amental characteristic, a shift of accent: the tone being 
now upon the ending, and now upon the root or the class- 
sign. Along with this goes a variation in the stem itself, 
which has a stronger or fuller form when the accent rests 
upon it, and a weaker or briefer form when the accent is 
on the ending: these forms are to be distinguished as the 
strong stem and the weak stem respectively (in part, both 
have been given above). The classes also form their opta- 
tive active, their 2d sing, imperative, and their 3d pi. 
middle, in a different manner from the others. 

605. In the classes of the SECOND CONJUGATION, the 
present-stem ends in , and the accent has a fixed place, 
remaining always upon the same syllable of the stem, and 
never shifted to the endings. Also, the optative, the 2d 
sing, impv., and the 3d pi. middle are (as just stated) un- 
like those of the other conjugation. 

V 606. The classes of this conjugation are as follows: 

VI. Thea-class, or unaccented a-class (first 

Whitney, Grammar. 14 


or J/m-class) ; the added class-sign is a simply ; and the 
root, which has the accent, is strengthened by guna 
throughout: thus. *R bhdva from y*\ bhu, 'be'; ^J nay a 
from y^\m, 'lead'; SJTEI bodha from y^3j)udh, 'wake'; 
^vdda from y^ vad^ 'speak'. 

VII. The a-class, or accented a-class (sixth or 
tud-cl&ss) ; the added class-sign is a, as in the preceding 
class; but it has the accent, and the unaccented root 
remains unstrengthened : thus, c^tuddfrom^ff^. 'thrust'; 
HsT srjd from /TO srj. 'let loose' ; R suvd from /H su, 

c t "v *" o c\ 

'give birth'. 

VIII. The /#-class (fourth or cfoiJ-class) ; ya is added 
to the root, which has the accent : thus, ^oT d^vya from 
Y^( div (more properly ^fa div: see 765); ^3f ndhya 
from y 7 ^ nah, 'bind'; ^F^TT krudhya from y&m krudh, 
'be angry'. 

IX. The passive conjugation is also properly a 
present-system only, having a class-sign which is not 
extended into the other systems ; though it differs mark- 
edly from the remaining classes in having a specific 
meaning, and in being formable in the middle voice 
(only) from all transitive verbs. Its inflection may 
therefore best be treated next to that of the ?/a-class, 
with which it is most nearly connected, differing from 
it as the a-class from the a-class. It forms its stem, 
namely, by adding an accented yd to the root : thus, 
^7J adyd from y*$Z[ ad; "^EET rudhyd from i/^T rudh ; 
budhyd from y'SFl budh; H" 2 ^ tudyd from y"^ tud. 

607. The Hindu grammarians reckon a tenth class or cur- 
class, having a class-sign aya added to a strengthened root (thus, 
cor&ya from ycur), and an inflection like that of the other a- 
stems. Since, however, this stem is not limited to the present- 
system, but extends also into the rest of the conjugation 
while it also has to a great extent a causative value, and may 


be formed in that value from a large number of roots it will 
be best treated along with the derivative conjugations (chap. XIV.). 

608. A small number of roots add in the present-system 
a ch, or substitute a ch for their final consonant, and form a 
stem ending in cha or chd, which is then inflected like an a- 
stem. This is historically, doubtless, a true class-sign, analogous 
with the rest ; but the verbs showing it are so few, and in 
formation partly so irregular, that they are not well to be put 
together into a class, but may best be treated as special cases 
falling under the other classes. 

Roots adding ch are r and yu, which make the stems rchd and yucha. 

Roots substituting ch for their final are is, us (or vas 'shine'), gam, 
yam, which make the stems ichd, uchd, gdcha, ydcha. 

Of so-called roots ending in ch, several are more or less clearly stems, 
whose use has been extended from the present to other systems of tenses. 

609. Roots are not wholly limited, even in the later language, to one 
mode of formation of their present-stem, but are sometimes reckoned as be- 
longing to two or more different conjugation-classes. And such variety of 
formation is especially frequent in the Veda, being exhibited by a consider- 
able proportion of the roots there occurring; already in the Brahmanas, how- 
ever, a condition is reached nearly agreeing in this respect with the classical 
language. The different present-formations sometimes have differences of 
meaning ; yet not more important ones than are often found belonging to 
the same formation, nor of a kind to show a difference of value as originally 
belonging to the separate classes of presents. If anything of this kind is to 
be established, it must be from the derivative conjugations, which are separ- 
ated by no fixed line from the present-systems. 

610. We take up now the different classes, in the order 
in which they have been arranged above, to describe more in 
detail, and with illustration, the formation of their present- 
systems, and to notice the irregularities belonging under each 

I. Root-class (second, ad-c\ass). 

\ 611. In this class there is no class-sign; the root itself 
is also present-stem, and to it are added directly the per- 
sonal endings - - but combined in subjunctive and optative 
with the respective mode-signs, and in the imperfect taking 
the augment prefixed to the root. 

The accented endings (552) regularly take the accent except 
in the imperfect, where it falls on the augment and before 



them the root remains unchanged ; before the unaccented endings, 
the root takes the guna- strengthening. 

It is only in the first three classes that the endings come immediately 
in contact with a final consonant of the root, and that the rules for consonant 
combination have to be noted and applied. 

1. Present Indicative. 

"612. The endings are the primary (with f?T ate in 3d 

pi. mid.), added to the hare root. The root takes the accent, 

and has guna, if capable of it, in the three persons sing. act. 

Examples of inflection : a. root ^ e, 'go': strong 

form of root-stem, ^ e; weak form, ^ i. 

active. f* middle.* P 

s. d. p. s. 

emi ivds imds iye ivdhe intake 

ithds ithd ise iyathe idhve 

eti Ltds ydnti ite iyhte iydte 

b. root fir^T dvis, 'hate'; strong stem-form, "^ dves ; 
weak, fer dvis. 

For rules of combination for the final s, see 226. 
i &RH I^^IH I^Q-HH T^" |&fe|t(. f^a-H 1 ^ 

dvesmi dvisvds dvismds dvise dvisvdhe dvismdhe 

dveksi dvisthds dvisthd dvikse dvisathe dviddhve 

l i dvistds dvisdnti dviste dvisate dvisdte 

c. root ~^^duh t 'milk': strong stem-form ^^doh; weak, 

For rules of combination for the final ft, and for the conversion of the 
initial to dft, see 222, 155, 160. 

dohmi duhvds duhmds duhe duhvdhe duhmdhe 

* Used in the middle with the preposition adhi, to signify 'go over for one's self, 
i. e. 'repeat, learn, read'. 

616] I. ROOT-CLASS (SECOND, #G?-CLASS). 213 

2 MUSI IT^H^ I^M" uif ^c^isj qrij 

dugdhds dugdhd dhukse duhhthe dhugdhve 

^JMH^ .S^lTi 3"^ ^c^(rt Js^H 
dogdhi dugdhds duhdnti dugdhe duhhte duhdte 
613. Examples of the 3d sing. mid. coincident in form with the 1st sing, 
are not rare in the older language (both V. and Br.): the most frequent 
examples are ice, duhe, vide, cdye: more sporadic are cite, bruve, huve. 
The irregular accent of the 3d pi. mid. is found in RV. in rihate, duhate, 
Examples of the same person in re and rate also occur: thus (besides those 
mentioned below, 629 30, 635), vidre, and, with auxiliary vowel, arhire 
(unless these are to be ranked, rather, as perfect forms without reduplica- 
tion: 790 b). 

2. Present Subjunctive. 

614. Subjunctive forms of this class are not uncommon in 
the older language, and nearly all those which the formation 
anywhere admits are quotable, from Veda or from Brahmana. 
A complete paradigm, accordingly, is given below, with the few 
forms not actually quotable for this class enclosed in brackets. 
We may take as models, for the active the root i, 'go', and for 
the middle the root as, 'sit', of both of which numerous forms 
are met with (although neither for these nor for any others can 
the whole series be found in actual use). 

The mode-stems are dya (e 

-f-fl) and asa respectively. 


























\ dsantai 

615. The RV. has no middle forms in di except those of the first per- 
son. The 1st sing. act. in d occurs only in RV., in aya, brava, stdvd. 
The 2d and 3d sing. act. with primary endings are very unusual in the 
Brahmanas. Forms irregularly made with long a, like those from present- 
stems in a, are not rare in AV. and Br. : thus, ayds, aydt, dydn; dsdt, 
brdvdt ,- asatha, bravatha, handtha; dddn, dohdn. Of middle forms with sec- 
ondary endings are found hdnanta, 3d pi., and Ifata, 3d sing, (after ma 
prohibitive), which is an isolated example. The only dual person in dite is 

3. Present Optative. 

616. The personal endings combined with the mode- 
signs of this mode (7JT ya in act., ^i in mid.) have been 


given in full above (566). The stem-form is the unaccented 
and unstrengthened root. The whole formation is so regular 
that a single example of inflection will be enough. 

active. middle. 

a. d. p. s. d. p. 


dvisyam dvisy&va dvisy&ma dvisiyd dvisivdhi dvisimdhi 


dvisy&s dvisy&tam dvisy&ta dvisiihks dvisiyatham dvisldhvdm 


dvisyat dvisyatam dvisyus dvisitd dvisiyatam dvisirdri 

So likewise, from j/i, iydm and iylyd; from yduh, duhydm and duhiyd; 
and so on. 

The RV. has once tana in 2d pi. act. (in syatana). 

4. Present Imperative. 

"^617. The imperative adds, in second and third persons, 
its own endings (with SfrTFT atum in 3d pi. mid.) directly 
to the root-stem. The stem is accented and strengthened 
in 3d sing, act.; elsewhere, the accent is on the ending 
and the root remains unchanged. The first persons, so called, 
of the later language are from the old subjunctive, and 
have its strengthened stem and accent; they are repeated 
here from where they were given above (614). In the 2d 
sing, act., the ending is regularly (as in the two following 
classes) ftj dhi if the root end with a consonant, and f% hi 
if it end with a vowel. As examples we take some of the 
roots already used for the purpose. r 

active. middle. 

s- d. p. g. d. p. 

dyani dyava dyama asai asavahai asamahai 

^p ^rFT ^rT MlfH MIHIMM t4i^H 

ihi itdm ltd assvd ashtham addhvdm 

3 yr\ 

etu itam ydntu astam asatam asatam 

620] I. ROOT-CLASS [SECOND, ad-CLASx. 215 

dohani doJiava dohama dohai dbhavahai dohdmahdi 

jfiri p~^ jrcr g^r ^IMIH^ gnj\ 

dugdlii dugdhdm dugdhd dhuksvd duhhtham dhugdhvdm 


dogdhu dugdhum duhdntu dugdlihm duhktam duhdtdm 

618. The 2d sing. act. ending tat is found in the older language in a 
few verbs of this class: namely, vittdt, vitat, brutat. In 3d sing, mid., two 
or three verbs have in the older language the ending am : thus, duhdm (only 
RV. case), uidara, fayam; and in 3d pi. mid. AV. has duhrdm and duhratdm. 
The use of tana for ta in 2d pi. act. is quite frequent in the Veda: thus, 
itana, yatdna, hantana, etc. And in stota, etana, sofana, we have examples 
in the same person of a strong (and accented) stem. 

5. Present Participle. 

619. The active participle has the ending *3fi[jint (weak 
stem-form 5IH at] added to the unstrengthened root. Me- 
chanically, it may be formed from the 3d pi. by dropping 
the final ^ i. Thus, for the verbs inflected above, the active 
participles are EPFT ydnt, p^H duhdnt, f^Nri dmsdnt. The 
feminine stem ends usually in 5^ ati : thus, ZTrft yati-, T^^ft 
duhati, f^Mrfl dvisati: but, from roots in , in 5Ttf?ft anti 
or 5flrft ati (449). 

The middle participle has the ending ETR and, added 
to the unstrengthened root: thus, ^ETR iyand, 3^11 duhand^ 

But a number of these participles in the older language 
have a double accent, either on the ending or on the radical 
syllable : thus, duhdnd and duhana (also dughana], vidand and 
vidana, stivand and siivana, stuvand and stdvdna the last having 
also a stronger form of the root when accented. The root as, 
'sit', forms the unique Usma (along with, in the Veda, dsand}. 

6. Imperfect. 

620. This tense adds the secondary endings to the root 
as increased by prefixion of the augment. The root has the 
^w^a-strengthening (if capable of it) in the three persons of 


the singular, although the accent is always upon the augment. 
Examples of inflection are: 

active. middle, 

s. d. p. s. d. p. 

at/am aiva aima Usi asvahi asmahi 

ais aitam aita asthas Usatham addhvam 

ait aitam ayan asta asatam mata 

and, from the root 

ddoham dduhva dduhma dduhi dduhvahi dduhmahi 

ddugdham ddugdha ddugdhas dduhatham ddhugdhvam 

ddhok ddugdham dduhan ddugdha dduhatam dduhata 

621. Roots ending in a may in the later language option- 
ally take us instead of an in 3d pi. act. (the a being lost before 
it) ; and in the older they always do so : thus, dyus from Vya, 
dpus from Ypa 'protect'. The same ending is also allowed and 
met with in the case of a few roots ending in consonants : 
namely vid, 'know', caks, dvis, duh, mrj. 

The ending tana, 2d pi. act., is found in the Veda in dyatana, dsastana, 

To save the characteristic endings in 2d and 3d sing, act., the root ad 
inserts a: thus, ddas, ddat; the root as inserts 1: thus, dsis, dsit (see 
below, 636). 

622. The use of the persons of this tense, without augment, in the 
older language, either in the same sense as with augment, or as subjunctives, 
has been noticed above (587). Augmentless imperfects of this class are 
rather uncommon in the Veda : thus, h&n, ves, 2d sing. ; ftan, vet, staut, 
ddn (?), 3d sing.; bruvan, duhus, eafesus, 3d pi.; vasta, suta, 3d sing. mid. 

623. The first or root-form of aorist is identical in its formation with 
this imperfect: see below, 829 ff. 

624. In the Veda (but almost limited to RV.) are found certain second 
persons singular, made by adding the ending si to the (accented and streng- 
thened) root, and having an imperative value. There is some difference of 
view as to their formal character; but the most acceptable opinion regards 
them as isolated indicative persons of this class, used imperatively. They 

629] I. ROOT-CLASS (SECOND, a</-CLASs). 217 

are : ksesi (1 fcai 'rule'), jesi, jdsi (for jossi, from yjus), ddrsi, dhaksi, naksi 
(2naf 'attain'), nesi, pdrsi (2pr 'set across'), prdsi, bhaksi, mdtsi, mdsi 
(2 ma 'measure'), ydksi, ydrhsi, yasi, y6tsi, rdtsi, rdsi, vdksi, vesi (ivi 'strive 
after'), frfoj, safest, sdtsi, hosi. 

Irregularities of the Root-class. 

625. It is impossible (at least at present) to determine with 
accuracy how many of the actually used roots of the language 
are inflected in the present-system according to this class, or 
according to any of the other classes, because the older language 
especially, and the later in less degree, has sporadic forms which 
are either of doubtful classification or too isolated to determine 
the character of the root to which they belong. The root-class 
may be said, however, to include from seventy to ninety roots. 
A considerable number of them present irregularities of inflection, 
a brief account of which (not claiming exhaustive completeness) 
is given in the following paragraphs. 

626. The roots of the class ending in u have in their 
strong forms the vrddhi instead of the ywwa-strengthening before 
an ending beginning with a consonant : thus, from \stu, staumi, 
dstaut, and the like; but dstavam, stdvani, etc. 

Roots found to exhibit this peculiarity in actual use are ksnu, yu, sfcit, 
stu, snu (these five in the earlier language), nw, ru, su 'impel', and hnu. 

627. The root mrj also has the vrddhi-vowel in its strong 
forms: thus, marjmi, dmarjam, dmart; and the same streng- 
thening is allowed in weak forms before endings beginning with 
a vowel : thus, marjantu, amarjan : but this is not found to 
occur in the older language. 

In the other tense-systems, also, and in derivation, mrj shows often 
the vrddhi instead of the puna-strengthening. 

628. A number of roots accent the radical syllable through- 
out, both in strong and in weak forms : thus, all those begin- 
ning with a long vowel, as, id, ir, iq ; and also caks, taks, tra, 
nihs, vas 'clothe', cinj, cl, and su 'generate'. All these, except 
taks and tra (and tra in the only Vedic forms), are ordinarily 
conjugated in middle voice only. Forms with the same irregular 
accent occur now and then in the Veda from other verbs : thus, 
mdtsva, fdhat. Middle participles so accented have been noticed 
above (619). 

629. Of the roots mentioned in the last paragraph, ci has 
the ^wwa-strengthening throughout: thus, cdye, cese, qayiya, 
cdyana, and so on. Other irregularities in its inflection (in 
part already noticed) are the 3d pi. persons cerate (AV. etc. 

218 IX. PRESENT- SYSTEM. [629 

have also cere], ceratdm, dcerata {RV. has also dceran], the 3d 
sing, pres/ cdye (R.) and impv. cdydm. The isolated active form 
dcayat is common in the older language. 

630. Of the same roots, Id and if insert a union-vowel i before endings 
beginning with s, sv, th : thus, ifise, ifidhve, idisva (these three being the 
only forms noted in the older language); but RV. has ikse beside If we. The 
3d pi. ifire (on account of its accent) is also apparently present rather than 

631. The roots rud (not in Veda), svap, an, and cvas insert 
a union-vowel i before all the endings beginning with a con- 
sonant, except the s and t of 2d and 3d sing, impf., where they 
insert instead either a or i: thus, svdpimi, cvdstsi, dniti, and 
mat or unit. And in the remaining forms, the last three are 
allowed to accent either root or ending : thus, svdpantu and 
cvdsantu (AV.), or svapdntu etc. 

In the older language, yvam makes the same insertions : thus, vamiti, 
avamit; and other cases occasionally occur: thus, janisva, vasisva (yvas 'aim'), 
fnathihi, stanihi (all RV.). On the other hand, an sometimes makes forms 
from an a-stem: thus, anati (AV.); pple dnant (QB.); opt. anet (AB.). 

632. The root bru (of very frequent use) takes the union- 
vowel i after the root when strengthened, before the initial vowel 
of an ending : thus, brdvimi, brdvisi, brdviti, dbravts, dbravit; but 
brumds, bruyam, dbravam, dbruvan, etc. Special occasional irreg- 
ularities are brumi, bravihi, abruvam, abruvan, bruydt, and sporadic 
forms from an a-stem. The subj. dual brdvaite has been noticed 
above (615). .< 

633. Some of the roots in u are allowed to be inflected like bru: 
namely, few, tu, ru, and stu; and an occasional instance is met with of a 
form so made (in V., only tavlti noted). 

634. The root am (hardly found in the later language) takes I as union- 
vowel : thus, arriisi (RV.), amlti and amit and amisva (TS). From |/pam 
occur camwvo (VS.: TS. f amisva) and famidhvam (TB. etc.). 

635. The irregularities of yduh in the older language have been already 
in part noted: the 3d pi. indie, mid. duhate, duhre, and duhrdte; 3d sing, 
impv. duhdm, pi. duhrSm and duhratam; impf. act. 3d sing, dduhat (which 
is found also in the later language), 3d pi. aduhran (beside dduhan and 
duhtis); the mid. pple dtighana; and (quite unexampled elsewhere) the opt. 
forms duhiydt and duhiydn (RV. only). 

Some of the roots of this class are abbreviated or otherwise 
weakened in their weak forms : thus 

( 636. The root 5TFT as, 'be', loses its vowel in weak forms 
(except where protected by combination with the augment). 




Its 2d sing, indie, is gr dsi (instead of assi); its 2d sing, 
impv. is ^ftf edhi (irregularly from asdhi). The insertion of 
| i in 2d and 3d sing. impf. has been noticed already 

The forms of this extremely common verb are, then, as 

follows : 



sv as 
























dsani dsava dsama 




edhi stdm 





stam sdntu 
Participle H?T 

asit astam 

(fern. Hrft sail 


The Vedic subjunctive forms are the usual ones, made upon the stem 
daa. They are in frequent use, and appear (asat especially) even in late 
Brahmanas where the subjunctive is almost lost. The resolution sidm etc. 
(opt.) is common in Vedic verse. In 2d and 3d sing. impf. is a few times 
met with the more normal as (for as-s, ds-t}. 

Middle forms from yas are also given by the grammarians as allowed 
with certain prepositions (vi + ati), but they do not appear to have been met 
with in use. A middle present indicative is compounded (in 1st and 2d per- 
sons) with the nomen agentis in tr (tar] to form the periphrastic future in 
the middle voice (see below, 942 ff.). The 1st sing, indie, is he; the rest 
is in the usual relation of middle to active forms (in 2d pers., e, dhve, 
sva, dhvam, with total loss of the root itself). 

The only other tense of this verb in use is the perfect, which is entirely 
regular in its inflection. 

637. The root han, 'strike, slay', is treated somewhat after 
the manner of noun-stems in an in declension (421): in weak 


forms, it loses its n before an initial consonant (except m and v) 
of an ending, and its a before an initial vowel and in the 
latter case its A, in contact with the n, is changed to gh (com- 
pare 402). Thus, for example : 

Present Indicative. Imperfect. 

s. d. p. s. d. p. 

1 hdnmi hanvds hanmds dhanam dhanva dhanma 

2 hdnsi hathds hathd dhan dhatam dhata 

3 hdnti hatds ghndnti dhan dhatam dghnan 
Its participle is ghndnt (fern, g/matij. Its 2d sing. impv. 

is jaM (by anomalous dissimilation, on the model of redupli- 
cating forms). 

Middle forms from this root are frequent in the Brahmanas, and those 
that occur are formed according to the same rules : thus, hate, hanmahe, 
ghnate; ahata, aghnalam. aghnata (in AB., also ahata)] ghnlta (but also hanita). 

638. The root vac, 'wish', is in the weak forms regularly 
and usually contracted to uc (as in the perfect: see chap. X.): 
thus, ucmdsi (V.: once apparently abbreviated in RV. to cmasi), 
ucdnti; pple ucdnt, ucand. Middle forms (except the pple) do not 
occur ; nor do the weak forms of the imperfect, which are given 
as aufva, austam, etc. 

RV. has in like manner the participle usand from the root vas 'clothe'. 

639. The root cas, 'command', shows some of the peculi- 
arities of a reduplicated verb, lacking (646) the n before t in 
all 3d perss. pi. and in the active participle. A part of its 
active forms namely, the weak forms having endings begin- 
ning with consonants (including the optative) come from a 
stem with weakened vowel, fis (as do the aorist, 854, and some 
of the derivatives). Thus, for example : 

Present Indicative. Imperfect. 

s. d. p. s. d. p. 

1 fasmi tfsvds fismds dfasam dfisva dfisma 

2 fassi cisthds fisthd d?as dfistam dfista 

3 fasti fistds fasati dfat dfistam dfdsus 
In 2d sing, impf., afdt is said to be also allowed. If it is actually so 

used, the t must be the sporadic conversion of final radical s to t (167); 
and then it would be open to question whether the t of 3d sing, is radical 
or of the ending (according to 555). The optative is fisydm etc. The 2d 
sing. impv. is $adhf (with total loss of the s); and RV. has the strong 
2d pi. fastdna (with anomalous accent). But no pis-forms of the present- 
system occur in the Veda. 

The middle inflection is regular, and the accent (apparently) 
always upon the radical syllable (pple casana, RV. etc.). 

The root dap, 'worship', has in like manner (RV.) the pple ddfat (not 


640. The double so-called root jaks, 'eat, laugh', is an evident re- 
duplication of ghas and has. It has the absence of n in act. 3d perss. pi. 
and pple, and the accent on the root before vowel-endings, which belong 
to reduplicated verbs; and it also takes the union-vowel i in the manner of 
rud etc. (above, 631). A 2d pers. impv. jagdhi from it occurs. 

641. Other obviously reduplicated verbs are treated by the 
native grammarians as if simple, and referred to this conjugation : 
such are the intensively reduplicated jagr, daridra, and vevl 

(chap. XIV.), didhi etc. (676), and cakas (677). 

II. Reduplicating Class (third, /m-class). 

642. This class forms its present-stem by prefixing a 
reduplication to the root. 

\ 643. a. As regards the consonant of. the reduplication, 
the general rules which have already been given above (590) 
are followed. 

b. A long vowel is shortened in the reduplicating syl- 
lable: thus, ^J dada from i/^T da; fspft bibhl from y$ bhl; 
st^juhu from yj^ hu. The vowel fj r never appears in the 
reduplication, but is replaced by ^ i: thus. f^T bibhr from 
yy{ bhr ; fqtfrf piprc from yvft pro. 

For verbs in which a and a also are irregularly represented in the re- 
duplication by i, see below, 660. 

c. The only root of this class with initial vowel is r (or 
ar]\ it takes as reduplication t, which is held apart from the 
root by an interposed y: thus, iyar and iyr (the latter has not 
been found in actual use). 

\ 644. The present-stem of this class (as of the other 
classes belonging to the first conjugation) has a double 
form : a stronger form, with gunated root-vowel ; and a 
weaker form, without guna: thus, from j/^" hu, the two forms 
are ^^Ij'uho and 3^ juhu; from j/~>ft bhi, they are i^H bibhe 
and fipft bibhl. And the rule for their use is the same as 
in the other classes of this conjugation: the strong stem is 
found before the unaccented endings (552), and the weak 
stem before the accented. 


645. According to all the analogies of the first general 
conjugation, we should expect to find the accent upon the root- 
syllable when this is strengthened. That is actually the case, 
however, only in a small minority of the roots composing the 
class : namely, in hu, bhi (no test-forms in the older language), 
hri (not found in the older language), mad and dhan (both very 
rare), jan (no forms of this class found to occur), ci 'notice' 
(in V.), yu 'separate' (in older language only), and in bhr in 
the later language (in V. it goes with the majority : but RV. 
has bibharti once ; and this, the later accentuation, is found also 
in the Brahmanas). In all the rest apparently, by a recent 
transfer it rests upon the reduplicating instead of upon the 
radical syllable. And in both classes alike, the accent is anom- 
alously thrown back upon the reduplication in those weak 
forms of which the ending begins with a vowel ; while in the 
other weak forms it is upon the ending. 

Apparently (the cases with written accent are too few to determine the 
point satisfactorily) the middle optative endings, lya etc. (566), are reckoned 
throughout as endings with initial vowel, and throw back the accent upon 
the reduplication. 

\ 646. The verbs of this class lose the ? n in the 3d 


pl. endings in active as well as middle, and in the imper- 
fect have 3^ws instead of 3R an and before this a final 
radical vowel has guna, 

1. Present Indicative. 

" 647. The combination of stem and endings is as in 
the preceding class. 

Examples of inflection: a. y"^ hu, 'sacrifice': 
strong stem-form, sl^T/ttAo/ weak form, sC^ juhu (or juhu). 

active. middle. 

s w _ d. p. s. d. p. 


juhoti juhutas juhvati juhute juhvate juhvate 
b. Root *j[ bhr, 'bear' (given with Vedic accentuation): 
strong stem-form, fsPTf bibhar ; weak, ^R bibhr (or bibhr}. 

juhomi juhuvds juhumds juhve juhuvdhe juhumdhe 

juhosi juhuthds juhutha juhuse juhvathe juhudhve 



bibharmi bibhrvds bibhrmds bibhre bibhrvdhe bibhrmdhc 

C. *\ C. c. C. 

bibharsi bibhrthds bibhrthd bibhrse bibhrathe bibhrdhve 


bibharti bibhrtds bibhrati bibhrte bibhrate bibhrate 

The u of hu (like that of the class-signs nu and u: see below, 697) 
is said to be omissible before v and m of the endings of 1st du. and pi.: 
thus, juhvds, juhvdhe. etc. 

2. Present Subjunctive. 

648. It is not possible (at least, at present) to draw a distinct line 
between those subjunctive forms of the older language which should be 
reckoned as belonging to the present-system and those which should be as- 
signed to the perfect even, in some cases, to the reduplicated aorist and 
intensive. Here will be noticed only those which most clearly belong to this 
class; the more doubtful cases will be treated under the perfect-system. 
Except in first persons (which continue in use as "imperatives" down to the 
later language), subjunctives from roots having unmistakably a reduplicated 
present-system are of far from frequent occurrence. 

649. The subjunctive mode-stem is formed in the usual 
manner, with the mode-sign a and guna of the root-vowel, if 
this is capable of such strengthening. The evidence of the few 
accented forms met with indicates that the accent is laid in ac- 
cordance with that of the strong indicative forms : thus, from 
yhu, the stem would be juhdva : from ybhr, it would be bibhara 
(but bibhdra later). Before the mode-sign, final radical a would 
be, in accordance with analogies elsewhere, dropped : thus ddda 
from yda, dddha from ydha (all the forms actually occurring 
would be derivable from secondary roots, as dad and dadh}. 

650. Instead of giving a theoretically complete scheme of 
inflection, it will be better to note all the examples quotable 
from the older language (accented when found so occurring). 

Thus, of 1st persons, we have in the active juhdvani, bibhardni, dadani, 
dadhdni, jahani; juhavdma, dddhama, jdhama; in the middle, dadhai, 
mimdij dadhdvahdi; juhavdmahdi, daddmahe, daddmahai, dadhdmahdi. 

Of other persons, we have with primary endings in the active bibhardsi 
(with double mode-sign: 560, end), dddhathas, juhavatha (do.) and juhavatha ; 
in the middle, dddhase; dddhate, rdrate, dddhdtdi, daddtdi: with second- 
ary endings, dddhas, vfvesas, juhavat, bibharat, yuydvat, dddhat, dadhdnat, 
babhasat ; dadhan^ yuyavan, juhavan. 


3. Present Optative. 

\ 651. To form this mode, the optative endings given 
above (566), as made up of mode-sign and personal endings, 
are added to the unstrengthened stem. The accent is as 
already stated (645). The inflection is so regular that it is 
unnecessary to give here more than the first persons of a 
single verb : thus, 

active. middle. 

s. d. p. s. d. 

juhuyhm juhuy&va juhuy&ma juhvlya juhvwahi juhvimahi 
etc. etc. etc. etc. etc. etc. 

4. Present Imperative. 

\ 652. The endings, and the mode of their combination 
with the root, have been already given. In 2d sing, act., 
the ending is f% hi after a vowel, but f?I dhi after a con- 
sonant : ^ huj however, forms sT^Rl juhudhi (apparently. 
in order to avoid the recurrence of ^ h in two successive 
syllables): and other examples of fa dhi after a vowel are 
found in the Veda. 
{ 653. Example of inflection: 

active. middle. 

s. d. . s. d. p. 

juJidvani juhavava juhdvama juhdvai juhdvavahai juhdvamahai 

2 g^iu g^riH^ g^H spra ^lyiH^ 5^^ 

juhudhi juhut&m juhuta juhusvd juhvatham juhudhvdm 

3 3^ '5 ^niH^ g^g ^diH^ g^iniH^ sj^riiH^ 

juhdtu juhutnm jtihvatu juhuthm juhvatam juhvatam 

The other division of this class differ here, as in the in- 
dicative, in the accentuation of their strong forms only : namely, 
in all the first persons (borrowed subjunctives), and in the 3d 
sing, act.: thus (in the older language) Ubharani etc., Ubhartu, 
libharai etc. 

654. Vedic irregularities of inflection are: a. the occasional use of 
strong forms in 2d persons: thus, yuyodhi, fifadhi (beside tifihf); yuyotam 
(beside yuyutdm}-, {yarta, dddata, dddhata and dddhatana (see below, 673), 




pipartana, juh6ta and juhdtana, yuyotana; b. the use of dhi instead of hi 
after a vowel (only in the two instances just quoted) ; c. the ending tana 
in 2d pi. act. (in the instances just quoted and in others, as mamaltana, 
jigatana, dhattana, etc.); d. the ending tat in 2d sing, act., in dattdt, 
dhattat, piprtat. 

5. Present Participle. 

655. As elsewhere, the active participle-stem may be 
made mechanically from the 3d pi. indie, by dropping ^ i: 
thus, ^^juAvat, fsffi^bibhrat. In inflection, it has no dis- 
tinction of strong and weak forms (444). The feminine 
stem ends in Efi-fj 1 ati. The middle participles are regularly 
made: thus, sl^M Jukvana, I^WU blbhrana. 

6. Imperfect. 

656. As already pointed out, the 3d pi. act. of this 
class takes the ending 3?T us, and a final radical vowel has 
guna before it. The strong forms are, as in pres. indie., 
the three sing. act. persons. 
f* 657. Example of inflection: 



djuhavam djuhuva ajuhuma djuhvi ajuhuvahi ajuhumahi 

djuhos djuhutam djuhuta djuhuthas djuhvatham djiihudhvam 

djuhot djuhutam djuhavus djufiuta djuhvatam djuhvata 

From }/H bhr, the 2d and 3d sing. act. are 
dbibhar (for abibhar-s and abibhar-t] and so in all other 
cases where the strong stem ends in a consonant. The 3d 
pi. act. is ^^^dbibharus; and from y^\ bhi, it is tjf^tHH^ 


658. The usual Vedic irregularities in 2d pi. act. strong forms, and 
the ending tana occur in this tense also : thus, adadata, ddadhata ; ddat- 
tana, djagantana, djahatana. The RV. has also once apiprata for apiprta 
in 3d sing, mid., and abibhran for abibharus in 3d pi. act. Examples of 
augmentless forms are pi'?as, vives, jtgat; jthlta, ffylta, jihata. 
Whitney , Grammar. 15 

226 IX - PRESENT-SYSTEM. [659 

Irregularities of the Reduplicating Class. 

659. It is still more difficult to determine the precise limits 
of this class than of the root-class, because of the impossibility 
(referred to above, under subjunctive : 648) of always separating 
its forms from those of other reduplicating conjugations and 
parts of conjugations. In the RV., about forty roots may be 
confidently assigned to it; in the AV., less than thirty; many 
of them have irregularities (besides those in tense-inflection 
already pointed out). 

660. Besides the roots in r or ar namely, r, ghr (usually 
written ghar\ tr, pr, Ihr, sr, pro the following roots having 
a or a as radical vowel take t instead of a in the reduplicating 
syllable: ga 'go', Spa 'rise', ma 'measure', ma 'bellow', ca, ha 
'remove' (mid.), vac, sac; vac has both i and a; ra has * once 
in RV.: for stha, pa 'drink', gkra, Jian, see below (670 4). 

661. Several roots of this class in final a change the a in 
weak forms to * (occasionally even to i), and then drop it alto- 
gether before endings beginning with a vowel. 

This is in close analogy with the treatment of the vowel of the class - 
sign of the no-class: below, 717. 
These roots are : 

662. fa, act. and mid.: thus, fifati, fifimasi, fiflhi (also fifadhi : above, 
654), fifatu, afifat, tf$lte. 

663. ma 'bellow', act., and ma 'measure', mid. (rarely also act.): thus, 
mimati, mimanti, mimiyat ; mmuie, mimafc, amimlta; mimihi, mtmatu. 

664. ha 'remove', mid.: thus, jtfiite, jihidhve, jtfiate; jihisva, jihatam; 
djihlta, ajihata. 

665. ha 'quit', act. (originally identical with the former), may further 
shorten the I to i: thus, jahati, jahita, jahltat (AV.); jahimas (AV.), jahitas 
(TB.), jahitam (TA.), ajahitam (TS.). In the optative, the radical vowel is 
lost altogether; thus, jahyam, jahyus (AV.). The 2d sing. impv. is jahlhi 
or jahihi. 

Compare with this the forms in dhi from ydha (below, 669). 

666. ra 'give', mid.: thus, ranaTivam, ranthas (impf. without augment): 
and, with i in reduplication, ririhi. 

In all these verbs, the accent is constant on the reduplicating syllable. 

667. The two roots da and dha (the commonest of the 
class) lose their radical vowel altogether in the weak forms, 
being shortened to dad and dadh. In 2d sing. impv. act., they 
form respectively dehi and dhehi. In combination with a fol- 
lowing t or th, the final dh of dadh does not follow the special 
rule of combination of a final sonant aspirate (becoming ddh 
with the t or th : 160), but as also before s and dhv the 




more general rules of aspirate and of surd and sonant com- 
bination ; and its lost aspiration is thrown back upon the initial 
of the root. 

668. The inflection of ydha is. then, as follows: 

Present Indicative, 

s. d. p. 

1 dddhdmi dadhvds dadhmds 

2 dddhdsi dhatthds dhatthd 

3 dddhati dhattds dddhati 







i dadhyam 

1 dddhani 
2 dhehf 
3 dddhatu 


Present Optative. 
dadhydma dddhlya 





Present Imperative. 

dddhdva dddhdma dddhdi dddhdvahai dddhdmahdi 

dhattdm dhattd dhatsva dadhdthdm dhaddhvam 

dhattdm dddhatu dhattdm dadhdtdm dadhatdm 

ddadhvahi ddadhmahi 
ddadhdthdm ddhaddhvam 
ddadhdtdm ddadhata 


i ddadhdm ddadhva ddadhma ddadhi 
2 ddadhds ddhattam ddhatta ddhatthds 
3 ddadhdt ddhattam ddadhus ddhatta 
Participles: act. dddhat; mid. dddhdna. 

In the middle (except impf.), only those forms are here accented for 
which there is authority in the accentuated texts, as there is discordance 
between the actual accent and that which the analogies of the class would 
lead us to expect. RV. has once dhdtse: dadhe and dadhdte might be per- 
fect, so far as the form is concerned. RV. accents dadhitd once (dddhlta 
thrice); TS. and TB. have dddhiran. 

The root da is inflected in precisely the same way, with 
change everywhere of dh to d. 

669. The older language has irregularities as follows: a. the usual 
strong forms in 2d pi., dddhdta and ddadhata, ddddta and ddaddta; b. the 
usual tana endings in the same person, dhattana, ddddtana, etc.; c. the 
3d sing, indie, act. dadhe (like 1st sing.); d. the 2d sing. impv. act. daddhf 
(for both dehi and dhehi}; e. the middle forms dadhidhve, dadhisvd, dadhi- 
dhvam (in RV.), with vowel weakened to i instead of being dropped. 

670. A number of roots have been transferred from this 
to the a-class (class VI., below, 749), their reduplicated root 
becoming a stereotyped stem inflected after the manner of ci- 
sterns. These roots are as follows: 

671. In all periods of the language, from the roots stha 
'stand', pa 'drink', and ghra 'smell', are made the presents 


228 IX - PRESENT-SYSTEM. [671 

tisthami, pibami (with irregular sonantizing of the second p : 
later often written pivami), and jiyhrami which then are in- 
flected not like mimami, but like bMvami, as if from the present- 
stems tistha, piba, jighra. 

672. In the Veda, the reduplicated roots da and dha are also sometimes 
turned into the a-stems ddda and dddha, or inflected as if roots dad and 
dadh of class VI.; and single forms of the same character are made from 
other roots: thus, mimanti (jAna 'bellow'), rdrate (j/>a 'give': 3d sing. mid.). 

673. In the Veda, also, a like secondary root, jighn, is made from yhan 
(with omission of the radical vowel, and conversion, usual in this root, of h 
to gh when in contact with n); and some of the forms of safe, from y~sac, 
show the same conversion to an a-stem, safca. 

674. In AB. (viii. 28), a similar secondary form, jighy, is given to j/^/u : 
thus, jighy ati, jighy atu. 

675. A few so-called roots of the first or root-class are the products of 
reduplication, more or less obvious: thus, jaks (640), and probably fas (from 
y~fas) and caks (from j/fcap or a lost root fcas, 'see'). In the Veda is found 
also sa$c, from ysac. 

676. The grammarians reckon (as already noticed, 641) several roots of 
the most evidently reduplicate character as simple, and belonging to the root- 
class. Some of these (jagr, daridra, vevl) are regular intensive stems, and 
will be described below under Intensives (chap. XIV.); didhi, 'shine', along 
with Vedic dldl 'shine' and plpl 'swell', are sometimes also classed as inten- 
sives; but they have not the proper reduplication of such, and may perhaps 
be best noticed here, as reduplicated present-stems with irregularly long 
reduplicating vowel. 

Of pres. indie, occurs in the older language only didyati, 3d pi., with 
the pples dZdyat and didhyat, and mid. didye, didhye, dldhyatham, with the 
pples didyana, didhyana, pipyana. The subj. stems are diddya, dldhaya, 
plpdya, and from them are made forms with both primary (from dlddya] and 
secondary endings (and the irregularly accented didayat and dldayat and 
dldhayan}. No opt. occurs. In impv. we have didiln (and didihi) and pipihi, 
and pipyatam, pipyatam, pipyata. In impf., adldes and pipes, ddldet and 
ddldhet and aplpet (with augmentless forms), aplpema (with strong form of 
root), and adidhayus and (irregular) apipyan. 

A few forms from all the three show transfer to an a-inflection : thus, 
dldhaya and plpaya (impv.), dpipayat, etc. 

Similar forms from y~mi 'bellow' are amimet and mimayat. 

677. The stem cafcas (sometimes cafeap) is also regarded by the gram- 
marians as a root, and supplied as such with tenses outside the present- 
system which, however, hardly occur in genuine use. It is not known 
in the older language. 

678. The root bhas, 'chew', loses its radical vowel in weak forms, taking 
the form baps: thus, bdbhasti, but bdpsati (3 pi.), bdpsat (pple). 

679. The root bhl, 'fear', is allowed by the grammarians to shorten its 
vowel in weak forms : thus, bibhlmas or bibhimas, bibhlyam or bibhiyam. 




680. Forms of this class from yjan, 'give birth', with added i thus, 
jajnise, jajnidhve are given by the grammarians, but do not appear to 
have been found in use. 

681. The roots ci and cit have in the Veda reversion of e to k in the 
root-syllable after the reduplication: thus, cikeai, cikethe (anomalous, for 
cikydthe], cikitam, aciket, cikyat (pple); cikiddhi. 

682. The root vyac has i in the reduplication (from the y), and is con- 
tracted to vie in weak forms : thus, viviktds, dviviktam. So the root hvar 
(if its forms are to be reckoned here) has u in reduplication, and contracts 
to hur : thus, juhurthas. 

III. Nasal Class (seventh, ra///-class). 

683. The roots of this class all end in consonants. And 
their class-sign is a nasal preceding the final consonant : in 
the weak forms, a nasal simply, adapted in character to the 
consonant ; but in the strong forms expanded to the syllable 
^ ndj which has the accent. 

In a few of the verbs of the class, the nasal extends also into other 
tense-systems: they are an}, bhanj, hins : see below, 694. 

1. Present Indicative. 

684. Examples of inflection: a. the root Jsf yuj, 
'join': strong stem-form, 7H5T yundj ; weak, ET^" yunj. 

For the rules of combination of final j, see 219. 



yundjmi yunjvas yunjmds yunje yunjvdhe yunjmahe 

2 yrf^i UC^H^ g^r nf" i^w\ jnr 

yundhsi yunkthds yunkthd yunhse yunjathe yungdhve 

3 uni^ Urh^ rsin g%" ^iTi g^ 1 

**yundkti yunktds yunjdnti yunkte yunjhte yunjdte 

b. the root "^T rudh, 'obstruct'; bases 
For rules of combination of final dh. see 153^ 160. 


runddhmi rundhvds rundhmds rundhe rundhvdhe rundhmahe 



rundtsi runddhds runddhd runtse rundhathe runddhve 


rundddhi runddhds rundhdnti runddhe rundhnte rundhdte 

Instead of yunkthas, yungdhve, and the like (here and in 
the impv. and impf.), it is allowed and more usual (231) to 
write yunthas, yundhve, etc.; and, in like manner, rundhas, rundhe, 
for runddhas, runddhe; and so in other like cases. 

685. Vedic irregularities of inflection are : a. the ordinary use of a 
3d sing. mid. like the 1st sing., as vrnje; b. the accent on te of 3d pi. mid. 
in anjate, indhate, bhunjate. 

2. Present Subjunctive. 

686. The stem is made, as usual, by adding a to the strong 
present-stem : thus, yundja, runddha. Below are given as if 
made from ~\'yuj all the forms for which examples have been 
noted as actually occurring in the older language. 

active. middle. 

s. d. p. s. d. p. 

1 yunajani yundjava yundjama yunajai yundjamahai 

2 yundjas yunajadhvai 

3 yundjat yundjatas yundjan yundjate 

687. The RV. has once anjatas, which is anomalous as being made 
from the weak tense-stem. Forms with double mode-sign are met with: 
thus, trnahan (AV.), yunajan (QB.); and the only quotable example of 3d du. 
act. (besides anjatas) is hinasdtas (^B.). QB. has also hinasavas as 1st du. 
act.: an elsewhere unexampled form. 

3. Present Optative. 

688. The optative is made, as elsewhere, by adding the 
compounded mode-endings to the weak form of present- 
stem. Thus : 

active. middle. 

s. d. . s. d. . 

yunjyhm yunjy&va yunjyhma yunjiyd yunjivdhi yunjimdhi 

etc. etc. etc. etc. etc. etc. 

4. Present Imperative. 

689. In this class (as the roots all end in consonants) 
the ending of the 2d sing. act. is always ftf dhi, 

693] III. NASAL CLASS (SEVENTH, rudh-CLAss}. 231 

active. middle. 

s. d. p. s. d. p. 

i yteiiPs JHsTra u^sllH Zpfff UHslN< UHsllH^ 

yun&jani yundjava yundjama yundjai yundjavahai yimdjamahai 

2 gfrr 1^ J3T 3?^ y^iyiH^ H 7 ^ 

yungdhi yunktam yunktd yunksvd yunjatham yungdhvam 

yundktu yunktam yunjdntu yunktam yunjatam yunjdtam 

690. There is no occurrence, so far as noted, of the ending tat, in verbs 
of this class. The Veda has, as usual, sometimes strong forms, and some- 
times the ending tana, in the 2d pi. act.: thus, undtta, yundkta, anaktana, 

5. Present Participle. 

f 691. The participles are made in this class as in the 
preceding ones: thus, act. ^c^yunjdnt (fern, y^rfl yunjati) ; 
mid. ET^FT yunjand (but RV. has indhana). 

6. Imperfect. 

692. The example of the regular inflection of this tense 
needs no introduction: 

active. middle. 

s. d. p. s. d. p. 


dy uajam dyunjva dyunjma dyunji dyunjvahi dywymahi 

dyunak dyunktam dyunkta dyunkthas dyunjatham dyungdhvam 

dyunak dyunktam dyunjan dyunkta dyunjatam ayunjata 

The endings s and / are necessarily lost in the nasal class 
throughout in 2d and 3d sing, act., unless saved (555) at the 
expense of the final radical consonant : which is a case of very 
rare occurrence : in the older language have been noted only 
ahinat (TB.), 3d sing., for ahinas (\/hins\ and once in AV. 
alhanas, 2d sing., for abhanak (ybhanj)', this last is a case of 
the utmost rarity. 

693. The Veda shows no irregularities in this tense. Occurrences of 
augmentless forms are found, especially in 2d and 3d sing, act., showing 
an accent like that of the present: for example, bhindt, prndk, vrnak, pindk, 


Irregularities of the Nasal Class. 

694. The roots which thus expand a penultimate nasal in 
the strong forms of the present-system into a syllable nd are 
about twenty-five: namely, tac, pro, ric, vie, anj, bhanj, vrj, 
bhuj, yuj, krt 'spin', chrd, trd, chid, bhid, ud, rdh, idh, rudh, 
ubh, ac (anacamahai, once, RV.), pis, cis, hihs, trh. Those here 
written with the nasal namely anj, bhanj, hihs - - have that 
addition also in the other tense-systems. Two, rdh and ubh, 
make present-systems also of other classes with nasal class-signs : 
thus, rdhnoti (cl. IV.), ubhnati (cl. V.). Several have a-stems 
with penultimate nasal: thus, prncd, cihsd, trhhd, umbhd ; and 
occasional a-forms, especially in the later language, are met 
with from others : thus, bhunjet, chindeta, apinsat, arundhat (com- 
pare the nasalized roots of the d-class, below, 758). 

695. The root trh combines trnah with U, tu, etc. into trnedhi, trnedhu ; 
and, according to the grammarians, has also such forms as trnehmi: see 
above, 224 b. 

696. The root hihs (by origin apparently a desiderative from yhari) 
accents irregularly the root syllable in the weak forms : thus, hfhsanti, hfhste 
(but hindsat etc.). 

IV. Nu and ^-classes (fifth and eighth, su and to-classes). 

\ 697. A. The present-stem of the ww-class is made by 
adding to the root the syllable ^ nu, which then in the 
strong forms receives the accent, and is strengthened to ^fino. 
B. The few roots of the w-class (about half-a-dozen) 
end in ^ n, with the exception of the later irregular Sfi kr 
(or kar] for which, see below, 714. The two classes, 
then, are closely correspondent in form ; and they are wholly 
accordant in inflection. 

The u of either class-sign is allowed to be dropped before 
v and m of the 1st du. and 1st pi. endings, except when the 
root (raw-class) ends in a consonant; and the u before a vowel- 
ending becomes v or uv, according as it is preceded by one or 
by two consonants (129). 

1. Present Indicative. 

r 698. Examples of inflection: A. raw-class; roof 
H su, 'press out': strong form of stem, RT suno ; weak form, 
R sunu. 


active. middle. 

s- d. p. s . d. n. 

oo _ 

sunomi sunuvds sunumds sunve sunuvdhe 

H^lfo 0^^ oo^ oo^ U-^lS) 

sunosi sunuthas sunuthd sunuse sunv&the 


sunoti sunutds sunvdnti sunute sunvhte 

The forms sunvds, sunmds, sunvdhe, sunmdhe are alternative 
with those given here for 1st du. and pi., and in practice are 
more common (no examples of the fuller forms have been noted 
from the older language). From }/ap, however (for example), 
only the forms with u can occur : thus, apnuvds, apnumdhe ; and 
also only apnuvdnti, apnuve, apnuvdte. 

B. w-class; root cH tan, 'stretch': strong form of stem. 
rPTT tano ; weak, rT'T tanu. 

tanomi tanvds tanmds tanve tanvdhe tanmdhe 
etc. etc. etc. etc. etc. etc. 

The inflection is so precisely like that given above that it 
is not worth writing out in full. The abbreviated forms in 1st 
du. and pi. are presented here, instead of the fuller, which rarely 
occur (as no double consonant ever precedes). 

699.' In the older language, no strong 2d perss. du. or pi., and no 
t/iana-endings, chance to occur (but they are numerous in the impv. and 
impf.: see below). The RV. has several cases of the irregular accent in 
3d pi. mid.: thus, tanvate, manvate, sprnvate. 

In RV. occur also several 3d pll. mid. in ire from present-stems of this 
class : thus, invire, rnvire, pinvire, crnvire, sunvire, hinvire. Of these, 
pinvire and hinvire might be perfects without reduplication from the second- 
ary roots pinv and hinv (below, 716). The 2d sing. mid. (with passive value) 
prnvise (RV.) is of anomalous and questionable character. 

2. Present Subjunctive. 

700. The subjunctive mode-stem is made in the usual 
manner, by adding a to the gunated and accented class-sign : 
thus, sundva, tandva. In the following scheme are given all the 
forms of which examples have been met with in actual use in 
the older language from either division of the class : some of 
them are quite numerously represented there. 




s. d. p. 

1 sundvani sundvava sundvdma 

2 sundvas sundvatha 

3 sundvat sundvan 

s. d. 

sundvai sundvavahai 
sundvase sundvdithe 
\ sundvdtdi 



701. Of the briefer 1st sing, act., RV. has krnava and hinava. Forms 
with double mode-sign occur (not in RV.): thus, krndvdt and karavdt (AV.); 
afnavdtha (K.), krnavatha (VS.; but -vatha in Kanva-text), karavatha (QB.). 
On the other hand, apnavatai is found once (in TS.). RV. has in a single 
passage krnvdtte (instead of krndvdile): the only form in dithe is apndvaithe. 

3. Present Optative. 

702. The combined endings (566) are added, as usual. 
to the weak tense-stem: thus, 





sunuyama sunviyd sunvivdhi sunvimdhi 

etc. etc. etc. etc. 

From yap, the middle optative would be apnuviyd and 
so in other like cases. 

4. Present Imperative. 

703. The inflection of the imperative is in general like 
that in the preceding classes. As regards the 2d sing, act., 
the rule of the later language is that the ending f^T hi is 
taken whenever the root itself ends in a consonant; other- 
wise, the tense- (or mode-) stem stands by itself as 2d per- 
son (for the earlier usage, see below, 704). Example of in- 
flection is : 



*- - 

sunavani sundvava sundvama sundvai sundvavahai sundvamahai 

sunu sunutam sunutd sunusvd sunvaiham sunudhvdm 


'^ ^5 ^OO j * S S. ^~r -*** >*/>^ -"X N/ --V ^-^ "S 

sunotu sunutam sunvdntu sunutam sunvatam sunvdtam 

707] IV. Nil- AND U- (FIFTH AND EIGHTH, SU- AND tan-} CLASSES. 235 

From yap, the 2d sing. act. would be apnuhi; from > ' ag, 
acnuhi; from ydJirs, dhrsnuhi ; and so on. From yap, too, 
would be made apnuvdntu, cipnuvatham, apnuvatam, apnuvatam. 

704. In the earliest language, the rule as to the omission of hi after 
a root with final vowel does not hold good: in RV., such forms as inuhi, 
krnuhf, cinuhf, dhunuhi, crnuhf, sprnuhi, hinuhi, and tanuhi, sanuhi, are 
thrice as frequent in use as inu, prnu, sunn, tanu, and their like; in AV., 
however, they are not more than one third as frequent; and in the Brahmanas 
they appear only sporadically ; even frnudhi (with dhi) occurs several times 
in RV. The ending tat is found in krnutat and hinutat, and kurutat. The 
strong stem-form is found in 2d du. act. in hinotam; and in 2d pi. act. in 
krnota and krnotana, crnota and crnotana, sun6ta and suntitana, hintita and 
hinotana, and tanota, karota. The ending tana occurs only in the forms 
just quoted. 

5. Present Participle. 

^ 705. The endings 5FFT ant and TR and are added to the 
weak form of tense stem : thus, from ]/g su come act. 
sunvdnt (fern. M^lrfl sunvati), mid. H-^fM sunmmd\ from 
tan, rT^Trr tanvdnt (fern, ri^ril tanvatfy rT^FT tanvand. From 
/>, they are 5fio|ti ojowwt?a< and tUirH apnuvand. 

6. Imperfect. 

/" 706. The combination of augmented stem and endings 
is according to the rules already stated: thus, 

active. middle. 

s. d. p. s. d. p 

dsunavam dsunuva dsunuma dsunvi dsumivahi dsunumahi 

dsunos dsunutam dsunuta dsumithas dsunvatham dsunudhvam 

3 -^ MHHHIH^ yyH^ MH^ri ^H^lrilH^ ^H-^ri 
dsunot dsunutam dsunvan dsunuta dsunvatam dsunvata 

Here, as elsewhere, the briefer forms dsunva, dsunma, 
dsunvahi, dsunmahi are allowed, and more usual, except from 
roots with final consonant, as dhrs; which makes, for example, 
always ddhrsnuma etc., and also ddhrsnuvan, ddhrsnuvi, ddhrmu- 
vatham, ddhrsnuvatam, ddhrsnuvata. 

707. Strong stem-forms and Jana-ending are found only in RV., in 
akrnota, akrnotana. Augmentless forms with accent are minvan, hinvdn, rnutd. 

236 1X - PRESENT-SYSTEM. [708 

Irregularities of the nu and w-classes. 

708. Less than thirty roots form their present-system in 
the manner set forth above, by the addition of the class-sign 
nu to the root : they are aks, ac 'attain', talcs, dabh, cak, sagh, 
ap, dac, r, kr 'make', vr (ur), str, spr, rdh, trp, dhrs, i 'send 1 
(or in: see below, 716), ksi 'destroy', ci, dhi, mi 'prop', hi, u, 
du, cru, su, sku, prus, dhu: and of these, several (as taks, sagh, 
dac, u, skii) have only isolated forms of this class. 

709. The root trp, 'enjoy', is said by the grammarians to retain the n 
of its class-sign unlingualized in the later language where, however, forms 
of conjugation of this class hardly occur; while in the Veda the regular 
change is made: thus, trpnu. 

710. The root cru, 'hear', is contracted to cr before the 
class-sign, forming crno and crnu as stem. Its forms . crnvise 
and crnvire have been noted above (699). 

711. The root dhu in the later language shortens its vowel, 
making the stem-forms dhuno and dhunu (earlier dhuno, dhunu). 

712. The so-called root urnu, treated by the native grammarians as 
dissyllabic and belonging to the root-class (I.), is properly a present-stem of 
this class, with anomalous contraction, from the root vr (or var}. In the 
Veda, it has no forms which are not regularly made according to the nu- 
class; but in the Brahmana language are found sometimes such forms as 
urnauti, as if from an w-root of cl. I. (626); and the grammarians make for 
it a perfect, aorist, future, etc. Its 2d sing. impv. act. is urnu or urnuhi; 
its impf. , aurnos, aurnot; its opt. mid., urnuvita (K.) or urnvltd (TS.). 

713. The roots of the other division, or of the ej-class, 
are extremely few : they are tan, man, van, san ; also ksan (not 
in V.: in (TB., and very rarely later), and kr 'make' (in late 
Vedic and later) ; and BR. assume in of the w-class instead of i 
of the raw-class. 

714. The extremely common root fi kr (or kar\ 'make', 
is in the later language inflected in the present-system ex- 
clusively according to the w-class (being the only root of 
that class not ending in ^ n}. It has the irregularity that in 
the strong form of stem it (as well as the class-sign) has 
the ^ww-strengthening, and that in the weak form it is 
changed to kur, so that the two forms of stem are ^T(T karo 
and gT^ kuru. The class-sign 3 u is always dropped be- 
fore ^ v and q m of the 1st du. and pi., and also before 

v -\ 

IT y of the opt. act. Thus : 

741T VI. ^4-CLASS (FIRST, $Aw-CLA88). 243 

such as bhdvanta (which are very common) are, of course, properly aug- 
mentless imperfects. The Brahmanas (especially QB.) prefer the 2d sing, 
act. in asi and the 3d in at. A 3d pi. in antai (vartantai, KB.) has been 
noted once. RV. has an example, area, of the briefer 1st sing. act. 

3. Present Optative. 

738. The scheme of optative endings as combined with 
the final of an ez-stem was given in full above (566. 

active. middle. 

s. d. p. s. d. p. 


bhdveyam bhdveva bhdvema bhdveya bhdvevahi bhdvemahi 

bhdves bhdvetam bhdveta bhdvethas bhdveyatham bhdvedhvam 

bhdvet bhdvetam bhdveyus bhdveta bhdveyatam bhdveran 

The RV. has once the 3d pi. mid. bharerata (for one other example, 
see 752). 

4. Present Imperative^ 
739. An example of the imperative inflection is : 

active. middle. 

s. d. p. s. d. p. 

bhdvani bhdvava bhdvama bhdvai bhdvavahai bhdvamahai 

- *N ^^ 

"S "V "N. 

bhdva bhdvatam bhdvata bhdvasva bhdvetham bhdvadhvam 

3 WT H^ri IH^ ^ ^ 

bhdvatu bhdvatam bhdvantu bhdvatam bhdvetam bJidvantam 

740. The ending tana in 2d pi. act. is as rare in this whole conjuga- 
tion as is iftana in the present : the V. affords only bhajatana in the a-class 
(and nahyatana in the 7/a-class : 760). The ending tat of 2d sing, act., on 
the other hand, is not rare; the RV. has avatat, osatat, dahatdt, bhavatdt, 
yachatat, yacatat, rdksatat, vahatdt; to which AV. adds jinvatat, dhavatat; 
and the Brahmanas bring other examples. 

5. Present Participle. 

741. The endings ^rT ant and qH mana are added to 
the present-stem, with loss, before the former, of the final 
stem-vowel: thus, act. HcfH bhdvant (fern. ^T^tft Ihdvanti}-, 

mid. HCHM bhdvamana. 



6. Imperfect. 
742. An example of the imperfect inflection is: 

active. middle. 

s. d. p. s. d. p. 

dbhavam dbhavava dbhavama dbhave dbhavavahi dbhavamahi 

dbhavas dbhavatam dbhavata dbhavathas dbhavetham dbhavadhvam 

dbJiavat dbhavatam dbhavan dbhavata dbhavetam dbhavanta 

743. No forms in fana are made in this tense from any o-class. Ex- 
amples of augmentless forms (which are not uncommon) are: cydvam, dvas, 
ddhas, bddhat, bhdrat, cdran, ndpan; badhathas, vdrdhata, fdcanta. The sub- 
junctively used forms of 2d and 3d sing. act. are more frequent than those 
of either of the mote proper subjunctive persons. 

Irregularities of the -class. 

744. A far larger number of roots form their present- 
system according to the a-class than according to any of the 
other classes: in the RV., they are about two hundred and forty 
(nearly two fifths of the whole body of roots) ; in the AV. , about 
two hundred (nearly the same proportion); to tell precisely, or 
very nearly, how many they are in the later language is not 
possible (of the number "about a thousand", as usually stated, 
the greater part are fictitious : see 108 a). Among them are no 
roots ending in long d except a few which make an a-stem 
,in some anomalous way: below, 749 a. 

745. A few verbs have irregular vowel-changes in forming 
the present-stem: thus, 

a. uh 'notice' has ^wna-strengthening (against 240): thus, ohate. 

b. krp (or fcrop), 'lament', on the contrary, remains unchanged : thus, 

c. guh, 'hide', has prolongation instead of guna: thus, guhati. 

d. kram, 'stride', lengthens its vowel in the active, but not in the 
middle: thus, kramati, krdmate; klam, 'tire', is said to form klamati etc., 
but hardly occurs; cam with the preposition a, 'rinse the mouth', forms 
a cam at i. 

e. In the later language are found occasional forms of this class from 
mrj, 'wipe'; and they show the same vrddhi (instead of guna} which belongs 
to the root in its more proper inflection (627): thus, marjasva. 

f. The grammarians give a number of roots in urv, which they declare 
to lengthen the u in the present-stem. Only three are found in (quite 

751] VI. A-CLASS (FIRST, b/lU-CLASS). 245 

limited) use, and they show no forms anywhere with short u. All appear 
to be of secondary formation from roots in r or ar. The root murch or 
murch, 'coagulate', has likewise only u in quotable forms. 

g. The onomatopoetic root athiv, 'spew', is written by the grammarians 
as rthiv. and declared to lengthen its vowel in the present-system. 

746. The roots danc, 'bite', ranj, 'color', sanj, 'hang', 
svanj, ''embrace'., of which the nasal is in other parts of the 
conjugation not constant, lose it in the present-system : thus, 
ddcati etc.; sanj forms both sajati and sajjati (probably for sajyati, 
or for sasjati from sasajati] ; math or manth has mathati later. 
In general, as the present of this class is a strengthening form- 
ation, a root that has such a nasal anywhere has it here also. 

747. The roots gam, 'go', and yam, 'furnish', make the 
present-stems gacha and ydcha: thus, gdchami etc.: see 608. 

748. The root sad, 'sit', forms sida (conjectured to be 
contracted from stsd for sisad) : thus, sidami etc. 

749. Transfers to this class from other classes are not rare, 
as has been already pointed out above, both throughout the 
present-system and in occasional forms. The most noticeable 
cases are the following : 

a. The roots in a, stha, 'stand', pa, 'drink', and ghra, 
'smell', form the present-stems tistha (tisthami etc.), piba or (later) 
piva (pibami etc.), and jighra (jighrami etc.); and, in the Veda, 
da, 'give', and dha, 'place', form sometimes ddda and dddha, 
han, 'slay', forms sometimes jighna, and hi, 'impel', forms jighya 
all these by transfer from the reduplicating class : see 671 4. 

b. Secondary root-forms like inv, jinv, pinv, from simpler 
roots of the raw-class, are either found alongside their originals, 
or have crowded these out of use: see 716. 

750. On the other hand, the root dham or dhma, 'blow', 
forms its present-stem from the more original form of the root : 
thus, dhdmati etc. 

VII. Accented -class (sixth, ?W-class). 

' 751. The present-stem of this class has the accent on 
the class-sign ^ a, and the root remains unstrengthened. In 
its whole inflection, it follows so closely the model of the 
preceding class, that to give the paradigm in full will be 
unnecessary (only for the subjunctive, all the forms found 
to occur will be instanced). 


752. Example of inflection : root fifST vi$, 'enter'; stem, 


1. Present Indicative. 

active. middle. 

s. d. p. s. d. p. 

vicGmi victlvas vicamas vice vicavahe vicamahe 

etc. etc. etc. etc. etc. etc. 

2. Present Subjunctive. 

i vifdni vifdva vifdma vi$ai vi$dvahai vifdmahai 

{vifdsi . i, (vi?dse 

, vi(atha { , _ vicaithe 

vifas \vi?a8ai 

(vicdti j_ . _i fvifate , 

3 : , vi^atas vi$an < , _ vifatte vicantai 

[vifat [vifatai 

A single example of the briefer 1st sing. act. is mrksd. The only forms 
in aithe and aite are prnafthe and yuvafte. 

3. Present Optative. 

viffyam viceva vicema viceya vicevahi viqemahi 

etc. etc. etc. etc. etc. etc. 

The RV. has the ending tana once in tiretana 2d pi. act., and rata in 
juserata 3d pi. mid. 

4. Present Imperative. 

The first persons having been given above as subjunc- 
tives, the second are added here: 

2 f%5F felrFT^ felrT N^IH N^I&IIH^ feiy^ 
mca vicdtam vicdta vicAsva vicetham vicddhvam 

etc. etc. etc. etc. etc. etc. 

, The ending tat is found in RV. and AV. in mrdatat, vrhatdt, suvatat; 
other examples are not infrequent in the Brahmana language : thus, khidatat, 

5. Present Participle. 

The active participle is fek\i\^vi$dntf the middle is 
i^^WM vigdmana. 

The feminine of the active participle is usually made from the strong 
stem-form: thus, vifdnti; but sometimes from the weak: thus, sincantl and 
sincati (RV. and AV.), tuddntl and tudati (AV.): see above, 449 b. 


6. Imperfect. 

i ^foSR^ ^fciN tjfoiiH ^rf^tl 4iief!MNf^ ytomJ^ 

dvicam dvicava dvicama dvice dvicavahi dvicamahi 

etc. etc. etc. etc. etc. etc. 

Examples of augmentless forms accented are srjds, srjdt, tirdnta. 
The a-aorist (846 ff.) is in general the equivalent, as regards its form, 
of an imperfect of this class. 

Irregularities of the a-class. 

753. It is impossible to determine closely the limits of 
this class, partly because of the occurrence of forms unaccented, 
or in unaccentuated texts, which might belong either to it or 
to the preceding class, partly because its modes and imperfect 
are accordant in form with those of the a-aorist (below, chap. 
XI.), and their separation is not always practicable, and partly 
for other reasons. With considerable confidence may be reckoned 
as belonging to it about seventy roots: namely, ksi, yu 'join', 
ru 'roar', su (or su) 'stir up', dhu, hu, kr 'strew', gr 'swallow', 
tr, rikh or likh, sic, ich, vij, khid, vid 'find', vidh, ksip, lip, riph, 
die, pic, ric, vie, is, tvis, mis, muc, uch, ubj, tuj, ruj, khud, tud, 
nud, rud, lup, ubh, cubh, gur, jur, tur, bhur, sphur, jus, prus, 
rus, cus, uks, vrcc (or vracc], rch, prch (or prach], rnj, srj, bhrjj 
(or hhrajf), mrd. prn, mrn, krt 'cut', crt, rd, trp, vnrc, sprc, rs 
'push', krs 'plough', mrks, vrs, drh, vrh or brh. Some even of 
these have either only isolated or very rare occurrences of a-forms. 
The roots ich, uch, and rch are reckoned as substitutes in the 
present-system for is 'wish', vas 'shine', and r 'go to' (608). 
Prn and mrn have been noticed above (731) as secondary roots 
from present-stems of the na-class (V.). 

754. Certain peculiarities of this "body of roots are very noticeable: it 
contains only one or two roots with long vowels, and none with long interior 
vowels ; very few with final vowels ; and none with a as radical vowel, except 
as this forms a combination with r, which is then reduced in the present- 
system, as in the weak forms generally, to r or some of the usual sub- 
stitutes of r. 

755. The roots in i and u and u change those vowels into 
iy and uv before the class-sign : thus, ksiydti, suvdti (sva instead 
of suva occurs in AV.; and the Brahmanas have forms in ksya 
from ksi}. 

756. The three roots in r form the present-stems kird, 
gird, tird, and they are sometimes written as kir etc.; and gur, 
jur, tur are really only varieties of gr, jr, tr; and bhur and 

sphur are evidently related with other ar or r root-forms. 


757. Two other roots which are used only in middle forms, and in 
combination with the preposition a (sometimes further combined), make the 
present-stems a-driya and a-priyd, and are reckoned as r or ar roots: dr, 
'regard', and pr, 'be busy' (neither is found in V.). It is a question whether 
they are more properly reckoned to this class or as passives ; and the same 
question arises as to the stems mriyd and dhriyd, from the roots mr, 'die', 
and dhr, 'hold': see below, 773. 

758. Although the present-stem of this class shows in 
general a weak form of the root, there are nevertheless a number 
of roots belonging to it which are strengthened by a penultimate 
nasal. Thus, the stem muncd is made from ymuc, 'release'; 
sincd from l/sic, 'sprinkle'; vindd from yvid 'find'; krntd from 
ykrt 'cut'; pined from ypig, 'adorn'; trmpd from }/trp, 'enjoy': 
lumpd from yiup, 'break'; limpd from yiip, 'smear'; and occa- 
sional forms of the same character are met with from a few 
others, as tundd from ytud, 'thrust'; umbha from yttbh, 'hold'; 
brn/id from ybr/i 'strengthen'; drnhd (beside drnhd] from ydrh, 
'make firm'; cumbhd (beside qumbha] from yculh, 'shine'. TS. 
has crnthati from y$rath (instead of crathriati). 

VIII. Fa-class (fourth, <%'-class). 

( 759. The present stem of this class adds TJ ya to the 
accented but unstrengthened root. Its inflection is also pre- 
cisely like that of the #-class 7 and may be presented in the 
same abbreviated form as that of the a-class. 

^760. Example of inflection: root ^ na ^^ 'bind'; 
stem ^T^T ndhya. 

1. Present Indicative. 

active. middle. 

d. . s. d. 

ndhyami ndhyavas ndhyamas ndhye ndhyavahe ndhyamahe 

etc. etc. etc. etc. etc. etc. 

2. Present Subjunctive. 

1 ndhydni ndhyama ndhyai ndhyavahai ndhydmahai 

fndhyasi ,, _,, _. 

2 { ndhyasai nahyadhvai 

3-1 , ndhyatas ndhyan ndhydtdi ndhydntdi 

\ncthy at 

A 3d pi. mid. in antai (jdyantdi) occurs once in TS. 

76 !] VHI. Ytf-CLASS (FOURTH. rfw/'-CLASS . 249 

3. Present Optative. 

ndhyeyam ndhyeva ndhyema ndhyeya ndhyevahi ndhyemahi 

etc - etc - etc. etc. etc. etc. 

4. Present Imperative. 

2 ^ 

ndhya ndhyatam ndhyata ndhyasva ndhyetkdm ndhyadhvam 

etc. etc. etc. etc. etc. etc. 

Of the ending tana, RV. has one example, nahyatana; the ending tat 
is found in asyatdt, chyatdt. 

5. Present Participle. 

The active participle is *1$lvX ndhyant (fern, ^rft nd- 
hyantl)', the middle is ^RH ndhyamana. 

6. Imperfect 

dnahyam dmhyava dnahyama dnahye dnahydvahi dnahyamahi 

etc. etc. etc. etc. etc. etc. 

Examples of augmentless forms showing the accent belonging to the 
present-system are gdyat, pdfyat, pdfyan, jdyathds. 

Irregularities of the ya-el&sa. 

761. The roots of the ya-class are more than a hundred 
in number. They may be grouped as follows : 

a. Roots signifying a state of feeling, or a condition of mind or body. 
These are nearly half the whole class. They are (alphabetically) as follows : 
we, 'be pleased'; kup, 'be angry'; fcrp, 'be lean'; klam, 'be weary'; krudh, 
'be angry'; fcsam, 'be patient'; ksudh, 'be hungry'; ksubh, 'be agitated 1 ; grdh, 
'be greedy'; jas, 'be worn out 1 ; tarn, 'be exhausted 1 ; tus, 'be satisfied'; trp, 
'be pleased'; trs, 'be thirsty'; tras, 'be alarmed'; dam, 'be submissive 1 ; dus, 
'be spoiled 1 ; drp, 'be crazed'; druh, 'be hostile'; pus, 'be in good condition'; 
budh, 'be awake'; bhram, 'be unsteady'; mad, 'be excited 1 ; man, 'be minded'; 
muh, 'be confused'; mrs, 'be forgetful'; mrit, 'be in ill condition'; med, 
'be fat'; yas, 'be ardent'; yudh, 'be hostile'; raj, 'be colored'; radh, 'be sub- 
ject'; ran, 'be happy'; rddh, 'be successful'; ris, 'be hurt'; rup, 'be in pain'; 
lubh, 'be lustful'; pam, 'be quiet'; pwc, 'be in pain'; fus, 'be dry'; pram, 
'be weary'; ftar, 'be gratified'; hrs, 'be excited 1 ; and we may perhaps add 
das, 'be deficient', and nap, 'be missing 1 . Some of these are of only early- 
use, some only of later ; and some have only sporadic forms of this class, 
made perhaps under the influence of the analogy of the others. 

250 IX - PRESENT-SYSTEM. [761 

b. Roots which have a more or less distinctly passive sense, and which 
are in part evident and in part presumable transfers from the passive or 
i/a-class, with change of accent, and sometimes also with assumption of active 
endings. It is not possible to draw precisely the limits of the division, or 
determine in all cases where passive form and meaning pass into intransitive ; 
but there are a number of clear cases, where in the older language the accent 
wavers and changes, and the others are to be judged by their analogy. Thus, 
muc forms mucyate once or twice, beside usual mucydte, in RV. and AV.; 
and in the Brahmanas the former is the regular accent : and similar changes 
are found in other verbs : thus, ji or jya, ksi 'destroy', ha 'leave', pac, dr 
'burst', chid, bhid. Cases closely analogous with these are miyate etc. from 
ymi or ml, 'lessen'; ricyate etc. from yric, 'leave'; viyate etc. from j/w, 
'impregnate'; fiyante from y^ya, 'coagulate'; tfsyate etc. from Yfis, 'leave'; 
drhyasva from ydrhh, 'make firm'; puryate etc. from j/pr, 'fill'; and Itipyate, 
tdpyate, tlryate, klfyyate, fdhyate, may be ranked along with them. Active 
forms are early made sporadically from some of these thus, drhya (RV.), 
ksiyati and puryati (TA.); and dlryati, klipyati, and other like cases, are 
found later. The AV. has jlryati, 'grows old' (later also jiryate)', and QB. 
has aprusyat, 'was sprinkled'. And from the earliest period jdyate etc., 'is 
born', is either altered passive or original ya-lormation from y'ja, serving as 
complement to yjan, 'give birth'. 

c. A small body of roots are either transitive, or not intransitive in a 
way that clearly connects them with either of the above classes : thus, as 
'throw'; is 'send'; fra, 'save'; nah, 'bind'; pap, 'see'; vyadh, 'split'; slv, 'sew'; 
diu, 'play'; tur, 'overcome' (RV., once); tan, 'thunder' (RV., once), r/, 'press 
on'; nri, 'dance'; pad, 'go'; t>ap, 'bleat'; dl, 'hover'; n, 'flow'; srlv, 'fail'; 
ylifj 'hang on'; 6/irap, 'fall'; sidh, 'succeed'; dip, 'shine' (and perhaps das 
and nap are better classed here than under a). 

d. A body of roots, of various meaning, and of somewhat questionable 
character and relations, which are by the native grammarians reckoned as ending 
with diphthongs : thus, 

1. Roots reckoned as ending in ai and belonging to the a-class : thus, 
gdyati from ygai. As these show abundantly (and in most cases exclusively) 
a-forms outside the present-system, there seems no reason why they should 
not be regarded as a-roots of the They are: ga 'sing'; .gla, 'be 
wearied or disgusted': dhya, 'think'; pya, 'swell'; mia, 'wither'; ra, 'bark'; 
va, 'droop'; fj/a, 'coagulate'; pra, 'cook'; stya, 'be coagulated'; and, in one 
or two sporadic forms, fca, 'burn'; da 'cleanse'; sta, 'be hidden'; spha, 'be 
fat'. Tra, 'save', was given in the preceding division. Many of these are 
evident extensions of simpler roots with added a. With them may be 
mentioned tay, 'extend' (compare pass, tayate from ytan: 772), and c%, 
'be shy or anxious' (which connects itself with uses of yd}. 

2. Roots reckoned as ending in e and belonging to the a-class: thus, 
dhdyati from ydhe. These, too, have a-forms, and sometimes t-forms, outside 
the present-system, and must be regarded as a-roots, either with a weakened 
to a before the class-sign of this class, or with a weakened to I or i and 

767] VIII. Ya-CLAss (FOURTH, div-CLAss). 251 

inflected according to the a-class. They are : dha, 'suck'; va, 'weave'; vya, 
'hide'; hva, 'call' (one of the forms of y7w); and a late example or two are 
found from ma, 'exchange'. With them may be mentioned day, 'share, 
sympathize, pity'; vyay, 'be 'wasted' (denom. of vyaya?}-, cay, 'visit with 
retribution' (probably a form of ci), 

3. Roots artificially marked with a final o (108c) and reckoned to this 
class, the radical vowel being declared dropped before the class-sign : thus, 
dydti from do. They have, as showing an accented yd, no real right to be 
classed here at all, but seem more accordant in formation with the present- 
stems sva and ksya, noticed under the preceding class (755). Outside these 
present-systems, they show a and i-forms; and the ya in the only RV. oc- 
currence, and in most of the AV. occurrences, is resolved into ia which 
in the true class-sign ya is the case only in very rare and purely sporadic 
instances. They might, then, perhaps be best viewed as a-roots with a weak- 
ened to i, and inflected by the a-class, but without the usual conversion of 
i to iy (755). They are: da 'cut', da 'bind'; ca, 'sharpen'; sa, 'bind'; cha, 
'cut off'. 

762. The j/a-class is the only one thus far described which shows any 
tendency toward a restriction to a certain variety of meaning. In this ten- 
dency, as well as in the form of its sign, it appears related with the class 
of distinctly defined meaning which is next to be taken up the passive, 
with t/a-sign. Though very far from being as widely used as the latter beside 
other present-systems, it is in no very small number of cases an intransitive 
conjugation by the side of a transitive of some other class. 

763. The roots of this class ending in am lengthen their 
vowel in forming the present-stem : they are tarn, dam, bhram 
(but bhramyat also occurs), cam, cram, klam (hardly found in 
use), and ksam (but ksamyate also) : for example, camyati, crhmyati. 

764. The root mad has the same lengthening: thus, madyati. 

765. The roots in w namely, div, siv, sriv (or criv) 
are written by the grammarians with iv, and a similar lengthening 
in the present-system is prescribed for them. 

They appear to be properly dm etc., since their vocalized final in other 
forms is always u : div is by this proved to have nothing to do with the 
assumed root div, 'shine', which changes to dyu (361 d): compare also the 
desiderative stem jujyusa from yjlv (1028h). 

766. From the roots jr and tr (also written as jur and tir or tur] come 
the stems jirya and tirya, and j&rya and tUrya (the last two only in RV.); 
from pr comes pUrya. 

767. The root vyadh is abbreviated to vidh: thus, vfdhyati. And any 
root which in other forms has a penultimate nasal loses it here : thus, drhya 
from drnh or drh; bhrdfya (also bhfcya) from bhrahf or bhraf; rajya from 
ran? or raj. 


IX. Accented //</-class: Passive conjugation. 
768. A certain form of present-stem, inflected with 
middle endings, is used only with a passive meaning, and 
is formed from all roots for which there is occasion to make 
a passive conjugation. Its sign is an accented 7J yd added 
to the root: thus, t^*y hanyd from j/^T han, ETTQJ apyd 
from j/ETPT ap } JT^T grhyd from J/JT^n grh (or grah}\ and 

"X C. (. 

so on, without any reference to the class according to which 
the active and middle forms are made. 

\ 769. The form of the root to which the passive-sign is 
added is (since the accent is on the sign) the weak one : thus, 
a penultimate nasal is dropped, and any abbreviation which is 
made in the weak forms of the perfect (794), in the aorist 
optative (922 b), or before to, of the passive participle (954), is 
made also in the passive present-system : thus, ajya from j/aw/, 
badhyd from ybandh, ucya from j/wzc, ijyd from yyaj. 

770. On the other hand, a final vowel of a root is in 
general liable to the same changes as in other parts of the 
verbal system where it is followed by y: thus., 

a. i and u final are lengthened: thus, miya from ymi; suyd from ysu; 

b. a final is usually changed to i : thus, dlyd from yda ; hiyd from 
yha: but jnayd from yjna, khyayd from ykhyd; 

C. r final is in general changed to ri: thus, kriyd from ykr- but if 
preceded by two consonants (and also, it is claimed, in the root r), it has 
instead the pwna-strengthening: thus, smart/a from j/smr, staryd from ytr; 
and in those roots which show a change of r to ir and ur (so-called f- 
verbs : see 242), that change is made here also, and the vowel is lengthened : 
thus, firyd from y$r; puryd from ypr. 

\ 771. The inflection of the passive-stem is precisely like 
that of the other a- stems ; it differs only in accent from 
that of the class last given. It may be here presented, 
therefore, in the same abbreviated form: 

Example of inflection: root ^i kr, 'make'; passive- 
stem T5FTET kriyd: 

1. Present Indicative. 

d. p. 

kriye kriyavahe kriy&mahe 

etc. etc. etc. 


2. Present Subjunctive. 

The forms noticed as occurring in the older language are 
alone here instanced : 

a. p. 

1 kriyai kriydmahai 

2 kriyddhvai 

3 \kriydtai 

The 3d pi. ending antai is found once (ucyantai, K.). 

3. Present Optative. 

kriyeya kriyevahi kriyemahi 
etc. etc. etc. 

No forms of the passive optative chance to occur in RV. or AV.; they 
are found, however, in the Brahmanas. 

4. Present Imperative. 

kriydsva hriyetham kriyddhvam 

etc. etc. etc. 

5. Present Participle. 

This is made with the suffix JTR mana: thus, 

In use, this participle is well distinguished from the other passive par- 
ticiple hy its distinctively present meaning: thus, krtd, 'done'; but kriydmana, 
'in process of doing', or 'being done'. 

6. Imperfect. 

dkriye dkriyavahi dkriyamahi 

etc. etc. etc. 

The passive-sign is never resolved into ia in the Veda. 

772. The roots tan and khan usually form their passives 
from parallel roots in a: thus, taydte, khaydte (but also tanydte, 
khanydte). The corresponding form to |//aw, namely jhyate (above, 
761 b), is apparently a transfer to the preceding class. 

773. By their form, mriydie, 'he dies', and dhriydte, 'he 
maintains himself, is steadfast', are passives from the roots mr, 
'die', and dAr, 'hold'; although neither is used in a proper 
passive sense, and mr is not transitive except in the derivative 
form mrn (above, 731). With them are to be compared the 
stems a-driyd and a-priyd (above, 757), which may possibly be 


peculiar adaptations of meaning of passives from the roots pr, 
'fill', and dr, 'scatter . 

774. Instances are occasionally found in the later language of an 
apparent assumption of active instead of middle endings by passive persons 
of the present-system. Probably, however, these are rather to be regarded 
as examples of transfer to the t/a-class, such as were considered above (761 b). 

775. As was pointed out above (607), the formation and 
inflection of stems in dya (the tenth or cwr-class of the Hindu 
grammarians) will be treated under the head of secondary con- 
jugation (chap. XIV.), along with the intensive and desiderative 
formations, because, in all alike, the stem is not a present-stem 
merely, but has been extended also into other tense-systems. 

Uses of the Present and Imperfect. 

776. The uses of the mode-forms of the present-system 
have been already briefly treated in the preceding chapter (572 ff.). 
The tense-uses of the two indicative tenses, present and imper- 
fect, call here for only a word or two of explanation. 

777. The present has, besides its strictly present use, the 
same side-uses which belong in general to the tense : namely, 
the expression of habitual action, of future action, and of past 
in lively narration. 

a. Examples of future meaning are: abruvan hrstd gachdmo vayam apt 
(MBh.), 'they said with gladness, "we will go too'"; agnir atmabhavam prdddd 
yatra vdnchati ndisadhah (MBh.), 'Agni gave his own presence wherever the 
Nishadhan should desire'. 

b. Examples of past meaning are : tittard sur ddharah putra dsld ddnuh 
$aye sahdvatsd nd dhenuh (RV.), 'the mother was over, the son under; there 
Danu lies, like a cow with her calf; prahasanti ca tdm kecid abhyasuyanti 
cd 'pare akurvata daydm kecit (MBh.), 'some ridicule her, some revile her, 
some pitied her'; tato yasya vacandt tatrd 'valambitds tarn sarve tiraskurvanti 
(H.), 'thereupon they all fall to reproaching him by whose advice they had 
alighted there'. 

778. In connection with certain particles, the present has 
rather more definitely the value of a past tense. Thus: 

a. With pwrcf, 'formerly': thus, saptarsin u ha sma vdi purd rkshd ity 
dcaksate (QB.), 'the seven sages, namely, are of old called the bears'; tan- 
mdtram api cen mahyam na daddti purd bhavdn (MBh.), 'if you have never 
before given me even an atom". 

b. With the asseverative particle sma.' thus, frdmena ha sma vdt tad 
devd jayanti ydd esdm jdyyam dsd rsayaf ca (QB.), 'for, in truth, both gods 
and sages were wont to win by penance what was to be won'; dvistah kalind 
dyute jlyate sma nalas tadd (MBh.), 'then Nala, being possessed by Kali, 
was beaten in play'. 


No example of this construction is found in either RV. or AV., or 
elsewhere in the metrical parts of the Veda. In the Brahmanas, only habitual 
action is expressed by it (Delbruck). In all periods of the language, the 
use of sma with a verb as pure asseverative particle, with no effect on the 
tense-meaning, is very common; and the examples later are hardly to be 
distinguished from the present of lively narration of which the whole 
construction is doubtless a form. 

779. The imperfect has remained unchanged in value through 
the whole history of the language : it is the tense of narration ; 
it expresses simple past time, without any other implication. 

Compare what is said later (end of chap. X. and chap. XI.) as to the 
value of the other past tenses, the perfect and aorist. 



L 780. THE perfect-system in the later language, as has 
been seen above (535), consists only of an indicative tense 
and a participle both of them in the two voices, active and 

In the oldest language, the perfect has also its modes and 
its augment-preterit, or pluperfect, or is not less full in its 
apparatus of forms than is the present-system. 

^ 781. The formation of the perfect is essentially alike 
in all verbs, differences among tbem being of only subord- 
inate consequence, or having the character of irregularities. 
The characteristics of the formation are these : 

a. a stem made by reduplication of the root; 

b. a distinction between stronger and weaker forms of 
stem, the former being used (as in presents of the First 
conjugation) in the singular active, the latter in all other 
persons ; 

c. endings in some respects peculiar, unlike those of 
the present; 

d. the frequent use, especially in the later language, 
of a union-vowel * between stem and endings. 

256 x - PERFECT-SYSTEM. [732 

\ 782. Reduplication. In roots beginning with a con- 
sonant, the reduplication which forms the perfect-stem is 
of the same character with that which forms the present- 
stem of the reduplicating conjugation-class (II. : see 643) 
hut with this exception, that radical 5[ a and 5TT ft and 
ff r (or 3^|" ar) have only 5f a, and never ^ i, as vowel of 
the reduplicating syllable : thus, from y^ pr, 'fill', comes 
the present-stem fqq pip?, but the perfect-stem qq papr ; 
from j/JTT, 'measure', comes the present-stem MHI mima, 
but the perfect-stem r^fT mama; and so on. 

Irregularities of roots with initial consonants will be given below, 784. 

^ 783. For roots beginning with a vowel, the rules of 
reduplication are these : 

a. A root with initial f a before a single final con- 
sonant repeats the f a, which then fuses with the radical 
vowel to 51T a (throughout the whole inflection): thus, 

ad from y^ ad; and in like manner ENs^a/, 5TR an, 
as, ETFJ ah. 

The root TR r forms likewise throughout Eff^" ar (as if 
from ET^" ar]. 

b. A root with ^ i or 3 u before a single final conso- 
nant follows the same analogy, except in the strong forms 
(sing, act.), where the vowel of the radical syllable has 
guna, becoming ^ e or EfT o; before this, the reduplicating 
vowel maintains its independent form, and is separated from 
the radical syllable by its own semivowel: thus, from y '^T 
35 comes ^T is in weak forms, but tm iyes in strong; from 
1/3*1 uc, in like manner, come ^T uc and 3^Trf uvoc. 

*N -\ -V 

The root 5 i, a single vowel, falls under this rule, and 
forms ^1 ly and ^ET iye. 

c. Roots which begin with vowels long by nature or by 
position do not in general make a perfect-system, but use 
instead a periphrastic formation, in which the perfect tense 




of an auxiliary verb is added to the accusative of a verbal 
noun (see below, chap. XV.: 1070 ff.). 

To this rule, however, yap (probably originally ap: 1087 fj 
constitutes an exception, making the constant perfect-stem ap 
(as if from ap : above., a). 

For the peculiar reduplication an, belonging to certain roots with initial 
vowels, see below, 788. 

784. A number of roots beginning with va and ending 
with a single consonant, which in various of their verbal forms 
and derivatives abbreviate the va to u, do it also in the perfect, 
and are treated like roots with initial u (above, 783 b), except 
that they retain the full form of root in the strong persons of 
the singular active. Thus, from yvac come uc and uvac ; from 
\/vas come us and uvas ; and so on. 

The roots showing this abbreviation are vac, vad, vap, vac, 
cas, vah ; and va, 'weave' (so-called ve : 761 d, 2), is said to 
follow the same rule. 

A single root beginning with ya, namely yaj, f offer', has 
the same contraction, forming the stems iyaj and If. 

785. A number of roots having ya after a first initial con- 
sonant take t (from the y] instead of a in the reduplicating 
syllable : thus, from yvyac comes vivyac ; from Ypya comes pipya. 

These roots are vyac, vyath, vyadh, vya, jya, pya; and, in the Veda, 
also tyaj, with cyu and dyu, which have the root-vowel u. 

A single root with va is treated in the same way : namely 
svap, which forms susvap. 

These roots are for the most part abbreviated in the weak forms : see 
below, 794. 

786. A considerable number of roots have in the Veda a 
long vowel in their reduplication. 

Thus, of roots reduplicating with a: kan, kip, gr 'wake', grdh, tan, trp, 
trs, dhr, dhrs, nam, mdh, mrj, mrp, ran, radh, rabh, vale, van, vat, vas 
'attack', vrj, vrt, vrdli, pad, sah, skambh. Some of these occur only in 
isolated cases ; some have also forms with short vowel. Most are Vedic only ; but 
dadhdra is common also in the Brahmana language, and is even found later. 
For jagr, see 1020 below. 

Of roots reduplicating with I: the so-called roots (676) dldhl and dldl, 
which make the perfect from the same stem with the present : thus, dldetha, 
diddya ; dldhima, didhyus (also didhiyus, dldiyus). But plpl has pipye, pipyus, 
etc., with short i. In AV. occurs once jlhida. 

Of roots reduplicating with u: ju and fit (or pt>5). 

787. A few roots beginning with the (derivative: 42) palatal mutes 
and aspiration show a reversion to the more original guttural in the radical 
Whitney, Grammar. 17 

258 x - PERFECT-SYSTEM. [737 

syllable after the reduplication: thus, yd forms ciki; }/cit forms cikit; yji 
forms jigi; yhi forms jighi; yhan forms jaghan (and the same reversions 
appear in other reduplicated forms of these roots). A root da, 'protect', is 
said by the grammarians to form digi: but neither root nor perfect is quotable. 

788. A small number of roots with initial a or r (ar) show 
the anomalous reduplication an in the perfect. 

Thus, in the Veda: 

yanj or aj, which forms the pres. andkti (cl. HI.), has the perfect 
anaje etc. (with anaja and anajyat); 

]/ap, 'attain' (from which comes once anacamahai], has the weak forms 
anapma etc. (with opt. anapt/am), and the strong forms ananpa and anapa 
along with the regular apa etc.; 

yrdh (from which comes once rnddhat) has anrdhe; 

y<rc or arc has anrctis and anrce- 

yarh has (in TS.) anrhus; 

anaha (RV., once) has been referred to a root ah, elsewhere unknown, 
and explained as of this formation ; but with altogether doubtful propriety. 

The later grammar, then, sets up the rule that roots beginning with a 
and ending with more than one consonant have an as their regular redupli- 
cation ; and such perfects are taught from roots like afcs, ar;, and anc or ac ; 
but the only other quotable forms appear to be anarchat (MBh.) and anarsat 
(TA.); which are accordingly reckoned as "pluperfects". 

789. One or two individual cases of irregularity are the 
following : 

a. The extremely common root bhu, 'be', has the anoma- 
lous reduplication ba, forming the stem babhu; and, in the Veda, 
ysu forms in like manner sasu. 

b. The root bhr, 'bear', has in the Veda the anomalous reduplication 
ja (as also in intensive: 1002b): but RV. has once also the regular babhre. 

c. The root sthiv, 'spew', forms either tisthiv (B. et al.) or tisthlv. 

d. Vivakvan (RV., once) is doubtless participle of yvac, with irregular 
redublication (as in the present, 660). 

790. Absence of reduplication is met with in the follow- 
ing cases : 

a. The root vid 'know' has, from the earliest period to the 
latest, a perfect without reduplication, but otherwise regularly 
made and inflected: thus, veda, vettha, etc., pple vidv&ns. It 
has the meaning of a present. The root vid 'find' forms the 
regular viveda. 

b. A few other apparently perfect forms lacking a reduplication are 
found in RV.: they are taksathus, yamdtus, skambhdthus and skambhus, nindima 
(for ninidima?), dhise and dhire (? ydha), and vidre and arhireC? see 613). 
And AV. has cetatus. The participial words dafvdns, midhvdns, sahvahs are 
common in the oldest language. 


c. One or two sporadic cases have been quoted from the later language : 
namely, fahsus and pansire (MBh.). 

791. For an anomalous case or two of reduplicated preposition, see 
below, 1087f. 

792. Strong and weak stem-forms. In the three 
persons of the singular active, the root-syllable is accented, 
and exhibits usually a stronger form than in the rest of the 
tense-inflection. The difference is effected partly by strength- 
ening the root in the three persons referred to, partly by 
weakening it in the others, partly by doing both. 
/ 793. As regards the strengthening : 

a. A final vowel takes either the guna or vrddhi change 
in 1st sing, act., guna in 2d, and vrddhi in 3d: thus, from 
ytf bhl, 1st fsft bibhe or f^R bibhai ; 2d i^R bibhe; 3d 
bibhai; from y%\ kr, 1st ^Fu|" cakdr or *Jtil^ cakar, 2d 
cakdr, 3d ^37T|" cakdr. 

But the u of ybhu remains unchanged, and adds v before 
a vowel-ending : thus, babhuva etc. 

to. Medial % a before a single final consonant follows 
the analogy of a final vowel, and is lengthened or vriddhied 
in the 3d. sing., and optionally in the first: thus, from 
tap, 1st cfflWatdp or t\mj,atap, 2d t&V{tatdp, 3d 

In the Veda, however, the weaker of the two forms allowed by these 
rules in the first person is almost exclusively in use: thus, 1st only bibhdya, 
tatdpa, 3d bibhdya, tat&pa. The only exceptions noticed are cakdra and 
jagrdha (doubtful reading) in AV. 

c. A medial short vowel has in all three persons alike 
the guna- strengthening (where this is possible : 240) : thus, 
from j/gqj druh comes l^FJ dudroh; from j/fipT^V comes 
\d^\ viveg; from y^fift krt comes r^\^cakdrt. 

d. These rules are said by the grammarians to apply to 
the 2d sing, always when it has simple tha as ending ; if it has 
itha (below, 797), the accent is allowed to fall on any one of 
the syllables of the word, and the root-syllable if unaccented has 
sometimes the weak form (namely, in contracted stems with e for 
medial a : below, 794 e ; and in certain other verbs : thus, vivijithd . 


260 X. PERFECT-SYSTEM. [793 

The earlier language, however, appears to afford no example of a 2d 
sing., whatever its ending, accented on any other than the radical syllable, 
or failing to conform to the rules of strengthening as given above (in a, b, c). 

e. Sporadic instances of a strengthening in other than the singular 
persons are found in RV.: thus, yuyopimd, vivefus. And the roots pr, pr, 
and dr 'tear j are said by the grammarians to have the strong stem in the 
weak forms; and jr 'decay' to be allowed to do the same. 

f. The root mrj has (as in the present-system : 627) vrddhi instead of 
guna in strong forms : thus, mamdrja ; and yguh (also as in present : 745 c) 
has u instead of o. 

794. As regards the weakening in weak forms : 
a. It has been seen above (783 b) that roots beginning 
with i or u fuse reduplicating and radical syllable together to i 
or u in the weak forms; and (784) that roots contracting va 
and ya to u or i in the reduplication do it also in the root in 
weak forms, the two elements here also coalescing to u or i. 

b. A few roots having ya and va after a first initial consonant, and 
reduplicating from the semivowel (785), contract the ya and va to i and u: 
thus, vivic from yvyac, vividh from yvyadh, susup from ysvap. The ex- 
tended roots jya, pyd, vyd, $ya, hva show a similar apparent contraction, 
making their weak forms from the simpler roots ji, pi, m, CM, 7m, while hva 
must and fvd may get their strong forms also from the same (and it is 
questionable whether from the others strong forms occur). 

C. The root grabh or grah (if it be written thus) contracts to grh, making 
the three forms of stem jagrdh (1st and 2d sing, act.), jagrah (3d), and 
jagrh; but prach (if it be so written) remains unchanged throughout. 

d. A number of roots having medial a between single con- 
sonants drop that vowel. These are, in the later language, gam, 
Jthan, Jan, han, ghas: they form the weak stems jctgm, jakhn, 

jajn, jaghn (compare 637), jaks (compare 640) : but RV. has 
once jajanus. 

In the old language are found in like manner mamndthe and mamnate 
from }/man,- vavne from yvan; tatne, tatnise, tatnire from ytan (beside 
tatane, and tate, as if from yta); paptima and paptus and paptivdns from 
ypat (beside pet-forms; below, e); saccima and safcus, sacce and safdre 
from ysac. 

e. Roots in general having medial a before a single final 
consonant, and beginning also with a single consonant that is 
repeated unchanged in the reduplication that is, not an as- 
pirate, a guttural mute, or h contract their root and redupli- 
cation together into one syllable, having e as its vowel : thus, 
ysad forms the weak stem sed, ypac forms pec, yyam forms 
yem; and so on. 

Certain roots not having the form here defined are declared by the 
grammarians to undergo the same contraction most of them optionally; 




and examples of them all are of rare occurrence (of one only, bhaj, quotable 
from the older language). They are as follows: raj (occurs in MBh.) and 
radh (radh?), notwithstanding their long vowel; phan, phal, bhaj (occurs in 
RV. etc.), though their initial is changed in reduplication; trap, tras (occurs 
in MBh.), syam, svan, though they begin with more than one consonant; 
dambh (forming debh from the weaker dabh), though it ends with more than 
one; and bhram (occurs in KSS.), bhraj, granth, ^ranth, svanj, in spite of 
more reasons than one to the contrary. 

This contraction is allowed also in 2d sing. act. when the ending is 
ithd: thus, tenitha beside tatantha (but no examples are quotable from the 
older language). 

The roots pap and dad (from da: 672) are said to reject the con- 
traction : but no perfect forms of either appear to have been met with in use. 

From ytr (or tar) occurs terus (R.); and jerus from yjr is authorized 
by the grammarians both against the general analogy of roots in r. 

f. Roots ending in a lose their a before all endings be- 
ginning with a vowel, including those that assume the union- 
vowel i (796) unless in the latter case it be preferable to 
regard the i as a weakened form of the a. 

795. Endings, and their union with the stem. 
The general scheme of endings of the perfect indicative has 
been already given (553); and it has also been pointed out 
(543) that roots ending in TT a have ^ au in 1st and 3d 
sing, active. 

( 796. Those of the endings which begin with a con- 
sonant - - namely % tha, ^ va, JJ ma in active; H se, 3sc 
vahe, R% make, 9 dhve, \ re in middle are very often, 
and in the later language usually, joined to the base with 
the help of an interposed union-vowel ^ i. 

The union-vowel i is found widely used also in other parts of the general 
verbal system: namely, in the sibilant aorist, the futures, and the verbal 
nouns and adjectives (as also in other classes of derivative stems). In the 
later language, a certain degree of correspondence is seen among the different 
parts of the same verb, as regards their use or non-use of the connective; 
but this correspondence is not so close that general rules respecting it can 
be given with advantage ; and it will be best to treat each formation by itself. 
The perfect is the tense in which the use of t has estab- 
lished itself most widely and firmly in the later language. 

797. The most important rules as to the use of ^ * in 
the later language are as follows: 

262 X. PERFECT-SYSTEM. [797 

a. The"^ re of 3d pi. mid. has it always. 

b. The other consonant-endings, except 5J iha of 2d 
sing, act., take it in nearly all verbs. 

But it is rejected throughout by eight verbs namely Ar 
'make', Ihr 'bear', sr 'go', vr 'choose', dru 'run', cru 'hear', 
stu 'praise', sru 'flow'; and it is allowably (not usually) rejected 
by some others, in general accordance with their usage in other 

c. In 2d sing, act., it is rejected not only by the eight 
verbs just given, but also by many others, ending in vow- 
els or in consonants, which in other formations have no 
^ i; but it is also taken by many verbs which reject it in 
other formations ; and it is optional in many verbs, in- 
cluding those in %R a (of which the 5TT is lost when the 
ending is ^2f itha], and most of those in ^ ', 3 , and 3 u. 

The rules of the grammarians, especially as regards the use of iha or 
itha, run out into infinite detail, and "2-tre not wholly consistent with one 
another; and, as the forms are by no means frequent, it is not possible at 
present to criticise the statements made, and to tell how far they are founded 
on the facts of usage. 

With this i, a final radical i or i is not combined, but 
changed into y or iy. The u of ~\/bhu becomes uv throughout 
before a vowel. 

798. In the older language, the usage is in part quite 
otherwise . Thus : 

a. In the RV., the union-vowel i is taken by roots ending in con- 
sonants provided the last syllable of the stem is a heavy one, but not other- 
wise: thus, dsitha, uvocitha, viveditha, but tatdntha and vivydktha; ucimd, 
paptima, sedima, yuyopimdj but jaganma and yuyujma; ucise, jajnise, 
sasahise, but vivitse and dadrkse ; bubhujmdhe and $a$admdhe etc. (no ex- 
amples of ivahe or imahe chance to occur, nor any of either idhve or dhve}\ 
ijire, jajnire, yetire, tataksire, but caklpre, vividre, duduhre, pasprdhre, 
tatasre (and so on : twenty-two forms). The only exception in RV. is vettha 
from yvid, without i (in Br., also attha from yah: below, 801 a). The 
other Vedic texts present nothing inconsistent with this rule, but in the 
Brahmanas 3d pi. forms in ire are made after light syllables also : thus, 
sasrjire, bubudhire. 

b. In roots ending with a vowel, the early usage is more nearly like 
the later. Thus: for roots in a the rule is the same (except that no 2d sing, 
in itha is met with), as dadhimd, dadhise, dadhidhve, dadhire (the only per- 
sons with i quotable from RV. and AV.); roots in r appear also to follow 


the later rule : as cakrse, paprse, vavrse, vavrmdhe, but dadhrise and jabhrise, 
and in 3d pi. mid. both cakrire and dadhrire; ybhu has both babhutha 
(usually) and babhuvitha, but only babhuvimd (AV.). But there are found 
against the later rules, susuma, cicyuse, juhure, and juhure, without i: the 
instances are too few to found a rule upon. 

799. The ending rire of 3d pi. mid. is found in RV. in six forms : 
namely, cikitrire, jagrbhrire, dadrire, bubhujrire, vividrire, sasrjrire ; to which 
SV. adds duduhrire. 

800. Examples of inflection. By way of illustra- 
tion of the rules given above, may be given in full the per- 
fect indicative inflection of the following verbs : 

a. As example of the normal inflection of a root with 
final consonant, we take the root ^TJ budh, 'know': its strong 
form of perfect-stem is ^t^T bubodh; weak form, SRTJ bu- 

active. middle. 

d. . s. d. 

bubodha bubudhiva bubudhimd bubudhe bubudhivdhe btibudhim&he 

2 ^^ ^ ^^ ^_ ^^ 

bubodhitha bubudhdthus bubtidhd bubudhise bubudhathe bubudhidhve 

y , . 

bubodha bubudhdtus bubudhiis bubudhe bubudhate bubudhire 

The asserted variety of possible accent in 2d sing. act. (above, 793d) 
needs to be noted both in this and in the remaining paradigms. 

b. As example of the normal inflection of a root with 
final i or w-vowel, we may take the root ^ft m, 'lead': its 
forms of stem are HTO nindy or pRTTT ninay, and ft^ ninl. 

nindy a, ninhya ninyivd ninyimd ninye ninyivdhe ninyimdhe 

2 i^r, PnRj^ in-dy^ SRI PiP^5 PFOT ftfefij 

nine f ha, nindyitha ninydthus ninyd ninyise ninyhthe ninyidhve 

3 PRUT PKIHU PRR f^F& Pi^icJ Mp^l( 

t o " v ." s ., L ' 

ninaya ninydtus ninyus ninye ninyate ninyire 

The root kri would make in weak forms cikriyivd, cikriydtus, 

cikriyus, etc.; ybhu makes babhnva, babhutha (V.) or babhuvitha, 

264 X. PERFECT-SYSTEM. [800 

babhuvivd, babkuvtts; bab/mve, babhuvire, etc.; other roots in u or 
u change this to uv before the initial vowel of an ending. 

c. As example of the inflection of a root ending in 5TT 
, we may take ^T da, 'give' : its forms of stem are ^T dada 
and ([J dad (or dadi): see above, 794 f. 

dadau dadivd dadima, dade dadivdhe dadimdhe 

3$m,ffi ^m^ ^ ^f ^ ^ 

dadatha, dadithd daddthus dadd dadise dadhthe dadidhve 

dadau daddtus dadus dade dadate dadire 

The RV. has once paprd for paprcrf* (and Bafts' for jahati?). 

d. As example of a root with medial ^ a showing fu- 
sion of root and reduplication, resulting in medial ^ e, in 
the weak forms (794 e), we may take rR tan, ; stretch': its 
forms of stem are nTrH tatdn or H'HH tatan, and rR ^ew. 

, rTrTR 

tatdna, tatnna tenivd tenimd tene tenivdhe tenimdhe 

tatdntha, tenithd tendthus tend tenise tenathe tenidhve 

3 rTrTH HHHH^ ^5^ ^ ^^ ^T 

tatana tendtus tenus tene tenate tenire 

The root Jan, with the others which expel medial a in weak 
forms (794 d), makes jajdntha or jajnithd, jajnivd, jajnus ; jajne, 
jajnimdhe, jajnire ; and so on. 

e. As example of a root with initial 3f va contracted 
to 3 u in the reduplication, and contracted with the redu- 
plication to 3T u in weak forms (784), we may take 3frl vac, 
'speak': its forms of stem are 3^ uvdc or 331^ uvac, and 


uvdca, uvaca ucivd ucimd uce ucivdhe ucimdhe 


uvdktha, uvdcitha ucdthus ucd ucise ucathe ucidhve 


uvaca ucdtus ucus uce ucate 


In like manner, ^yaj forms iydja or iyaja, iydstha or iydjitha ; 
ye, yise, and so on ; j/wc has nvoca and uvocitha in the strong 
forms, and all the rest like vac. 

f. Of the four roots in ft r mentioned at 797 b, the first 
persons are made as follows: 

cakdra, cakara cakrvd cakrmd cakre cakrvdhe cakrmdhe 
The 2d sing. act. is cakdrtha; the 3d pi. mid. is cakrire, 
Of the roots in ft r in general, the first persons are 
made as follows : 

dadhdra,dadharadadhrivd dadhrimd dadhre dadhrivdhe dadhrimdhe 

801. A few miscellaneous irregularities call still for 
notice : 

a. The root ah, 'speak', occurs only in the perfect indic- 
ative, and only in the 3d persons of all numbers and in the 
2d sing, and du., in active (and in 2d sing, the h is irregularly 
changed to t before the ending): thus, attha, nha; ahathus, 
ahatus; ahus (in V., only nha and ahus are met with). 

b. From yva 'weave', the 3d pi. act. uvus occurs in RV., and no other 
perfect form appears to have been met with in use. It is allowed by the 
grammarians to be inflected regularly as va; and also as vay (the present- 
stem is vdya : 761 d, 2), with contraction of va to u in weak forms ; and 
further, in the weak forms, as simple w. 

c. The root vya, 'hide', has in RV. the perfect-forms vivyathus and 
vivye, and no others appear to have been met with in use; the grammarians 
require the strong forms to be made from vyay, and the weak from vi. 

d. The root i, 'go', forms in RV. and AV. the 2d sing. act. iydtha 
beside the regular iyetha. 

e. The AV. has once vavrdhete (for -dhdte), and once jaharus (for 
jdhrus}: both are perhaps false readings. 

f. Persons of the perfect from the er-forms of roots in changeable f 
(242) are titirus and tistire (bothRV.); and they have corresponding participles. 

g. The bastard root urnu (712) is said by the grammarians to make 
the perfect-stem urnunu. 

h. The roots majj and nap are said to insert a nasal in the 2d sing. 

266 X. PERFECT-SYSTEM. [801 

active, when the ending is simple tha: thus, mamanktha, nanahstha (also 
mamajjitha and ne?itha). 

i. The anomalous ajagrabhaisarh (AB. vi. 35) seems a formation on the 
perfect-stem (but perhaps for ajigrabhisan, desid. ?). 

Perfect Participle. 

802 . The ending of the active participle is SfftT vans (that 
is to say, in the strong forms : it is contracted to 3^ us in the 
weakest, and replaced by cffiT vat in the middle forms : see 
above, 458 if.). It is added to the weak form of the perfect 
stem as shown, for example, in the dual and plural of 
the active inflection of the given verb; and, mechanically, 
the weakest participle-stem is identical with the 3d pi. 
active. Thus, sHMlU bubudhvans, PHl^l^ nimvans, rloNfH 

\ 803. If the weak form of the perfect stem is monosyl- 
labic, the ending takes the union-vowel S i (which, however, 
disappears in the weakest cases): thus, (Tf^lH tenivans, 
ucivans, sii^JH jajnivans, ^ii^cjtH adivans (from 
ad: 783 a), and so on; ^f^TCT dadivans and its like, 
from roots in 5TT a, is to be reckoned in the one class or 
the other according as we view its ^ i as weakened root- 
vowel or as union-vowel (794f). 

But participles of which the perfect-stem is monosyllabic by absence of 
the reduplication do not take the union-vowel: thus, vidvdns , and, in V., 
dafvahs, midhvans, sahvans ; and RV. has also dadvans (AV. dadivans and 
once dadavdhs) from ydd (or dad : 672) and AV. has vif ivdns and 
varjivahs (in negative fern, dvarjusi). 

804. Other Vedic irregularities calling for notice are few. The long 
vowel of the reduplication (786) appears in the participle as in the indicative : 
thus, vavrdhvdhs, sasahvdhs, jujuvdhs. RV. and AV. have sasavdns from 
ysan or sa. RV. makes the participial forms of ytr or tar from different 
modifications of the root: thus, titirvdhs, but tatarusas. Respecting the 
occasional exchanges of strong and weak stem in inflection, see above, 462 c. 

805. From roots gam and han the Veda makes the strong stems 
jaganvdhs (as to the n, see 212) and jaghanvdhs; the later language allows 
either these or the more regular jagmivdhs and jaghnivdhs (the weakest 
stem-forms being everywhere jagmus and jaghntis). 


806. From three roots, vid 'find', uip, and dry, the later language allows 
strong participle-stems to be made with the union-vowel, as well as in the 
regular manner without it: thus, vivifivdna or vivifvdns. PB. has once 

807. The ending of the middle participle is and. It 
is added to the weak form of perfect-stem, as this appears 
in the middle inflection : thus, 5RTJH bubudhand, 
ninyana, ^R dadand, ^F? tenand, sT^H jajnand, 

In the Veda, the long reduplicating vowel is shown by many middle 
participles : thus, vavrdhand, vdvasdnd, dddrhdnd, fucuvand, etc. RV. has 
?afayand from y$l (with irregular guna, as in the present-system: 629); 
tistirand from ]/sr,- and once, with mana, sasrmdnd from j/sr. 

Modes of the Perfect. 

808. Modes of the perfect belong only to the Vedic lan- 
guage, and are even rarely found outside of the Rig- Veda. 

To draw the line surely and distinctly between these and the mode-forms 
from other reduplicated tense-stems the present-stem of class II., the 
reduplicated aorist, and the intensive is not possible, since no criterion 
of form exists which does not in some cases fail, and since the general 
equivalence of modal forms from all stems (582), and the common use of 
the perfect as a present in the Veda (823), deprive us of a criterion of 
meaning. There can be no reasonable doubt, however, that a considerable 
body of forms are to be reckoned here : optatives like anacydm and babhuyds 
and babhuydt, imperatives like babhutu, subjunctives like jabhdrat, show such 
distinctive characteristics of the perfect formation that by their analogy other 
similar words are confidently classed as belonging to the perfect. 

809. The normal method of making such forms would 
appear to be as follows : from a reduplicated perfect-stem, as 
(for example) mumuc, an imperative would be made by simply 
appending, as usual, the imperative endings ; the derived sub- 
junctive mode-stem would be mumoca (accented after the analogy 
of the strong forms of the perfect indicative), and would take 
either primary or secondary endings ; and the optative mode- 
stems would be vnumucyh in the active, and mumuci (accent on 
personal endings) in the middle. 

And the great majority of the forms in question (about 
three quarters) are made in these ways. Thus : 

810. Examples of the regular subjunctive formation are : 
a. with secondary endings, active: 2d sing., paprdthas, mamdhas, 

piprdyas, bubodhas; 3d sing., jabhdrat, sasdhat, paspdr$at, piprdyat, ciketat; 

268 X. PERFECT-SYSTEM. [810 

1st pi., tatdndma, (u^dvdma; 3d pi., tatanan, paprdthan (other persons do 
not occur). This is the largest class of cases. 

b. with primary endings, active: here seem to belong only dadhdrshati 
and vavdrtati: compare the formation with different accent below, 81 la. 

c. of middle forms occur only the 3d sing, tatdpate, papamafe, yuyojate, 
jujosate (SV.; RV. has jujosate)-, and the 3d pi. tatdnanta (and perhaps two 
or three others: below, 81 1 b, end). 

811. But not a few subjunctives of other formation occur; thus: 

a. with strengthened root-syllable, as above, but with * accent on the 
reduplication (as in the majority of present-forms of the reduplicating class : 
above, 645). Here the forms with primary endings, active, preponderate, 
and are not very rare: for example, jtijosasi, jtijosati, jujosathas, jtijosatha 
(other persons do not occur). With secondary endings, jtijosas, jtijosat, and 
jujosan are the forms that belong most distinctly here (since ddda$as and 
stisudas etc. are perhaps rather aorists). And there is no middle form but 
jtijosate (RV.: see above, 81 Oc). 

b. with unstrengthened root-syllable occur a small body of forms, which 
are apparently also accented on the reduplication (accented examples are found 
only in 3d pi. mid.): thus, active, for example, mumucas ; vavrtat, vividat, 
ptipuuai; the only middle forms are dadhrsate, vdvrdhate, 3d sing.; and 
cdkramanta, dddhrsanta, rurucanta (with dadabhanta, paprathanta, mdmahanta, 
juhuranta, which might also belong elsewhere: 81 Oc). 

c. accented on the ending are vdvrdhdnta and cakrpdnta (which are 
rather to be called augmentless pluperfects). 

As to forms with double mode-sign, or transfers to an a-conjugation, 
see below, 815. 

812. Examples of the regular optative formation are: 

a. in active: 1st sing., dna$ydm, jagamydm, paprcydm, riricydm; 2d 
sing., vavrtydSj vivifyds, fUfruyas, babhuyds; 3d sing., jagamydt, vavrtydt, 
tutujydt, babhuydt; 2d du., jagamydtam, pucruydtam; 1st pi., sasahyama, 
vavrtyama, fu^uydma; 3d pi., tatanyus, vavrtyus. The forms are quite 

b. in middle, the forms are few: namely, 1st sing., vavrtlya; 2d sing., 
vawrdhithas, caksamlthds ; 3d sing., jagrasita, vavrtita, mdmrjlta, pufuclta; 
1st pi., vawrtimahi. And sdsahlsthds and ririsista appear to furnish examples 
of precative optative forms. 

There is no irregular mode of formation of perfect optatives. Individual 
irregularities are shown by certain forms : thus, cakriyds, paplydt, fu$ruyas 
and wruydtam, with treatment of the final as before the passive-sign yd 
(770); anajydt with short initial; jak&ydt is anomalous; ririses is the only 
form that shows a union-vowel a. 

813. Of regular imperative forms, only a very small number are to be 
quoted : namely, active, cikiddhi, mumugdM, pucugdhf, and piprihi; mumoktu 
and babhutu; mumuktam and vavrktam; jujustana and vavrttana (unless we 


are to add mamaddhf, mamattu, mamdttana); middle, vavrtsva and 

814. As irregular imperatives may be reckoned several which show a 
union- vowel a, or have been transferred to an a-conjugation. Such are, in 
the active, mum6catam and jujosatam (2d du.), and mumticata (2d pi.}; in 
the middle, piprdyasva (only one found with accent), and mdmahasva, vdvrdh- 
asva, vdvrsasva (2d sing.), and mdmahantdm (3d pi.: probably to be ac- 
cented -dsva and -dntam). 

815. Such imperatives as these, taken in connection with some of the 
subjunctives given above (and a few of the "pluperfect" forms: below, 820), 
suggest as plausible the assumption of a double present-stem, with redupli- 
cation arid added a (with which the desiderative stems would be comparable : 
below, chap. XIV.): for example, jujosa from YJUS, from which would come 
jujosasi etc. and jujosate (81 la) as indicative, jujosas etc. as subjunctively 
used augmentless imperfect, and jujosatam as imperative. Most of the forms 
given above as subjunctives with primary ending lack a marked and constant 
subjunctive character, and would pass fairly well as indicatives. And it 
appears tolerably certain that from one root at least, vrdh, such a double 
stem is to be recognized ; from vdvrdha come readily vdvrdhate, vdvrdhdnta , 
and from it alone can come regularly vdvrdhasva, vdvrdhete (above, 801 e), 
and vdvrdhdti (once, RV.) and, yet more, the participle vavrdhdnt (once, 
RV.: an isolated case): yet even here we have also vdvrdhltha's, not vdvrdh- 
ethds. To assume double present-stems, however, in all the cases would be 
highly implausible ; it is better to recognize the formation as one begun, 
but not carried out. 

Only one other subjunctive with double mode-sign namely, paprcdsi 
is found to set beside vdvrdhdti. 

816. Forms of different model are not very seldom made from the same 
root: for example, from j/mwc, the subjunctives mumticas, mumocati, and 
mumucas ; from ydhrs, dadhdrsati and dadhrsate , from j/pn, the imperatives 
piprlM and piprdyasva. 


817. Of an augment-preterit from the perfect-stem, to which 
the name of pluperfect is given on the ground of its formation 
(though not of its meaning), the Veda presents a few examples ; 
and one or two forms of the later language (mentioned above, 
788, end) have also been referred to it. 

There is something of the same difficulty in distinguishing the pluperfect 
as the perfect-modes from kindred reduplicated formations. Between it and 
the aorist, however, a difference of meaning helps to make a separation. 

818. The normal pluperfect should show a strong stem in the singular 
active, arid a weak one elsewhere thus, mumoc and mumuc with 
augment prefixed and secondary endings added (us in 3d pi. act., ata in 
3d pi. mid.). 

270 X. PERFECT-SYSTEM. [818 

Of forms made according to this model, we have, in the active: 1st sing., 
ajagrabham and acacaksam (which, by its form, might be aorist: 860); 
2d sing, djayan; 3d sing., ajagan and aciket; 2d du., amumuktam; 2d pi. 
ajaganta, and ajagantana and ajabhartana (a strong form, as so often in this 
person: 556); 3d pi. (perhaps), amamandus and amamadus . To these may 
be added the augmentless ciketam and cakaram. In the middle, the 3d pi. 
acakriran and ajagmiran (with Iran instead of ata), and the augmentless 
2d sing, jugurthas and susupthds, are the most regular forms to be found. 

819. Several forms from roots ending in consonants save the endings 
in 2d and 3d sing. act. by inserting an I (555 b): thus, dbubhojis, avive&s; 
arireclt, djagrabhlt (avavarit and avava^ltdm are rather intensives); and the 
augmentless jfhinszs (accent?) and dadharslt belong with them. 

820. A few forms show a stem ending in a: they are, in the active: 
3d sing., asasvajat, acakrat; in the middle: 3d sing., dpiprata; 2d du., 
dpasprdhetham ; 3d pi., atitvisanta (which by its form might be aorist), 
ddadrhanta; and cakradat, cakrpdnta, vdvrdhdnta, juhuranta, would perhaps 
be best classified here as augmentless forms (compare 811, above). 

Uses of the Perfect. 

821. In the later language, the perfect is simply a preterit 
or past tense, equivalent to the imperfect, and interchangeable 
with it. Except as coming from a few often used verbs (espe- 
cially aha and uvaca), it is much more rarely employed than the 

822. In the Brahmana language, very nearly the same thing is true. 
In most Brahmanas, the imperfect is the usual tense of narration, and the 
perfect only occasional; in the Qatapatha Brahmana, the perfect is much 
more widely used. 

823. In the Veda, the case is very different. The perfect is used as 
past tense in narration, but only rarely ; sometimes also it has a true "perfect 1 ' 
sense, or signifies a completed past; but oftenest it has a value not distin- 
guishable in point of time from the present. It is thus the equivalent of 
imperfect, aorist, and present; and it occurs coordinated with them all. 

Examples are : of perfect with present, nd frdmyanti nd vf muncanty 
ete vdyo nd paptuh (RV.), 'they weary not nor stop, they fly like birds'; se 
'd u rdjd ksayati carsanindm ardn nd nemth pdri td babhuva (RV.), 'he in 
truth rules king of men ; he embraces them all, as the wheel the spokes'; 
of perfect with aorist, tipo ruruce yuvatfr nd y6sd. . . dbhud agnih samtdhe 
mdnusanam dkar jyotir bddhamdnd tdmdhsi (RV.), 'she is come beaming like 
a young maiden ; Agni has appeared for the kindling of mortals ; she hath 
made light, driving away the darkness'; of perfect with imperfect, dhann 
dhim dnv apds tatarda (RV.), 'he slew the dragon, and penetrated to the 
waters'. This last combination is of constant occurrence in the later language. 

824] 271 



<^ 824. UNDER the name of aorist are included (as was 
pointed out above, 532) three quite distinct formations, 
each of which has its sub-varieties : namely, 

I. A SIMPLE-AORIST (equivalent to the Greek "second 
aorist"), analogous in all respects as to form and inflection 
with the imperfect. It has two varieties: 1. the root- 
aorist, with a tense-stem identical with the root (corres- 
ponding to an imperfect of the root-class, I.); 2. the 
a-aorist, with a tense-stem ending in f #, or with union- 
vowel 5f a before the endings (corresponding to an imper- 
fect of the a-class, VII.). 

II. 3. A REDUPLICATED AORIST, perhaps in origin ident- 
ical with an imperfect of the reduplicating class (II.), but 
having come to be separated from it by marked peculiarities 
of form. It usually has a union-vowel 51 a before the end- 
ings, or is inflected like an imperfect of one of the ^-classes ; 
but a few forms occur in the Veda without such vowel. 

III. A SIBILANT-AORIST (corresponding to the Greek "first 
aorist"), having for its tense-sign a H s added to the root, 
either directly or with a preceding auxiliary ^ */ its endings 
are usually added immediately to the tense-sign, but in a 
small number of roots with a union-vowel 5[ a ; a very few 
roots also are increased by H s for its formation ; and accord- 
ing to these differences it falls into four varieties : namely, 
A. without union-vowel ^ a before endings: 4. s-aorist, 
with H s alone added to the root; 5. as -aorist, the same 
with interposed ^ i; 6. s^s-aorist, the same as the pre- 
ceding with H s added at the end of the root; B. with 
union-vowel 5!" a, 7. s a- aorist. 

272 XI. AORIST-SYSTEMS. j 825 

825, All these varieties are bound together and made 
into a single complex system by certain correspondences of 
form and meaning. Thus, in regard to form, they are all 
alike, in the indicative, augment-preterits to which there 
does not exist any corresponding present; in regard to 
meaning, although in the later or classical language they 
are simply preterits, exchangeable with imperfects and per- 
fects, they all alike have in the older language the general 
value of a completed past or "perfect", translatable by 'have 
done' and the like. 

826. The aorist-system is a formation of very infrequent occurrence in 
the classical Sanskrit (its forms are found, for example, only twenty-one 
times in the Nala, eight in the Hitopadeca, seven in Manu, six each in the 
Bhagavad-Gita and Qakuntala), and it possesses no participle, nor any modes 
(excepting in the prohibitive use of its augmentless forms: see 580; and 
the so-called precative : see 921 ff.); in the older language, on the other 
hand, it is quite common, and has the whole variety of modes belonging to 
the present, and sometimes participles. Its description, accordingly, must be 
given mainly as that of a part of the older language, with due notice of its 
restriction in later use. 

827. In the RV., nearly half the roots occurring show aorist forms, 
of one or another class; in the AV., rather less than one third; and in the 
other texts of the older language comparatively few aorists occur which are 
not found in these two. 

More than fifty roots, in RV. and AV. together, make aorist forms of 
more than one class (not taking into account the reduplicated or "causative" 
aorist); but no law appears to underlie this variety; of any relation such as 
is taught by the grammarians, between active of one class and middle of 
another as correlative, there is no trace discoverable. 

Examples are : of classes 1 and 4, adham and dhasus from ydha, ayuji 
and ayuksata from yyuj ; of 1 and 5, agrabham and agrdbhxsma from 
ygrabh, mrsthas and marsisthas from ymrs ; of 1 and 2, arta and arat 
from yr ; of 2 and 4, avidam and avitsi from yvid 'find', anijam and 
anaikslt from j/ny; of 2 and 5, sanema and asanisam from ysan; of 
2 and 7, aruham and aruksat from yruh; of 4 and 5, amatsus and 
amadisus from ymad; of 4 and 6, hasmahi and hdsisus from yha; 
of 1 and 2 and 4, atnata and atanat and atan from ytan; of 1 and 4 
and 5, abudhran and abhutsi and bodhisat from ybudh, astar and strslya and 
astaris from ystr. Often the second, or second and third, class is repre- 
sented by only an isolated form or two. 


I. Simple Aorist. 

828. This is, of the three principal divisions of aorist, 
the one least removed from the analogy of forms already ex- 
plained ; it is like an imperfect, of the root-class or of the d- 
class, without a corresponding present indicative, but with (more 
or less fragmentarily) all the other parts which go to make up 
a complete present-system. 

1. Root-aorist. 

829. This formation is in the later language limited to 
a few roots in TT a and the root H bhu, and is allowed to 
be made in the active only, the middle using instead the 
s-aorist (4), or the is -aorist (5). 

The roots in ETF a take 3^T us as 3d pi. ending, and, as 
usual, lose their TT a before it ; *F bhu (as in the perfect : 
793 a) retains its vowel unchanged throughout, inserting ^ 
v after it before the endings FT am and R an of 1st sing, 
and 3d pi. Thus: 

ddam ddava ddama dbhuvam dbhuva dbhuma 


ddas ddatam ddata dbhus dbhutam dbhuta 


ddat ddatam ddus dbhut dbhutam dbhuvan 

For the classical Sanskrit, this is the whole story. 

830. In the Veda,, these same roots are decidedly the most 
frequent and conspicuous representatives of the formation : es- 
pecially the roots ffd, da, dha, pa 'drink', stha, bhu; while spo- 
radic forms are made from Jnd, pra, sa, ha. As to their middle 
forms, see below, 834 a. 

Instead of abhuvam, RV. has twice dbhuvam. 

831. But aorists of the same class are also made from a 
number of roots in r , and a few in i and u (short or long) 
with, as required by the analogy of the tense with an imperfect 
of the root-class, ywwa-strengthening in the three persons of the 

Thus (in the active), from j/?rw, afravam and afrot; from }/?n, acres 
and of ret; from ykr 'make', akaram and akar (for akars and afcar-<); from 
Whitney , Grammar. 18 


vr 'enclose', dvar (585); and so adar, aaiar, aspar. Dual and plural forms 
are much less frequent than singular; but for the most part they also show 
an irregular strengthening of the root-vowel: thus (including augmentless 
formsjj akarma and akarta, vartam, spartam, ahema and ahetana, bhema, 
ftoma, afravan; regular are only avran, afcrara, ahyan, afriyan, and anltdm. 

832. Further, from a few roots with medial (or initial) 
vowel capable of ywraz-strengthening, and having in general that 
strengthening only in the singular. 

Thus, abhet (2d and 3d sing.) from ybhid; amok (3d sing.) from j/mwc , 
avart from yvrt ; vark from yvrj ( AV. has once avrk) ; adarfam from }/drp , 
and adrfan, avrjan, afvitan. But chedma, with guna, from ychid. 

833. Again, from a larger number of roots with a as ra- 
dical vowel : 

Of these, gam (with n for m when final or followed by m: 143, 212) 
is of decidedly most frequent occurrence, and shows the greatest variety of 
forms : thus, agamam, agan (2d and 3d sing.), aganma, aganta (strong form), 
agman. The other cases are akran from ykram; atan from ytan; askan 
from yskand ; dsrat from y trans (? VS. ) ; dAafc und daghma from ydagh , 
ana (585) and anastam from }/nap (?) ; afcsan (for agh-san, like apman) from 
yghas; and the 3d pll. in w, dfcramws, kramus, ayamus, yamus, abddhus, 
dabhus, nrtus (impf . ?) : mandus and taksus are perhaps rather to be reckoned 
as perfect forms without reduplication (790 b). 

834. So far only active forms have been considered. In 
the middle, a considerable part of the forms are such as are 
held by the grammarians (881) to belong to the s-aorist, with 
omission of the s: they doubtless belong, however, mostly or 
altogether, here. Thus : 

a. From roots ending in vowels, we have adhithas and adhita; adita 
and adimahi (and adlmahi from ydd 'cut'); aathithas and asthita and dsthiran, 
forms of a-roots (arddhvam is doubtless for arasdhvam); of r-roots, afcri, 
akrthds, akrta, akratam, akrata; avri, avrthas, avrta; drta (with augmentless 
arta), drata; mrthds, amrta; dhrthds; drihds; astrta; asprta; gurta; of i 
and u roots, the only examples are dhvi (?AV., once) and acidhvam. The 
absence of any analogies whatever for the omission of a s in such forms, and 
the occurrence of avri and akri and akrata, show that their reference to the 
-s-aorist is without sufficient reason. 

b. As regards roots ending in consonants, the case is more questionable, 
since loss of s after a final consonant before thds and ta (and, of course, 
dhvam) would be in many oases required by euphonic rule (233). We find, 
however, such unmistakable middle inflection of the root-aorist as ayuji, 
ayukthds, ayukta, ayujmahi, ayugdhvam, dyujran; dsta and dfata; apadi 
(1st sing.) and apadmahi and apadran; amanmahi; aganmahi and agmata; 
atnata; ajani (1st sing.) and ajnata (3d pi.); from ygam are made agathds 
and agata, and from >/wan, amafa, with treatment of the final like that of 
han in present inflection (637). The ending ran is especially frequent in 


3d pi., being taken by a number of verbs which have no other person of 
this aorist: thus, agrbhran, dsrgran, adr^ran, abudhran, dvrtran, ajuaran, 
akrpran, asprdhran, avasran, dvifran; and ram is found beside ran in 
ddr$ram, dbudhram, dsrgram. 

c. From roots of which the final would combine with s to fes, it seems 
more probable that aorist-forms showing 7c (instead of a) before the ending 
belong to the root-aorist: such are amukthas (and amugdhvam], aprkthds and 
aprkta, abhakta, avrkta, asakthas and asakta, rikthas, vikthas and vikta : 
dspasta, asrsta, mrsthas would be the same in either case. 

d. There remain, as cases of doubtful belonging: amatta, arabdha, 
asrpta, atapthas, chitthas, patthas, and nutthas. 

Modes of the Root-aorist. 

835. In subjunctive use, forms identical with the augmentless indic- 
ative of this aorist are much more frequent than the more proper subjunc- 
tives. Those to which no corresponding form with augment occurs have 
been given above ; the others it is unnecessary to report in detail. 

836. Of true subjunctives, the forms with primary endings are quite 
few. In the active, gani is the only example of 1st sing, (as to bhuvani, 
see below); of 3d sing, occur sthati, and ddti and dhdti, which are almost 
indicative in use ; of dual persons, sthdthas, darfathas, fravathas and prd- 
vatas. In the middle, 3d sing, idhate (? anomalous accent), 2d du. dhethe 
and dhaithe, and 1st pi. dhamahe. 

Forms with secondary endings are, in the active, ddrfam; tdrdas, pdrcas, 
ydmas ; karat, garat, daghat, yamat^ yodhat, frdvat, spdrat, sdghat ; ddrpan, 
gar an, yaman. No middle forms are classifiable with confidence here. 

The series bhuvam, bhuvas, bhuvat, bhuvan, and bhuvani (compare 
abhuvam : 830, end), and the isolated friivat, are of doubtful belongings ; 
with a different accent, they would seem to be of the next class; here, a 
^Una-strengthening would be more regular (but note the absence of guna in 
the aorist indicative and the perfect of ybhu). 

837. As regards the optative, we have to distinguish between pure 
optatives and optatives with s inserted before the endings, or precatives. 

Pure optative forms, both active and middle, are made from a number 
of roots. From roots in a occur in the active, with change of a to e before 
the y, deyam, dheyam and dheyus, stheyama, and jneyas (which might also 
be precative); in the middle, only simahi and dhimahi (which might be aug- 
mentless preterit, as adhimahi also occurs once, and adhltam once : but ydha 
shows nowhere else conversion of its a to long 1], From bhu, bhuyas and 
bhuyat (possible precatives), and bhuyama. From roots in r, the middle 
forms arlta, muriya, vurlta. From roots ending in consonants, afyam and 
acydma and a?yus act., and aflya and afimdhi mid. (this optative is espe- 
cially common in the older texts) ; vrjydm ; mrdhyas, sahyas, gamyas and 
gamyat (possible precatives), and sahyama; indhlya, gmiya, ruciya, vaslya; 
idhimahi, nafimahi, nasimahi, prclmahi, mudlmahi, yamimahi. And rdhyam 
etc. and rdhimdhi belong perhaps here, instead of to a present-system (cl. I.). 



838. Precative optative forms of this aorist active are in the later 
language allowed to be made from every verb (922). In RV., however, 
they do not occur from a single root which does not show also other aorist 
forms of the same class. They are, indeed, of very limited use: thus, in 
3d sing. act. (ending in -yds for -ydst), we find avyds, cifyas, daghyds, 
bhuyds, yamyds, yuyds, vrjyds, ?ruyds, sahyds (RV. has no 3d sing, in -yat}; 
and besides these and the 2d sing, in yas (given in the preceding paragraph), 
which might be of either formation, occur in the active only bhuydsam (beside 
bhuyama) and kriyasma, each once. In the middle, RV. furnishes the three 
forms grabhlsta, padistd, muclsta; nothing additional is found in any other 
text. From the AV. on, the active precative forms are more frequent than 
the pure optative (which are not wholly unknown, however); they are nowhere 
common, excepting as made from ybhu; and from no other root is anything 
like a complete series of persons quotable (only bhuyasva and bhuydstdm being 
wanting; and these persons having no representative from any root). In- 
cluding the cases already given, they have been noticed as made from about 
twenty roots, as follows: fruyasam etc., kriyasam etc., priydsam, bhriydsam, 
saghydsam, bhrdjydsam, udydsam etc. (j/vad), rdhydsam etc., radhydsam etc., 
badhydsam etc., trpydsma, jivydsam, pusydsam etc. AV. has once bhuydstha, 
with primary ending, but it is doubtless a false reading (TB. has -sta in 
the corresponding passage). TS. has dldydsam, from the quasi-root dldl (676). 

839. Imperative forms of the root-aorist are not rare in the early lang- 
uage. In the middle, indeed, almost only the 2d sing, occurs : it is accented 
either regularly, on the ending, as fcrsua, dhisvd, yuksvd, or on the root, as 
mdtsva, ydksva, vdnsva, rdsva, sdksva ; mdsva is not found with accent ; the 
2d pi. is represented by krdhvam, vodhvam. In the active, all the persons 
(2d and 3d) are found in use; examples are: 2d sing., krdhi, prudhf, gadhi, 
yandhf, gahi, mafti, mogdhi; 3d sing., dato, a*u, s6tu ; 2d du., ddtam, 
jitam, frwtam, bhutdm, sprtdm, gatam, riktdm; 3d du., only gantam, vodhSm; 
2d -p\.,'gatd, bhutd, prwta, fcrta, gata, dhatana; 3d pi., only dhdntu, fru- 
vantu. These are the most regular forms ; but irregularities both as to accent 
and strengthening are not infrequent. Thus, strong forms in 2d du. and pi. 
are varktam, vartam; kdrta, gdnta, yanta, vartta, heta, yr6ta, stita; and, 
with tana, kdrtana, aanfana, t/antana, sotana. Much more irregular are 
yodhi (instead of yuddhf) from yyudh, and bodM from both ybudh and ybhu 
(instead of buddhi and bhudht). A single form (3d sing.) in tat is found, 
namely fastat. 

As to 2d persons singular in si from the simple root used in an im- 
perative sense, see above, 624. 

Participles of the Root-aorist. 

840. In the oldest language, of the E-V., are found a 
number of participles which must be reckoned as belonging to 
this formation. 

In the active, they are extremely few : namely ferdnt, sthdnt, and bhiddnt, 
and probably rdhdnt. 


In the middle, they are much more numerous : examples are arana, 
idhand, krand, jusana, drparad, prcand, bhiyand, vrand, sprdhand, hiydnd. 
Such formations are extremely rare in the later Veda and in the Brahmana. 

841 . All together, the roots exhibiting in the older language 
forms which are with fair probability to be reckoned to the root- 
aorist-system are very nearly a hundred (more than ninety) ; about 
eighty of them make such forms in the RV. 

Passive Aorist third person singular. 

842. A middle third person singular, of peculiar formation 
and prevailingly passive meaning, is made from many verbs in 
the older language, and has become a recognized part of the 
passive conjugation, being, according to the grammarians, to be 
substituted always for the regular third person of any aorist 
middle that is used in a passive sense. 

843. This person is formed by adding ^ i to the root, 
which takes also the augment, and is usually strengthened. 

The ending i belongs elsewhere only to the first person ; and this third 
person apparently stands in the same relation to a first in i as do, in the 
middle voice, the regular 3d sing, perfect, and the frequent Vedic 3d sing, 
present, which are identical in form with their respective first persons. That 
a fuller ending has been lost off is extremely improbable; and hence, as an 
aorist formation from the simple root, this is most properly treated here, in 
connection with the ordinary root-aorist. 

844. Before the ending ^ i. a final vowel, and usually 
also a medial % a before a single consonant, have the vrddhi- 
strengthening; other medial vowels have the guna-strengtli- 

ening if capable of it (240); after final TT # is added T y> 

Examples (all of them quotable from the older language) are : from roots 
ending in a, djnayi, ddhayi, dpayi ; in other vowels, a'prai/i, dstavi, dhavi, 
a/carz, astern , from roots with medial i, M, r, acefz, dchedi, apesi, dbodhi, 
dmofi, dyoji, ddarfi, asarji, varhi; from roots with medial a strengthened, 
agami, dpadi, at/ami, avaci, vapi, dsddi (these are all the earlier cases): 
with a unchanged, only djani (and RV. has once jdni), and, in heavy syl- 
lables, dmyaksi, vandi, fansi; with medial a. dbhraji, dradhi; from roots 
with initial vowel, ardhi (only case). 

According to the grammarians, certain roots in am, and ybadh, retain 
the a unchanged: thus, adami, abadhi; and there are noted besides, from 
roots sometimes showing a nasal, arambhi, arandhi, ajambhi, abhanji or 
abhaji, alambhi (always, with prepositions) or alabhi: QB. has asanji. 

Augmentless forms, as in all other like cases, are met with, with either 
indicative or subjunctive value: examples (besides the two or three already 




given) are : dhdyi, frdvi, bhari, red, vedi, rod, jdni, pddi, sddi, ardhi. The 
accent, when present, is always on the root- syllable. 

845. These forms are made in RV. from forty roots, and all the other 
earlier texts combined add only about fifteen to the number; in the later 
language they are (like all the kinds of aorist) very rare. When they come 
from roots of neuter meaning, like gam, pad, sad, bhraj, sanj, etc., they 
have (like the so-called passive participle in ta: 952) a value equivalent to 
that of other middle forms ; in a case or two (RV. vii. 73. 3[?]; VS. xxviii. 15) 
they appear even to be used transitively. 

2. The a-aorist. 

846. This aorist is in the later language allowed to be 
made from a large number of roots (near a hundred). It is 
made in both voices, but is rare in the middle, most of the 
roots forming their middle according to the s-class (4) or 
the /5-class (5). 

Its closest analogy is with the imperfect of the a-class 
(VII.); its inflection is the same with that in all particulars; 
and it takes in general a weak form of root save the 
roots in ft r (three or four only), which have the guna- 

As example of inflection may be taken the root 
sic, 'pour': thus, 



dsicam dsicava dsicama dsice 

dsicavahi dsicamahi 

dsicas dsicatam dsicata dsicathas dsicetham dsicadhvam 

dsicat dsicatam dsican dsicata dsicetam dsicanta 

847. The a-aorist makes in the RV. a small figure beside the root- 
aorist, being represented by less than half the latter's number of roots. It 
becomes, however, more common later (it is the only form of aorist which is 
made from more verbs in AV. than in RV.); and in Veda and Brahmana together 
about seventy roots exhibit the formation more or less fully. Of these a large 
number (fully half) are of the type of the roots which make their present- 
system according to the d-class (VII.), having a vowel capable of pwia-strength- 
ening before a final consonant (753): thus, with i, chid, bhid, nij, pis, 

849] SIMPLE AORIST: 2. -AORIST. 279 

ris, vid, if is (fds], 2pi, fvit, sic, sridh; with M, krudh, ksudh, dm, dyut, 
druh, pus, budh, bhuj, muc, mus, yuj, rudh, muh, ruh, pwc,- with r, rdh, 
fer, grdh, trp, trs, trh, drp, dYp, mrs, vrj, vrt, vrdh, srp. A small number 
end in vowels: thus, r, fcr, gr, sr (which have the gruna-strengthening 
throughout), hi (? ahyat once in AV.), stu (? stuvatdm, 3d sing. impv. mid., 
once in AV.); and several in a, apparent transfers from the root-class by 
the weakening of their a to a: thus, Mi/a, hvd, vyd, fvd, and (in RV. only) 
da and dha. A few have a penultimate nasal in the present and elsewhere, 
which in this aorist is lost: thus, bhrah?, srahs, krand, manth, randh. Of 
less classifiable character are feraw, gam, tarn, fam, fram, tan, van, san, 
sad, radh, dabh, sap, ap, IT, das, fafc. The anomalous astham is the aorist 
of as 'throw'. The roots pat, nap, vac form the tense-stems papta, ncpa, 
voca, of which the first is palpably and the other two are probably the result 
of reduplication ; but the language has lost the sense of their being such, 
and makes other reduplicated aorists (3) from the same roots (see below, 854). 

848. The inflection of this aorist is in general so regular that it will 
be sufficient to give only examples of its Vedic forms. We may take as 
model avidam, from yvid 'find', of which the various persons and modes are 
more frequent and in fuller variety than those of any other verb. Only the 
forms actually quotable are instanced; those of which the examples found 
are from other verbs than vid are bracketed. Thus : 

active. middle. 

s. d. p. s. d. p. 

1 avidam dvidava dvidama dvide [dvidavahi] dviddmahi 

2 dvidas [avidata] [dvidathds] 

3 dvidat dvidan [avidata] [avidetdm] dvidanta 

The middle forms are rare in the earlier language, as in the later: we 
have dhve etc., dkhye etc., dvide (?) and avidanta, avocathds and avocdvahi 
(and aviddmahe GB. and asicdmahe KB. are doubtless to be amended to mahi). 

Augmentless forms, with indicative or subjunctive value, are not in- 
frequent. Examples, showing accent on the tense-sign, according to the 
general analogies of the formation, are ruhdm, srpas, bhujdt, viddt, aratdm, 
vocata, fakan; vidata (3d sing.), ardmahi, fisdmahi, viddnta, budhdnta, 
mrsanta (for exceptions as regards accent, see below, 853). 

Modes of the a-aorist. 

849. The subjunctive forms of this aorist are few; those which occur 
are instanced below, in the method which was followed for the indicative : 

1 [vidava] viddma [vidamahe] 

2 < ,_, viddthas viddtha 

3 viddt [viddtdi ?] 

The ending ihana is found once, in risdthana. Of middle forms occur 
only tfsdtdi (AV.: but doubtless misreading for (fsydtdi) and fisdmahe (AV,, 
for RV. fisdmahi). 


850. The optatives are few in the oldest language, but become more 
frequent, and in the Brahman as are not rare. Examples are : in active, 
videyam, saneyam ; vides, games ; garnet, vocet ; gametam ; pafeema, sanema ; 
vareta; in middle, (only) videya; gamemahi, vanemahi. 

A single middle precative form occurs, namely videsta (AV., once); it 
is so isolated that how much may be inferred from it is very questionable. 

851. A complete series of active imperative forms are made from ysad 
(including sadatana, 2d pi.), and the middle sadantam. Other imperatives 
are very rare: namely, sdra, ruha; ruhdtam, viddtam; khydta ; and, in middle, 
stuvatam (?). 

Participles of the fl-aorist. 

852. The active participles trpdnt, rudhant, vrdhdnt, fisdnt, fucdnt, 
huvdnt, savant, and (in participial compounds, 1309) kftant-, guhant-, 
vidant- (all RV.) are to be assigned with plausibility to this aorist. 

Likewise the middle participles vrdhand, sridhdnd, huvand; and perhaps 
also vipand and cubhand, although no personal forms corresponding to them 

Irregularities of the -aorist. 

853. A few irregularities and peculiarities may be noticed here. 

The roots in r, which (847) show a strengthening like that of the 
present of the sixth or unaccented a-class, have also the accent on the 
radical syllable, like that class: thus, from j/r, dranta (augmentless 3d pi.), 
sdrat and sdra. The root sad follows the same rule: thus, sddatam; and 
from ysan are found sdnas and sdnat and sanema, beside saneyam and 
sanema. It is questionable whether these are not true analogues of the sixth- 
class (unaccented a-class) present-system. On the other hand, ruhat (beside 
ruhdm, ruhdva, ruhdtam} and rfsat (only accented form) are anomalies. From 
yvac, the optative is accented voceyam, vocema, voceyus ; elsewhere the 
accent is on the root-syllable : thus, vdce, v6cat, vdcati, vdcanta. 

854. The stem voc has in Vedic use well-nigh assumed the value of 
a root ; its forms are very various and of frequent use, in RV. especially far 
outnumbering in occurrences all other forms from yvac. Besides those already 
given, we find voca (1st sing, impv.) and vocdti, vocavahai; voces, voceya, 
vocemahi; vocatdt (2d sing.), vocatu, vocatam, vocata. 

Of the stem neya from j/nap only ne^at occurs. 

The root pas (as in some of its present forms : 639) is weakened to pz, 
and makes afisam. 

855. Isolated forms which have more or less completely 
the aspect of indicative presents are made from some roots be- 
side the aorist-systems of the first two classes. It must be left 
for maturer research to determine how far they may be relics 
of original presents, and how far recent productions, made in 
the way of conversion of the aorist-stem to a root in value. 




II. (3) Reduplicated Aorist. 

856. The reduplicated aorist is different from the other 
forms of aorist in that it has come to be attached in almost 
all cases to the derivative (causative etc.) conjugation in 
TCJ dya, as the aorist of that conjugation, and is therefore 
liable to be made from all roots which have such a conju- 
gation, beside the aorist or aorists which belong to their 
primary conjugation. Since, however, the connection of 
the two is no formal one (the aorist being made directly 
from the root, and not from the causative stem), but rather 
a matter of established association, owing to kinship of 
meaning, the formation and inflection of this kind of aor- 
ist is best treated here, along with the others. 

857. Its characteristic is a reduplication of the radical 
syllable, by which it is assimilated, on the one hand, to 
the imperfect of the reduplicating class (II.), and, on the 
other hand, to the so-called pluperfect. But the aorist re- 
duplication has taken on a quite peculiar character, with 
few tracas left even in the Veda of a different condition 
which may have preceded this. 

858. As regards, indeed, the consonant of the redupli- 
cation, it follows the general rules already given (590). And 
the quality of the reduplicated vowel is in general as in the 
formations already treated: it needs only to be noted that 
an a-vowel and r (or ar) are usually (for exceptions, see 
below, 860) repeated by an ^-vowel as they are, to a 
considerable extent, in the reduplicated present also (660). 

But in regard to quantity, this aorist aims always at 
establishing a diversity between the reduplicating and radi- 
cal syllables, making the one heavy and the other light. 
And the preference is very markedly for a heavy redupli- 
cation and a light root- syllable which relation is brought 
about wherever the conditions allow. Thus : 


859. If the root is a light syllable (having a short 
vowel followed by a single consonant), the reduplication is 
made heavy. 

And this, usually by lengthening the reduplicating vowel, 
with i for radical a or r or / (in the single root containing that 
vowel) : thus, arlnsam, adudusam, ajljanam, avivrdham, aciklpam. 
The great majority of aorists are of this form. 

If, however, the root begins with two consonants, so that 
the reduplicating syllable will be heavy whatever the quantity of 
its vowel, the vowel remains short : thus, aciksipam, acukrudham, 
atitrasam. apisprcam. 

860. If the root is a heavy syllable (having a long 
vowel, or a short before two consonants), the vowel of the 
reduplication is short : and, in this case 5f a or (T a, and 
fT r (if it occurs), are reduplicated by f a. 

Thus adidlksam, abubhusam, adadaksam, adadhavam, atatahsam. 
And, in the rare cases in which a root both begins and ends 
with two consonants, both syllables are necessarily heavy, not- 
withstanding the short vowel in the former : thus, apapraccham, 
acaskandam (neither, apparently, found in use). 

These aorists are not distinguishable in form from the so-called pluper- 
fects (817ff.). 

861. In order, however, to bring about the favored relation 
of heavy reduplication and light radical syllable, a heavy root is 
sometimes made light : either by shortening its vowel, as in 
avivacam from y'vac, aslsadham from ~\/sad/i, adidipam (K. and 
later: RV. has didlpas] from y dip, abibhisam from y bhis, asu- 
sucam from ]/ sue ; or by dropping a penultimate nasal, as in 
acikradam from ]/ krand, asisyadam from ~|/ syand. 

In those cases in which (1047) an aorist is formed directly 
from a causal stem in ap, the a is abbreviated to i : thus, atisthip- 
am etc., ajijnipat, jihipas, ajijipata (but VS. ajijapata]; but from 
crap comes acicrapama 

862. Examples of this aorist from roots with initial vowel are very rare; 
the older language has only amamat (or amamat) from ]/am, and arpipam 
(augmentless) from the causative stem arp of ]/r in which latter the root 
is excessively abbreviated. The grammarians give other similar formations, 
as arcicam from ]/arc, aubjijam from ]/a&/, arjiham from |/arfe, aiciksam 
from yiks, ardidham from yrdh. Compare the similar reduplication in de- 
siderative stems: 1029 b. 

863. Of special irregularities may be mentioned : 


From ydyut is made adidyutam (taking its reduplicating vowel from the 
y instead of the u); yplu makes apiplavam (QB. etc.). 

Some verbs with radical a or r are by the grammarians allowed to re- 
duplicate with either i (1} or a, or even with a only. Others are allowed 
either to retain or shorten a long root-vowel. Details are unnecessary, the 
whole formation being so rare, and the forms instanced having never been 
met with in use. 

As to apaptam, avocam, and anepam, see above, 847. 

864, The inflection of the reduplicated aorist is like 
that of an imperfect of the second general conjugation : that 
is to say, it has the union-vowel ^ a before the endings, 
with all the peculiarities w r hich the presence of that vowel 
conditions. Thus, from y^ftjan, 'give birth': 

active. middle. 

s. d. p. d. p. 

MsJlsMIH' MsflsH M 

djijanam djijanava djljanama djyane djijanavahl djijanamahi 

^ ^ 

djljanas djyanatam djyanata djtjanathas djijanetham dfyanadhvam 

fclsilswr^ ^siWlrilH^ WlsH*^ *tellstlrl WlsHcilH^ fctsflswri 

ajijanat ajijanatam djijanan djyanata dfljanetam djljananta 

865. The middle forms are quite rare in the older language 
(the 3d pi. is decidedly the most common of them, being made 
from eleven roots ; the 3d s. from seven) ; but all are quotable 
except 1st and 2d du. and of the 1st du. no active example 

Atitape appears to be once used (RV.) as 3d sing., with passive sense. 

866. A final r has the ^wwa-strengthening before the end- 
ings : thus, apiparam, atitaras, adidharat, avivaran. Of similar 
strengthened forms from i and w-roots are found apiprayan (TS.), 
abibhayanta (RV.), apiplavam (QB.), acucyavat (K.); of unstrength- 
ened, acucruvat (GB.). Few roots ending in other vowels' than r 
make this aorist : see below, 868. 

867. Forms of the inflection without union-vowel are occasionally met 
with: namely, from roots ending in consonants, sfsvap (2d sing., augmentless) 
from ysvap, and apipnat from y$nath; from roots in r or ar, didhar (2d sing.), 
and ajigar (2d and 3d sing.). Of 3d pi. in us are found almost only a form 
or two from i and w-roots, with guna before the ending: thus, apipraj/ws, 
acucyavus, apwpravws, asusavus (? AB.); but also ablbhajus (QB.). And the 
3d pi. mid. avavrtran and avavrtranta and asasrgram seem to belong here 
rather than to the pluperfect. 


868. In the later language, a few roots are said by the 
grammarians to make this aorist as a part of their primary con- 
jugation : they are cri and cvi, dru and sru, kam, and dha 'suck' 

(cvi and dha optionally). 

In the older language are found from y$ri a$i$ret and afifrayus (noticed 
in the preceding paragraph); from ydru, adudrot and adudruvat (TB.: not 
used as aorist); from j/sra, asusrot and (augmentless) susros and susrot. Of 
forms analogous with these occur a number from roots in u or u : thus, 
anunot and nunot from ynu; yuyot from yyu 'repel'; dudhot from ydhu; 
apupot from ]/pu,- tutos and tutot from ytu; and one or two from roots in i 
or i: thus, siset from ysi (or so) 'bind'; apipres (with apiprayan, noticed 
above) from yprl (and the "imperfects" from dldhl etc., 676, are of corres- 
ponding form). And from ycyu are made, with union-vowel z, acucyavit 
and acucyavltana. None of these forms possess a necessarily causative or a 
decidedly aoristic value, and it is very doubtful whether they should not all 
be assigned to the perfect-system. 

Modes of the Reduplicated Aorist. 

869. As in other preterit formations, the augmentless in- 
dicative persons of this aorist are used subjunctively, and they 
are very much more frequent than true subjunctives. 

Of the latter are found only riradha (1st sing.); titapasi; ciklpati and 
sigadhati, and pisprfati (as if corresponding to an indicative apisprk, like 
apifnat); and perhaps the 1st sing. mid. $a$vacaf. 

The augmentless indicative forms are accented in general on the redu- 
plication : thus, didharas. nma^as; jijanat, piparat; jijanan; also sisvap ; 
but, on the other hand, we have fifrdthas and $i$ndthat; and dudrdvat, 
pufrdvat, tustdvat (which perhaps belong rather to the perfect: compare 810). 
According to the native grammarians, the accent rests either on the radical 
syllable or on the one that follows it. 

870. Optative forms are even rarer. The least questionable case is the 
middle "precative" ririsista (ririsista has been ranked above with sasahlsta, 
as a perfect: 81 2 b). Cucyuvimahi and cucyavwata belong either here or to 
the perfect-system. 

871. Of imperatives, we have the indubitable forms pupurantu and 
Cifrathantu. And jigrtdm and jigrtd, and didhrtam and didhrtd (all RV. only), 
are doubtless to be referred hither, as corresponding to the indicatives (without 
union-vowel) ajigar and adldhar : their short reduplicating vowel and their 
accent assimilate them closely to the reduplicated imperfects (cl. II.), with 
which we are probably to regard this aorist as ultimately related. 

872. No participle is found belonging to the reduplicated aorist. 

873. The number of roots from which this aorist is met 
with in the earlier language is about ninety. In the classical 
Sanskrit it is very unusual ; in the whole series of later texts 
mentioned above (826) it occurs only once. 




III. Sibilant-Aorist. 

874. The common tense-sign of all the varieties of this 
aorist is a H s (convertible to ^ s) which is added to the 
root in forming the tense-stem. 

This sibilant has no analogues among the class-signs of the present- 
system ; but it is to be compared with that which appears (and likewise with 
or without the same union-vowel i) in the stems of the future tense- system 
(chap. XII.) and of the desiderative conjugation (chap. XIV.). 

To the root thus increased the augment is prefixed and 
the secondary endings are added. 

875. In the case of a few roots, the sibilant tense- 
stem (always ending in ^ ks] is further increased by an f a, 
and the inflection is nearly like that of an imperfect of the 
second general conjugation. 

876. In the vast majority of cases, the sibilant is the 
final of the tense-stem, and the inflection is like that of an 
imperfect of the first general conjugation. 

And these, again, fall into two nearly equal and strongly 
marked classes, according as the sibilant is added imme- 
diately to the final of the root, or with an auxiliary vowel 
^ i, making the tense-sign ^ w. Finally, before this "^ts 
the root is in a very small number of cases increased by a 
H s, making the whole addition ffpST^**. 

877. We have, then, the following classification for the 
varieties of sibilant-aorist : 

A. With endings added directly to the sibilant: 

4. with H s simply after the root: s-aorist; 

5. with ^ i before the "R s : *s-aorist ; 

6. the same, with 3f s at end of root: m-aorist. 

B. With Ef a added to the sibilant before the endings : 

7. with sibilant and 5f a: Sfl-aorist. 

As regards the distinction between the fourth and fifth forms, it may 
be said in a general way that those roots incline to take the auxiliary i in 
the aorist which take it also in other formations ; but it is impossible to lay 


down any strict rules as to this accordance, and also, considering the rarity 
of aorist-forms in the later language, practically quite useless to attempt the 
task. See below, 903. 

4. The s-aorist. 

878. The tense-stem of this aorist is made by adding 
^ s to the augmented root, of which also the vowel is usu- 
ally strengthened. 

879. The general rules as to the strengthening of the 
root-vowel are these : 

a. A final vowel (including ft r) has the vrddhi-ch&nge 
in the active, and (excepting ft r) guna in the middle : thus, 
from y^i active stem ERST anais, middle stem ^fa anes; 
from yy^fru, wmm agratis- and 5P2Tfa agros ; from |/5R kr, 
faiFSf akars and *33n^ akrs. 

^ c -s 

b. A medial vowel has the vrddhi-ch&uge in the active, 
and remains unaltered in the middle : thus, from y'^>*^ chand, 
active stem iirH achants, middle stem tj^rH achants ; from 

araiks and tii^rl ariks; from ]/"^I rudh, 
and 5T^fH aruts; from y'flsT srj] JblHW asraks 
and 5TH^T asr^s. 

C "S 

880. The endings are the usual secondary ones, with 
3TT us (not R ow) in 3d pi. act., and ^{ ata (not 3^tT wte) 
in 3d pi. mid. 

But before H s and cT ^ of 2d and 3d sing. act. is in 
the later language always inserted an ^ 2, making the end- 
ings ^T is and ^rT It. 

This insertion is unknown in the earliest language (of the RV.): see 
below, 888. 

881. Before endings beginning with t or th, the tense- 
sign s is (233 b) omitted after the final consonant of a root 
unless this be r, or n or m (converted to anusvara}. 

The same omission is of course made before dhvam, after either vowel 
or consonant; and the ending becomes dhvam, provided the sibilant, if retained, 
would have been s: thus, astodhvam and avrdhvam (beside astosata and 
avrsata), but aradhvam (beside arasata}. These three are the only test-cases 
for the form of the ending which have been noted in the older language, 




except drdhvam (ydr 'regard': ^B., once), which is to drthas (2d sing.) as 
avrdhvam and avrsata to avri and avrthas. 

According to the grammarians, the omission of s before t and th takes 
place also after a short vowel (the case can occur only in the 2d and 3d sing, 
mid.); but we have seen above (834 a) that this is to be viewed rather as 
a substitution in those persons of the forms of the root-aorist. Neither in 
the earlier nor in the later language, however, does any example occur of an 
aorist-form with s retained after a short vowel before these endings. 

882. As examples of the inflection of this variety of 
sibilant aorist we may take the roots "^J rudh, 'obstruct', 
and ^ rii, 'lead'. 'Thus: 

active. middle. 

s. d. p. s. d. p. 

drautsam drautsva drautsma drutsi drutsvahi drutsmahi 

drautsis drauttam drautta drutthas drutsatham druddhvam 

drautsit drauttam drautsus drutta drutsatam drutsata 

dnaisam, dnaisva dnaisma dnesi dnesvahi dnesmahi 

dnaisis dnaistam dnaista dnesthas dnesatham dnedhvam 


dnaisit dnaistam dnaisus dnesta dnesatam dnesata 

883. The omission of s in the active forms is a case of very rare oc- 
currence. In the older language have been noted only achantta (RV.; written 
achanta, by 231), abhakta (AB.), and the augmentless taptam and fapta 
(repeatedly in the Brahmanas). The middle forms with omission are in- 
distinguishable from the corresponding ones of the root-aorist; and whether, 
in the older language, a given form is to be assigned to the one aorist or to 
the other is a question difficult to settle. Above (834 b d) were given all 
the older forms of belongings thus questionable. 

884. Certain roots in a weaken the a in middle inflection 
to i (as also in the root-aorist: above, 834 a): these are said to 
be stha, da (da 'give', and da 'cut' and 'share'), and dha (dha 
'put', and dha 'suck'); in the older language have been noted 
only ddisi and adisata (from da 'give') and asthisata. TA. has 
aglsata from ]/y 'sing'. 

The middle inflection of the aorist of yda would be, then, 


according to the grammarians ; ddisi, ddithas, ddita ; ddisvahi, 
ddisatham, ddisatam; ddismahi, ddidhvam, ddisata. 

885. Roots ending in changeable r (so-called roots in f : 242) are said 
by the grammarians to convert this vowel to ir in middle forms : thus, astlrsi, 
astlrsthds, etc. (from ystr): hardly any such forms, however, have been found 
in the older language (only akirsata, PB.; and, on the contrary, astrsi occurs 
once, AB.). 

886. The s-aorist is made in the older language from some- 
what over a hundred roots (in RV., from about seventy; in 
AV., from about fifty, of which fifteen are additional to those 
in RV. ; and the other texts add about twenty more not count- 
ing in any case those of which the forms may be from the root- 
aorist). It has there certain peculiarities of stem-formation and in- 
flection, and also the full series of modes of which the optative 
middle is retained also in the later language as "precative". 

887. Irregularities of stem-formation are: 

a. The roots hu, dhu, and nu have u instead of o in the middle : thus, 
ahusata, adhusata, anusi and anusatam and anusata ; ydhur (or dhurv) makes 

b. From ymad occurs amatsus (RV., once), with unstrengthened vowel. 

c. From ygam occurs agasmahi, apparently for agansmahi (compare 
many a, below, 895). 

888. The principal peculiarity of the older language in re- 
gard to inflection is the frequent absence of ~i in the endings of 
2d and 3d sing, act., and the consequent loss of the consonant- 
ending, and sometimes of root-finals (150). The forms without 
are the only ones found in RV. and K., and they outnumber 
the others in AV. and TS.; in the Brahmanas they grow rarer 
(only one, adrak, occurs in GB. and QB.; PB. has none). 

889. If the root ends in a vowel, only the consonant of the ending is 
necessarily lost: thus, aprds (for both aprds-s and apras-t] from yprd: and 
in like manner ahas from yhd; ajais (for ajdis-t] from yji; and in like 
manner acdis from yd, and ndis (augmentless) from ym; and ydus (for 
aydus-t) from yyu. 

Bat (as in other like cases : 555) in 3d sing, the ending t is sometimes 
preserved at the expense of the tense-sign; and we have ajdit (beside ajais 
and ajdmt) from yji ; and in like manner acdit, afrdit, ahdit, ndit : no ex- 
amples have been noted except from roots in i and I. 

890. If the root (in either its simple or strengthened form) ends in a 
consonant, the tense-sign is lost with the ending. Thus, abhdr (for abhdrs-t : 
beside abharsam, abhdrstdm] from ybhr; other like cases are ahar, and (from 
roots in ar) aksar, atsar, asvar, hvar. Further, drdik (585, end: for ardiks-t) 
from yric; like cases are afvdit from ycvit, and (from roots with medial ) 
adydut from ydyut, ardut from yrudh, and mduk from ymuc. Further, 
from roots ending in the palatals and ft, aprdk from ]/prc, asrdk from |/sr;, 




abhdk from ybhaj, adrdk from }/drf, adhdk from ydah; but, with a different 
change of the final, ayat from yyaj, apra/ from j/prc/z, and avat from |/uafe , 
and (above, 146) sras appears to stand twice in AV. for sras-t from ysrj : 
RV. has also twice ay as from yyaj. Further, from roots ending in a nasal, 
atan from j/taw, khan from ykhan, aydn and anan from yyyam and nam 

If, again, the root ends in a double consonant, the latter of the two is 
lost along with tense-sign and ending: thus, achan (for achdnts-t; beside 
achantta and achdntsus] from ychand ; and other like cases are akrdn, asfcan, 
and asydn. 

A relic of this peculiarity of the older inflection has been preserved to 
the later language in the 2d sing, bhdis, from ybhi. 

891. Irregularities of inflection do not occur further: avesam from ym 
is too great an anomaly to be accepted. 

Modes of the s-aorist. 

892. The indicative forms without augment are used in a 
subjunctive sense, especially after mti prohibitive, and are not 
uncommon. Examples with accent, however, are extremely rare ; 
there has been noted only vdnsi, middle ; judging from this, 
the tone would be found on the radical syllable. According 
to the Hindu grammarians, it may be laid on either root or 

893. Proper subjunctive forms are not rare in RV., but 
are markedly less common in the later Vedic texts, and very 
seldom met with in the Brahmanas. They are regularly made 
with ^wm-strengthening of the radical vowel, in both active and 
middle, and with accent on the root. 

The forms with primary endings are : in active, stosdni ; darsasi ; nesati, 
parsati, pdsati, matsati, yosati, vaksati, saksati; dtisathas, dhdsathas, pdrsathas, 
vaksathas, varsathas ; pasatas, yaihsatas, yaksatas, vaksatas ; dhdsatha, nesatha, 
pdrsatha, mdtsatha; in middle, narhsdi, mahsdi; mdhsase ; krarhsate, trdsate, 
darsate, mdhsate, yaksate, rdsate, vahsate, sdksate, hdsate; trdsdthe (not 
trdsdithe, as we should rather expect); ndmsahte, mdhsante: and, with the 
fuller ending in 3d sing., mdsdtdi. 

The forms with secondary endings are (active only): jesas, vdksas; ddrsat, 
nesat, pdksat, ydksat, yosat, vdhsat, vdksat, vesat, sdtsat, chantsat, etc. 
(14 others); yaksatdm; stosdma; parsan, yamsan, yosan, rasan, vaksan, cesan, 
frdsan. Of these, yaksat and vaksat are found not rarely in the Brahmanas ; 
any others, hardly more than sporadically. 

894. Of irregularities are to be noted the following: 

a. The forms dfksase and prksase (2d sing, mid.) lack the #wna-strength- 

b. Jesam, stosam, and yosam (AV. yusam, with u for o as in anusata 
Whitney, Grammar. 19 


etc.) appear to be first persons formed under government of the analogy of 
the second and third unless they are relics of a state of things anterior 
to the vrddhi- strengthening: in which case jesma is to be compared with them 
(we should expect jdisma or jesdma}. 

c. From roots in d are made a few forms of problematic character : 
namely, yesam (only case in RV.), khyesam, jnesam, gesam and gesma, desma, 
sthesam and sthesus. Their value is optative. The analogy of jesam and 
jesma suggests the possibility of their derivation from i-forms of the a-roots ; 
or the sibilant might be of a precative character (thus, yd-i-s-am). That 
they are to be reckoned to the is-aorist appears highly improbable. 

d. The RV. has a few difficult first persons middle in se, which are 
perhaps best noted here. They are: 1. from the simple root, krse, hise (and 
ohise?), stuse ; 2. from present-stems, arcase, rnjase, yajase, gdyise, grnue 
and punlse. They have the value of indicative present. Compare below, 897. 

895. Optative forms of this aorist are made in the middle only, and 
they have in 2d and 3d sing, always the precative s before the endings. 
Those found to occur in the older language are : disiya, dhisiya, bhakslyd, 
masiya (for mahsiya], muksiya, rdsiya, sdksiya, strislya; mansisthds; darslsta, 
bhaksista, manslsta, mrkslsta; bhaksimahi, dhukstmdhi, mahslmdhi, vahsimdhi, 
saksimdhi; mahslrata. PB. has bhuksisiydj which should belong to a sz's-aorist. 
The RV. form trdsltham (for trdslydthdm or trdsdthdm) is an isolated anomaly. 

This optative makes a part of the accepted "precative" of the later 
language: see below, 921 ff. 

896. Imperative persons from this aorist are extremely rare : we find 
only the 2d sing. act. nesa and parsa (both from a-stems, and showing 
rather, therefore, a treatment of the aorist-s'tem a,s a root), and the 3d sing, 
mid. rdsatdm and pi. rdsantdm (of which the same may be said). 

Participles of the s- Aorist. 

897. Active participles are ddksat or dhdksat, and sdksat (both RV.). 

If rnjase (above, 894 d) is to be reckoned as an s-aorist form, rnjasdnd 
is an s-aorist participle ; and of a kindred character, apparently, are arfasdndj 
tihasdna, jrayasdnd, dhiyasdnd, mandasdnd, yamasdnd, rabhasdnd, vrdhasdnd, 
sahasdnd, cavasdnd, all in RV.; with namasdnd, bhiydsdna, in AV. 

5. The ^5-aorist. 

898. The tense-stem of this aorist adds the general 
tense-sign H s by help of a prefixed auxiliary vowel ^ , 
making ^T is, to the root, which is usually strengthened, 
and which has the augment. 

899. The rules as to the strengthening of the root are 
as follows: 

902] SIBILANT AORIST: 0. ^6 1 -AORIST. 291 

a. A final vowel has vrddhi in the active, and guna in 
the middle: thus, ^Mlf^N apavis and SOT^R" apams from yq 
pu; WH^^itaris. act., from y c[ #V ^lilUlN p<W, mid., 
from y 5TT f. 

b. An interior vowel has guna, if capable of it, in both 
voices: thus, ti^fem ale$is, act. and mid., from yf^5T lig; 

~x r*~ . r^r "^ 

i(llT(q arocts from y"^tf -rwc; Njq&j avarsts from 
but CfeiV|G|fc| ajlvis from y sffo jw. 

c. Medial f a is sometimes lengthened in the active; 
but it more usually remains unchanged in both voices. 

The roots in the older language which show the lengthening are kan, 
san, ran, stem, tans, vraj, vad, mad, car, tsar, hvar, jval, das. From svan 
and sah occur forms of both kinds. 

900. Of exceptions may be noted: yrnr? has (as elsewhere : 627) vrddhi 
instead of guna: thus, amarjisam; ystr has astaris, and j/fr has aparit (also 
a far ait in AV.), with guna in active. 

The root grabh or grah has (as in future etc., below, 935 d, 956 ) long 
i instead of i before the sibilant : thus, agrabhisma, agrahlsta, agrabhlsata. 
The roots in changeable r (so-called roots in f : 242), and yvr are said by 
the grammarians to do the same optionally; but no forms with long I from 
such roots are found in the older language. 

901. The endings are as in the preceding formation 
(3H us and ^rT ata in 3d pi.). But in 2d and 3d sing., 
the combination is-s and is-t are from the earliest period 
of the language contracted into "$3 is and ^rT^. 

The 2d pi. mid. should end always in idhvam (or iddhvam, 
from is-dhvam : 232) ; and this is in fact the form in the only 
examples quotable from the older language, namely ajanidhvam 
and artidhvam and aindhidhvam ; but the grammarians give rules 
by which the lingual dh is optional only, and that after i pre- 
ceded by g, r, I, v, h. 

902. As examples of the inflection of the ^ f s-aorist may 
be taken the roots ^pu, 'purify', and ^Uf^budh, 'awake 1 . 
Thus : 

active. middle. 

s. d. p. s. d. p. 

dpavisam dpavisva dpavisma dpavisi dpavisvahi dpavismahi 



dpavis dpavistam dpavista dpavisthas dpavi&atham dpavidhvam 

apamt dpavistam dpavisiis dpavista dpavisatam dpavisata 

dbodhisam dbodhisva dbodhisma dbodhisi dbodhisvahi dbodhismahi 

etc. etc. etc. etc. etc. etc. 

903. The number of roots from which forms of this aor- 
ist have been noted in the older language is about a hundred 
and twenty (in RV. , about eighty; in AV. more than thirty, of 
which a dozen are additional to those in RV. ; in the other texts, 
nearly thirty more). Among these are no roots in a ; but other- 
wise they are of every variety of form (rarest in final i and i). 
Active and middle persons are freely made ; but very sparingly 
from the same root (only about fifteen roots have both active 
and middle forms, and of these a part only exceptionally in the 
one voice or the other). 

No rule appears to govern the choice of usage between the 
is and the s-aorist ; and in no small number of cases (more than 
a fifth of all), the same root shows forms of both classes. 

904. Irregularities of the older language are to be noticed as follows : 

a. The contracted forms aferarram, agrabhlm, and avadhtm (with aug- 
mentless vddhim] are found in 1st sing. act. 

b. For afarlt occurs in AV. a$arait; also (in a part of the MSS.) 
farais for pans : agrahaisam is found in AB. (also the monstrous form 
ajagrabhaisarh : see 801 i). 

c. From yvad is found vadisma (once, AB.), with short root-vowel. AV. 
has nudisthas, without guna. 

d. The forms atarima (RV.) and avddiran (AV.), though they lack the 
sibilant, are perhaps to be referred to this aorist. 

e. Ajayit, with short i in the ending, occurs in TS. 

Modes of the /s-aorist. 

905. As usual, augmentless indicative forms of this aorist are more 
common than proper subjunctives. Examples, of all the persons found to 
occur (and including all the accented words), are, in the active: pdrmsam, 
vddhim; mathis, vddhls, ydvls, sdvls; dvlt, jurvit, mdthit, vddhlt,; 
mardhistam, dosistam, hihsistam; avistdm, jdnistam, bddhistdm; framima, 
vadisma; vadhista arid vadhistana, mathistana, hihsista; hvdrisus, grahisus ; 
in the middle : rddhisi ; jdnisthds, marsisthas, vyathisthds ; krdmista, janista, 
pavista, prdthista, mdndista; vyathismahi. The accent is on the root-syllable 
(tarisus, AV. once, is probably an error). 


906. Of subjunctive forms with primary endings occur only the 1st sing, 
act. davisani, and the 1st pi. mid. (with unstrengthened e) yacisamahe and 

Forms with secondary endings are almost limited to 2d and 3d sing. act. 
There are found : avisas, kanisas, tarisas, raksisas, vddhisas, vddisas, vesisas, 
fahsisas ; karisat, jambhisat, josisat, tarisat, nfndisat, pdrisat, Mdhisat, 
mdrdhisat, yacisat, yodhisat, raksisat, vanisat, vyathisat, fahsisat, sanisat, 
savisat. They are made, it will be noticed, with entire regularity, by adding 
a to the tense-stem in is before the endings. The only other persons found 
to occur are the 3d pi. act. sanisan and mid. sdnisanta, which are also 
regular. Bhavisat (AB. once) is a solitary example of a form with double 
mode-sign. The radical syllable always has the accent, and its vowel usually 
accords with that of the indicative : but we have san- in the subjunctive 
against asanisam (as to cay- and ran-, see below, 908). 

907. The middle optative of this aorist also forms a part of the ac- 
cepted "precative" of the later language. It is very rare at all periods, being 
made in RV. from only five roots, and in AV. from two of the same and 
from three additional ones (six of the eight have other w-forms); and the 
remaining texts add, so far as noticed, only four other roots. All the forms 
found to occur are as follows : janislya, indhislya, edhislyd, rucislya and 
rocisiya, gmisiya; modisisthas ; janisista, vanislsta ; sahisivahi- edhislmdhi, 
janisimahi, tarislmahi, mandisimahi, vandisimdh^ vardhisimdhi, sahislmahi 
and sahislmahi. The accent is on the ending, and this would lead us to 
expect a weak form of root throughout ; but the usage in this respect appears 
to be various, and the cases are too few to allow of setting up any rule. 

908. Of imperative forms, we have from yav a series : namely, aviddhf, 
avistu, avistdm, avitd (if this, as seems probable, stands anomalously for 
avistd) and avistdna : two of these are of unmistakably imperative form. 
Other forms occur only in 2d du. and 2d pi., and are accordingly such as 
might also be subjunctives used imperatively (which is further made probable 
for two of them by their accentuation on the root-syllable): they are kram- 
istam, gamistam, cayistam (against acayisam], tSristam, yodhistam, vadhistam, 
pnathistam; rdnistana (against aranisus), pnathistana. 

909. No words having a participial ending after is are found 
anywhere to occur. 

910. This is the only aorist of which forms are made in 
the secondary and denominative conjugations : see below, chap. 
XIV. (1019, 1035, 1048, 1068). 

6. The sss-aorist. 

911. According to the grammarians, this aorist is made 
from roots in 5(T a (including ft mi 'establish', ft mi 'di- 
minish', and ^ft ll 'cling', which substitute forms in a), and 


from ^t^nam, FTj/#m, and JT^ram, and is used only in the 
active; the corresponding middle being of the s-form (4). 
Its inflection is precisely like that of the ^s-aorist ; it is un- 
necessary, then, to give more than its first persons, which 
we may form from the roots ETT yci, 'go', and ^ nam, 'bend'. 

s. d. p. s. d. p. 

dyasisam dyasisva dyasisma dnamsisam dnamsisva dnamsisma 

etc. etc. etc. etc. etc. etc. 

912. The szs-aorist is properly only a sub-form of the zs-aorist, having 
the tense-sign and endings of the latter added to a form of root increased 
by an added s. It is of extreme rarity in the older language, being made 
in RV. only from the roots ga and yd, and in AV. only from fta, and probably 
from pya and van (see below, 914): the remaining texts add jna (TB.) and 
dhd (QB.); the forms adrdszt (K.) and ahvdslt (GB.) might "be either from 
this or from the s-aorist. 

The participle hdsamdna and causative hdsayanti (RV.) show that has 
had assumed, even at a very early period, the value of a secondary root 
beside hd for other forms than the aorist. 

913. The whole series of quotable indicative forms is as follows: 
aydsisam, adhdsisam; agdsls ; agdslt, aydsit (and adrdslt and ahvdslt ?); 
aydsistdm; ajndsisma; ajndsista, aydsista; agdsisus, aydsisus (dksisus is 
from yaks}. 

Forms without augment are these: hdsisam; hdsls ; hdslt, ydslt; hdsistam; 
hdsistdm; hdsista; hdsisus, gdsisus. The accent would doubtless be upon 
the root-syllable. 

914. Of proper subjunctives are found two, gdsisat and ydsisat (both RV.). 
Optatives are not less rare : namely, ydsislsthds and pydsisimahi (for 

which the AV. manuscripts read pydfisimahi, altered in the edition to pydyis-); 
and doubtless vahfisiya (AV., twice) is to be corrected to vohsisiya, and 
belongs here. 

The accent of ydsistdm (like avistdm, 908) shows it to be a true im- 
perative form; and ydsista (RV., once) is doubtless the same, with anomalous 
I for i. 

915. Middle forms of this aorist, it will be noticed, occur from the 
optative only; but, considering the great rarity of the whole formation, we 
are hardly justified in concluding that in the ancient language the middle 
persons in -sisi, -sisthds, etc., were not allowable, like those in -isi, -isthds, 
and the others of the i-aorist. 

7. The sa-aorist. 
916. In the later language, the roots allowed to form 

920] SIBILANT AORIST: 7. sa-AomsT. 295 

this aorist end in S^p, E[^, or ^ h - - all of them sounds 
which in combination with the tense-sign make ft ks ; and 
they have ^ i, 3 u, or ft r as radical vowel. 

They are : dip, r if , Zip , uip, A;i'p, femp, mp, mrp, #prp , tvis, dvis, flis, 
vis, krs; dih, mih, lih, guh, duh, ruh, trh, wft, strh (Kielhorn). Some of 
them may, or with certain meanings must, take aorists of other forms. And a 
few are allowed to drop both tense-sign and union-vowel a in certain persons 
of the middle : that is, they may make instead forms of the root-aorist (1). 

917. As the tense-stem ends in 5j 7 the inflection is 
in the main like that of an imperfect of the second general 
conjugation. But (according to the grammarians : the forms 
unfortunately have not been found in the older language) 
the 1st sing. mid. ends in ^ i instead of ^ e, and the 2d 
and 3d du. mid. in ETTSfFT atham and *JIHH atam, as in imper- 
fects of the other conjugation. Both active and middle in- 
flection is admitted. The root is throughout unstrength- 

918. As example of inflection we may take the root 
$pt dip, 'point out'. Thus : 

active. middle. 

s. d. p. s. d. p. 

ddiksam ddiksava adiksama ddiksi ddiksavahi ddiksamahi 


ddiksas ddiksatam ddiksata ddiksathas ddiksatham ddiksadhvam 

ddiksat ddiksatam ddiksan ddiksata ddiksatam ddiksanta 

919. In the earlier language, the forms of the sa-aorist are hardly more 
than sporadic, and are with much probability to be regarded as transfers of 
the s-aorist to an inflection after the manner of an a-stem. They are made 
in RV. from eight roots ; in AV. , from two of these and from two others ; 
and the remaining texts add six more, making sixteen in all. As later, 
all have i or u or r as root-vowel, and a final consonant which combines 
with s to ks; but there are in the list also two ending in j (unless the forms 
ascribed to mrj and vrj be forced under mrp and vrh). All the examples 
noted are given below. 

920. In the indicative, we find, in the active: avrksam; adruksas, 
aruksas, asprksas; adiksat, amiksat, aviksat, dkruksat, aghuksat, aduksat and 


ddhuksat, druksat, akrksat, dmrksat, dsprksat; aghuksatdm; aruksama, amr- 
ksdma, avrksdma; ddhuksan, apiksan (j/pis), aruksan; in the middle, only 
akrksathas (]/Ars), adhuksata, and amrksanta. 

Forms without augment (no true subjunctives occur) are, in the active : 
drksam, mrksam; duksas, ruksas, mrksas; dviksat ; mrksata; dhuksdn and 
duksan; in the middle, dviksata, duksata and dhuksata, dhuksdnta. 

There are no optative forms. 

Imperative are : in the active, mrksatam ; in the middle, dhuksdsva. 

The few accented forms without augment which occur have the tone on 
the tense-sign sd, in analogy with the a-aorist (2) and the imperfect of the 
a-class (VII.): a single exception is dhtiksata, which probably needs emend- 
ation to dhuksata. 

The aspiration of initial d and p, after loss of the aspirated quality of 
the root-final (155), is seen in forms from the roots duh and guh, but not 
from druh (only a single case, AB.); RV., however, has also aduksat and 
duksas, duksan, duksata. 


921. As the so-called precative is allowed by the gram- 
marians to be made in the later language from every root, and 
in an independent way, without reference to the mode of form- 
ation of the aorist from the same root, it is desirable to put 
together here a brief statement of the rules given for it. 

922. The precative active is madej by adding|the active 
precative endings (above, 568) directly to the root. But : 

a. Of final root-vowels (as before the passive-sign yd : 770), 
i and u are lengthened; r is usually changed to ri, but to ir 
and ur in those roots which elsewhere show ir and ur forms 
(so-called f-roots : 242), and to ar in r and smr ; a is changed 
to e in the roots da, dha, stha, pa 'drink', go, 'sing', and a few 
others, in part optionally. 

b. The root in general assumes its weakest form : a penulti- 
mate nasal is lost, as in badhyasam from y bandh ; the roots 
which are abbreviated in the weak persons of the perfect (794) 
have the same abbreviation here, as in ucyasam, ijyasam, vidhya- 
sam, supyasam, grhyasam; yeas forms cisyasam (compare 639, 
854) : and so on. 

In the older language also, as has been seen above (838), precative 
optative forms are made in the active only for the root-aorist, and in a 
manner accordant with that here described. 

923. The precative middle is made by adding the middle 
precative endings (above, 568) to the root increased by H s 

925] PRECATIVE. 297 

or "$& ts - - that is, to the tense-stem of an s-aorist or of 
an eis-aorist (but without augment). 

The root is strengthened according to the rules that 
apply in forming the middle-stem of the s and of the is- 
aorists respectively: in general, namely, a final vowel is 
gunated in both formations ; but a medial vowel, only be- 
fore ^T is. 

Other minor rules it is unnecessary to repeat here. 

In the older language, as has been pointed out in detail above, prec- 
ative optative forms of the middle voice are oftenest made from the s-aorist 
(895) and the w-aorist (907); but also from the root-aorist (838), the a- 
aorist (850), the reduplicated aorist (870), and the w-aorist (914); and 
even from the perfect (812 b). 

924. As example of inflection, we may take the root 
H bku, 'be', which is said (no middle aorist or precative is 
made from it in the older language) to form its middle on 
the s-stem . Thus : 

active. middle. 

s. d. p. s. d. p. 


c\_ , "s c\_ , <r\_ f _ . _ . . 

bhuyasam bhuyasva bhuyasma bhavisiyd bhaviswdhi bhavisimdhi 

_ . _ f _ , f _ 

bhuyas bhuyastam bhuyasta bhavisisthas bhavisiyastham bhavisidhvdm 


cx t "^ SNs _r "^ . _ L 

bhuyat bhuyastam bhuyasus bhavislstd bhavisiyastam bhavisirdn 

According to the grammarians, the dental or lingual character of the 
initial of the middle ending dhvam depends upon how the aorist tense-sign 
is preceded: in the s-form, it is dhvam if the si is preceded by any other 
vowel than a or d; in the z's-form, it may be optionally dhvam if the is is 
preceded by y, r, I , v, or h. This seems wholly irrational : the true question 
is, whether the precative s is to be regarded as really present in 2d pi. mid., 
as in all other 2d persons of both voices : if so which is altogether prob- 
able, but in the absence of quotable forms from the older language cannot 
be pronounced certain the ending is necessarily and always dhvam. 

925. The precative is a form of rare occurrence in the 
classical language. In each of the texts already more than once 
referred to (Manu, Nala, Bhagavad-GIta, Qakuntala, Hitopadeca) 


it occurs once and no more. Its value, as already stated (573), 
is purely optative : thus, sarvarastresv idam vaco bruyasta (Nala), 
'[I beg that you] speak these words in all kingdoms'. 

Uses of the Aorist. 

926. The uses of the aorist mode-forms (as has been 
already pointed out : 582) appear to accord in general with those 
of the mode-forms of the present-system. The predilection of 
the earlier language, continued sparingly in the later, for the 
augmentless forms in prohibitive expression after mU, was suffi- 
ciently stated and illustrated above (579). 

The tense-value of the aorist indicative has also been more than once 
referred to, and calls only for somewhat more of detail and for illustra- 
tion here. 

927 . The aorist of the later language is simply a pret- 
erit, equivalent to the imperfect and perfect, and frequently 
coordinated with them. 

Thus, tatah sa gardabham lagudena tddaydmdsa; tend J sdu pancatvam 
agamat (H.), 'thereupon he beat the donkey with a stick ; and hereof the 
latter died'; tatah sa vidarbhdn agamat punah , tarn tu bandhujanah sama- 
pujayat (MBh.), 'thereupon she went back to Vidarbha; and her kindred paid 
her reverence'; pritimdn abhut: uvdca cat J nam (MBh.), 'he was filled with 
affection, and said to hinr. 

928. The aorist of the older language has the value of a 
proper "perfect": that is, it signifies something past which is 
viewed as completed with reference to the present ; and it requires 
accordingly to be rendered by our tense made with the auxiliary 
have. In general, it indicates what has just taken place ; and 
oftenest something which the speaker has experienced. 

Examples from the Veda are : pdrl 'me gdm anesata pdry agn'fm ahrsata, 
devesv akrata frdvah kd imdn d~ dadharsati (RV.), 'these here have led about 
a cow, they have carried around the fire, they have done honor to the gods 
who shall venture anything against them V yam dichdma mdnasd sb 'yam 
a 'gat (RV.), 'he whom we (formerly, impf.) sought with our mind has (now, 
aor.) come'; yene 'ndro havtsd krtvy dbhavad dyumny uttamah, idam tad akri 
devd asapatndh kilo, 'bhuvam (RV.), 'that libation by which India, making it, 
became (impf.) of highest glory, I have now made, ye gods; I have become 
free from enemies'. 

Examples from the Brahmana language are : sd ha 'smih jydg uvdsa . . . 
tdto ha gandharvah sdm udire: jydg vd iydm urvdfl manusyesv avatslt (QB.), 
'she lived with him a long time. Then the Gandharvas said to one another, 
"this Urvac.i, forsooth, hath dwelt a long time among mortals'"; tasya ha 
dantah pedire: tarn ho J vaca: apatsata vd asya dantdh (AB.), 'his teeth fell 




out. He said to him: "his teeth truly have fallen out'"; indrasya vrtrdm 
jaghnusa indriydrh viryam prthivim dnu vy hrchat tad dshadhayo vlrudho 
l bhavan sd prajapatim upa 'dhdvad vrtrdm me jaghnusa indriydm viryam 
prthivim dnu vy arat tad dshadhayo vlrudho l bhuvann iti (TS.), 'of Indra, 
when he had slain Vritra, the force and might went away into the earth, 
and became the herbs and plants : he ran to Prajapati, saying : "my force 
and might, after slaying Vritra, have gone away into the earth, and have 
become the herbs and plants'"; svaydm enam abhyudetya bruydd vrdtya kvh 
'vatslh (AV., in prose passage), 'going up to him in person, let him say: 
"Vratya, where hast thou abode"?' ydd iddnlm dvau vivddamandv eydtam ahdm 
adar^am ahdm a^rdusam iti yd evd bruydd ahdm adar?am iti tdsmd evd 
yrdddadhydma (QB.), 'if now two should come disputing with one another, 
[the one] saying "I have seen", [the other] "I have heard", we should believe 
the one who said "I have seen'". 

929. This distinction of the aorist from the imperfect and perfect as 
tenses of narration is very common in the Brahmanas, and is closely observed : 
neglect of it is very rare, and is to be regarded as either due to corruption 
of text or indicative of a late origin. 

In the Vedic hymns, the same distinction is prevalent, but is both less 
clear and less strictly maintained : many passages would admit an inter- 
pretation implying either sense; and evident aorist-forms (especially of the 
simple aorist: Delbruck) are sometimes used narratively, while imperfect- 
forms are also occasionally employed in the aorist sense. 

930. The boundary between that which has just been and that which 
now is is occasionally overstepped, and the aorist becomes nearly the equiv- 
alent of a present. Not very rarely, in the Veda, it is convenient to render 
the former as if it were the latter; and in the Brahmana the same is true 
especially of the aorist a/car. 



931. THE verb has two futures, of very different age 
and character. The one has for tense-sign a sibilant follow- 
ed by IT yet, and is an inheritance from the time of Indo- 
European unity. The other is a periphrastic formation, made 
by appending an auxiliary verb to a derivative noun of 




agency, and it is a recent addition to the verb-system; its 
beginnings only are met with in the earliest language. The 
former may be called the s- future (or the old future, or 
simply the future,; the latter may be distinguished as the 
periphrastic future. 

I. The s-future. 

932. The tense-sign of this future is the syllable HT 
syd, added to the root either directly or by an auxiliary 
vowel ^ i (in the latter case becoming ^ET isyd). The root 
has the ^wwa-strengthening. Thus, from |/^T da is formed 
the future tense-stem ^THT dasyd; from ]/^ i, the stem ^T 
esyd ; from y~^> d*uh, the stem J7??J dhoksyd; from yH bhu, 
the stem Hf^W Wiavisyd; from y'fJEJ rdh, the stem ^[MCVU 
ardhisyd; and so on. 

But from yjw the stem is jivisyd, from y uks it is uksisyd, 

and so on (240). 

933. This tense-stem is then inflected precisely like a 
present-stem ending in f a (second general conjugation). 
We may take a# models of inflection the future of ]/^T da, 
'give', and that of ]/5R kr, 'make'. Thus: 



dasyhmi dasyhvas dasytimas dasye 

dasytivahe dasyamahe 

dasyasi dasydthas ddsydtha dasydse dasyetfie dasyddhv 

ddsydti dasydtas ddsydnti ddsydte ddsyete dasydnte 

karisyfimi karisyhvas karisyamas karisye karisyavahe karisyamahe 

etc. etc. etc. etc. etc. 


934. With regard to the use or non-use of the auxiliary 
vowel i before the sibilant, there is a degree of general accord- 
ance between this tense and the other future and the desidera- 

935] THE ^-FUTURE. 301 

live ; but it is by no means absolute, nor are any definite rules 
to be laid down with regard to it (and so much the less, because 
of the infrequency of the two latter formations in actual use) : 
between this and the aorist (s-aorist on the one side, or /s-aorist 
on the other), any correspondence is still less traceable. Prac- 
tically, it is necessary to learn, as a matter of usage, how any 
given root makes these various parts of its conjugational system. 

935. Below is added a statement of the usage, as regards the auxiliary 
vowel, of the roots observed to form the s-future in the older language (more 
than a hundred and fifty : the collection is believed to be tolerably complete) 

for the most part, in the form of a specification of the roots which add 
the tense-sign directly to the root; in brackets are further mentioned the 
other roots which according to the grammarians also refuse the auxiliary 

a. Of roots ending in vowels, the great majority (excepting those in r) 
take no i. Thus, all in a (numerous, and unnecessary to specify); all in i 

i, ksi, ci, ji except p ri [and cvi] ; all in I kri, ni, bhi, ml, vU 
except ci [and dl]; all in u cyu, dru, plu, cm, dru except su 'press', 
and stu, which follow either method, as stosyami and stavisyami [and except 
ksu, ksnu, nu, yu< ru, snu]. But all in r (numerous, and unnecessary to 
specify) take i [and those in changeable r, or so-called f-roots (242) are 
said to take either i or I; no I-forms, however, are found in the older 
language] ; and likewise those in u namely bhu, dhu. 

b. Of roots ending in mutes, two thirds add sya directly. Thus, of 
roots in k, fafc; in c (all but yac : namely), muc, ric, vac, vrace, sic 
[and pac, vie]; in ch, prach (only case); in j (all but vraj : namely), 
bhaj, majj (manksya), mrj (marksya), yaj, yuj, vrj, srj [also tyaj, bhrajj, 
bhanj, ran/, sari;, svanj, nij, vij, ruj, bhuj]; in t, krt and wt [krt, crt, nrt 
optionally]; in d, ad, pad, cad, sad, skand, syand, chid, bhid, vid 'find', 
nud [also had, khid, svid, ksud, tad, and chrd and trd optionally] : only 
observed exceptions, vad, and vid 'know'; in dh, bandh, radh, budh, 
yudh, rudh [also vyadh, sadh, sidh, krudh, ksudh, c.udh, vrdh\ : only observed 
exceptions, rdh and grdh , in n, tan and man (but man forms sometimes 
manisya); in p, tap, vap, ap, gup, drp, srp, kip [also cap, ksip, Up, 
lap]: svap forms both svapsya and svapisya; in bh, yabh and labh [also 
rabh] : no exceptions observed ; in m, nam, yam, ram : kram follows 
either method. 

c. Of roots ending in semivowels, all (they are very few) take the 
auxiliary . The roots va or vi ("ve") 'weave' and hva or hu ("hve") 'call' 
take the forms vay arid hvay, as before the a of their present-stem : thus, 
vayisya, hvayisya. 

d. Of roots ending in spirants, the minority (about a third) are without 
the auxiliary vowel. They are : roots in c , vie, drc (draksya), mrc (mraksya) 
[also danp, dip, ric, Mp, kruc, rue, sprf] ; in s, cis, krs (kraksya) [also 
tvis, dvis, pis, vis, clis, tus, dus, pus, pua] ; in 8, vas 'dwell' (vatsya : 


167) [also yhas]- in h, dah, vah, mi/*, dwft, ruh [also naft, dih, lih]: 
exceptions are yrah (grahlsya) and muh. 

In the older language, a decided, though not a large, majority of simple 
roots add the sya without auxiliary *; in the classical Sanskrit, doubtless the 
contrary is the case, as i is generally taken by any root of late origin and 
derivative character as it is also uniformly in secondary conjugation 
(chap. XIV.). 

936. As the root is strengthened in forming the stem of 
this future,- so, of a root that has a strong and a weak form, 
the strong form is used: thus, from y nap or nanc is made 
nanksya ; from ]/ sras or srans, sransisya, and the like. 

Special irregularities in forming stems from certain roots have been 
noted in connection with those roots above (see majj, mrj, vay and hvay, 

The QB. has once the monstrous form a?nuvisyamahe, made upon the 
present-stem apnw (el. IV.) of j/ap. 

937. This future is comparatively rare in the oldest language in 
p;irt, apparently, because the uses of a future are to a large extent answered 
by subjunctive forms but becomes more and more common later. Thus, 
the RV. has only seventeen occurrences of personal forms, from nine different 
roots (with participles from six additional roots) ; the AV. has fifty occurrences, 
from twenty-five roots (with participles from seven more); the TS. has oc- 
currences (personal forms and participles together) from over sixty roots ; and 
(as has been noticed above) forms frem more than a hundred and fifty roots 
are quotable from the older texts. 

Modes of the s-future. 

938. Mode-forms of the future occur only sporadically. The sole Vedic 
example is karisyds, 2d sing. subj. act. (in RV., once or twice); GB. has 
esyamahai, tansy amahai, sthasy amahai, all 1st pi. subj. mid.; and bhavi- 
syadhvam, vetsyadhvam, savisyadhvam, 2d pi. impv. mid., are quoted (Bopp) 
from MBh. 

Participles of the s-future. 

939. Participles are made from the future- stem pre- 
cisely as from a present-stem in 5f a: namely, by adding 
in the active the ending rT nt, in the middle the ending qH 
mana; the accent remains upon the stem. Thus, from the 

verbs instanced above, ^iHJtt dasydnt and ^IKIHM dasyd- 
mana, Sfjf^H karisydnt and ^i^MHim karisydmana. 

According to the grammarians, the feminine of the active participle is 
made either in anil or in ati ; but only the former has been noted as occur- 

942] CONDITIONAL. 303 

ring in the older language, and the latter, if met with at all, is very rare: 
see above, 449. 

In RV. occurs once susyanti, from |/su, with anomalous accentuation. 

Preterit of the s-future: Conditional. 

940. From the future-stem is made an augment-preterit, 
by prefixing the augment and adding the secondary endings. 
in precisely the same manner as an imperfect from a pres- 
ent-stem in Ef a. This preterit is called the conditional. 

It stands related to the future, in form and meaning, as the French 
conditional aurais to the future aurai, or as the English would have to will 
have nearly as the German wiirde haben to werde fiaben. 

Thus, from the roots already instanced : 

active. middle. 

s. d. p. s. d. p. 

ddasyam ddasyava ddasyama ddasye ddasyavahi ddasyamahi 

ddasyas ddasyatam ddasyata ddasyathds ddasyetham ddasyadhvam 

s ^(Kir *(^ I HI ri IH 

ddasyat ddasyatam ddasyan ddasyata ddasyetam ddasyanta 

dkarisyam dkarisyava dkarisyama dkarisye dkarisyavahi dkansyamahi 

etc. etc. etc. etc. etc. etc. 

941. The conditional is the rarest of all the forms of the Sanskrit verb. 

The RV. has but a single example, dbharisyat, 'was going to carry off, and 

none of the Vedic texts furnishes another. In the Brahmanas it is hardly 

more common (fifteen occurrences, of which several are in repetitions of the 

same passage) except in QB-> where it is met with more than fifty times. 

Nor does it, like the future, become more frequent later: not an example 

occurs in Nala, Bhagavad-Gita, or Hitopadeca ; only one in Manu ; and two 

in Qakuntala. 

II. The Periphrastic Future. 

942. This formation contains only a single indicative 
tense, active and middle, without modes, or participle, or 

Its consists in a derivative nomen agentis. having the 




value of a future active participle, and used, either with 
or without an accompanying auxiliary, in the office of a 
verbal tense with future meaning. 

943. The noun is formed by the suffix cT tr (or cT|~ 
tar}-, and this (as in its other than verbal uses: see chap. 
XVII.) is added to the root either directly or with a preced- 
ing auxiliary x^owel ^ i, the root itself being strengthened 
by guna, but the accent resting on the suffix: thus, ^TcT 
datr from y^TT da; SficT kartr from y^fi kr ; ^rf^rT bhavitr 
from yv( bhu. 


As regards the presence or absence of the vowel z, the usage is said by 
the grammarians to be generally the same as in the s-future from the same 
root (above, 935). The most important exception is that the roots in r take 
no i: thus, kartr (against karisya); roots han and gam show the same dif- 
ference ; while i?rt, vrdh, and syand have i here, though not in the s-future. 
The few forms which occur in the older language agree with these statements. 

944. In the third persons of both voices, the nom. 
masc. of the noun, in the three numbers respectively (373), 
is used without auxiliary : thus, Hf^FTT bhavita, 'he or she or 
it will be'; HftHI|f bhavitarau, 'both will be'; H&cU(H bha- 
vitdras, 'they will be'. In the other persons, the first and 
second persons present of y'ERf as 'be' (636) are used as 
auxiliary; and they are combined, in all numbers, with the 
singular nom. masc. of the noun. As an independent verb, 
33TT as has no middle forms ; but for this auxiliary use middle 
persons have been made by analogy, ^T lie being used in 
1st sing. 

Thus, from y^J da, 'give': 



datasmi datasvas 


datahe datasvahe datasmahe 

datasi datasthas datastha datase datasathe datadhve 


data datarau daturas data datarau dataras 




Very rarely, other persons than the third are used without the auxiliary 
verb: thus, aham drasta, 'I shall see' (MBh.); tvam bhavita (MBh. Megh.), 
'thou shalt be'; and examples are not unknown of the auxiliary in the 3d 
person: thus, vakta 'sti (MBh.), 'he will speak'; and of the use in dual and 
plural of the proper number-form with the auxiliary : thus, kartarau svah 
(MBh.), 'we two shall do'. 

945. The accent in these combinations, as in all the or- 
dinary cases of collocation of a verb with a preceding predicate 
noun or adjective (592), is on the noun itself; and, unlike all 
the true verbal forms, the combination retains its accent every- 
where even in an independent clause : thus, tdrhi vri atinastro 
bhavitUsmi (QB.), 'then I shall be out of danger' (where lhavisyami, 
if used, would be accentless). Whether in a dependent clause 
the auxiliary verb would take an accent (595), and whether, if 
so, at the expense of the accent of the noun (as in the case of 
a preposition compounded with a verb-form : 1083), we are 
without the means of determining. 

940. In the Veda, the nomina agentis in tr or tar, like various other 
derivative nouns (271), but with especial frequency, are used in participial 
construction, governing the accusative if they come from roots whose verbal 
forms do so. Often, also, they are used predicatively, with or without ac- 
companying copula; yet without any implication of time; they are not the 
beginnings, but only the forerunners, of a new tense-formation. The tense- 
use begins, but rather sparingly, in the Brahmanas (from which over thirty 
occurrences are quotable), and grows more common later, though the peri- 
phrastic future is nowhere so frequent as the s-future. 

947. Middle forms are extremely few in the older language. TS. has 
once prayoktdse, which seems to be 1st sing, (the usual ending e added to 
the abbreviated root '); but TA. (i. 11) has once the later form yastdhe; 
fayitdse in QB. is clearly 2d sing.; TB. has once yastdsmahe, 1st pi. 

Uses of the Futures and Conditional. 

948. As the s-future is the commoner, so also it is the 
one more indefinitely used. It expresses in general what is go- 
ing to take place at some time to come but often, as in 
other languages, adding on the one hand an implication of will 
or intention, or on the other hand that of promise or threat- 

A few examples are: varsisydty aisdmah parjdnyo vrstiman bhavisyati 
(B.), 'it is going to rain; Parjanya is going to be rich in rain this year'; 
yds tan nd veda Mm red karisyati (RV.), 'whoever does not know that, what 
will he do with verse?' d va{ vaydm agni dhasyamaha dtha yuydrh kfrh 
karisyatha (B.), 'we are going to build the two fires; then what will you 
do?' tarn fndro 'bhyddudrava hanisydn (QB.), 'him Indra ran at, intending 
to slay'; yddy evd karisyatha sakdih devafr yajnfyaso bhavisyatha (RV.), 'if 
Whitney, Grammar. 20 


ye will do thus, ye shall be worthy of the sacrifice along with the gods'; 
ddntds te catsyanti (AV.), 'thy teeth will fall out'; na marisyasi md bibheh 
(AV.), 'thou shalt not die; be not afraid'; bruhi kva ydsyasi (MBh.), 'tell us; 
where are you going to go?' yadi mam pratydkhydsyasi visam dsthdsye (MBh.), 
'if you shall reject me, I will resort to poison'. As in other languages, the 
tense is also sometimes used for the expression of a conjecture: thus: ko 
'yam devo gandharvo vd bhavisyati (MBh.), 'who is this? he is doubtless a 
god, or a Gandharva'. 

949. The periphrastic future is defined by the grammarians 
as expressing something to be done at a definite time to come. 
And this, though but faintly traceable in later use, is a distinct 
characteristic of the formation in the language where it first 
makes its appearance (Delbrtick). It is especially often used 
along with $vds, 'tomorrow'. 

A few examples are : yatardn vd ime fvah kamitdras te jetdras (K.), 
'whichever of the two parties these shall choose tomorrow, they will conquer'; 
prdtar yastdsmahe (TB.), 'we shall sacrifice tomorrow morning'; ityahe vah 
paktdsmi (QB.), 'on such and such a day I will cook for you'; tan ma ekdm 
rdtrim dnte fayitdse jdtd u te 'yam tdrhi putrd bhavita (QB.), 'then you shall 
lie with me one night, and at that time this son of yours will be born'. 
In other cases, this definiteness of time is wanting, but an emphasis, as of 
special certainty, seems perhaps to belong to the form; thus, bibhrhf md 
pdrayisydmi tve 'U: kdsmdn md pdrayisyasi 'ty dughd imdh sdrvdh prajd 
nirvodhd, tdtas tvd pdrayitdsmi J ti (QB.), 'support me and I will save you, 
said it. From what will you save me? said he. A flood is going to carry 
off all these creatures; from that I will save you, said it'; paridevaydm cakrire 
mahac chokabhayam prdptdsmas (GB.), 'they set up a lamentation : "we are 
going to meet with great pain and dread"'; yaje l yaksi yastdhe ca (TA.), 
'I sacrifice, I have sacrificed, and I shall sacrifice'. In yet other cases, in 
the older language even, and yet more in the later, this future appears to 
be equivalent to the other : thus, prajdydm enam vijndtdsmo yadi vidvdn vd 
juhoty avidvdn vd (AB.), 'we shall know him in his children whether he is 
one that sacrifices with knowledge or without knowledge'; vaktdsmo vd idam 
devebhydh (AB.), 'we shall tell this to the gods'; yadi svdrtho mama 'pi 
bhavitd tata evam svdrtham karisydmi (MBh.), 'if later my own affair shall 
come up, then I will attend to my own affair'; katham tu bhavitdsy eka Hi 
tvdm nrpa yocimi (MBh.), 'but how will you get along alone? that, king, 
is the cause of my grief about you'. 

950. The conditional would seem to be most originally and 
properly used to signify that something 'was going to' be done. 
And this value it has in its only Vedic ocurrence, and occasion- 
ally elsewhere. But usually it has the sense ordinarily called 
"conditional"; and in the great majority of its occurrences it is 
found (like the subjunctive and the optative, when used with 
the same value) in both clauses of a conditional sentence. 


Thus, yo vrtraya sinam dtra 'bharisyat prd tdm jdnitri vidUsa uvdca 
(RV.)i 'him, who was going here to carry off Vritra's wealth, his mother pro- 
claimed to the knowing one'; fatdyum gam akarisyam (AB.), 'I was going to 
make (shonld have made) the cow live a hundred years' (in other versions 
of the same story is added the other clause, in which the conditional has a 
value more removed from its original : thus, in GB., 'if you, villain, had not 
stopped [prdgrahlsyah] my mouth'); tdta evd J sya bhaydm v\ 'ydya kdsmdd 
dhy dbhesyad dvitiydd vdf bhaydm bhavati (QB.), 'thereupon his fear departed ; 
for of whom was he to be afraid? occasion of fear arises from a second 
person'; titpapdta dram tan mene ydd vdsah paryddhdsyata (QB.), 'he leaped 
up ; he thought it long that he should put on a garment'; sd tad evd nd 
'vindat prajdpatir ydtrd 'hosyat (MS.), 'Prajapati, verily, did not then find 
where he was to (should) sacrifice'; evam cen nd 'vaksyo murdhd te vyapatisyat 
(GB.), 'if you should not speak thus, your head would fly off'; ad ydd dhdi 
^tdvad evd 'bhavisyad ydvatyo hai 'vd 'gre prajdh srstds tdvatyo hdi 'v<T 
'bhavisyan nd prd 'janisyanta (QB.), 'if he had been only so much, there 
would have been only so many living creatures as were created at first ; they 
would have had no progeny'; kirn vd 'bhavisyad arunas tamasdm vibhettd 
tarn cet sahasrakirano dhuri nd 'karisyat (Qak.), 'would the Dawn, forsooth, 
be the scatterer of the darkness, if the thousand-rayed one did not set her 
on the front of his chariot?' 



951. THOSE verbal adjectives, or participles, which are 
made from, tense-stems, and so constitute a part of the various 
tense-systems, have been already treated. It remains to describe 
certain others, which, being made directly from the root itself, 
belong to the verbal system as a whole, and not to any partic- 
ular part of it. 

The infinitive (with a few sporadic exceptions in the older 
language) also comes in all cases from the root directly, and 
not from any of the derived tense-stems. 

The same is true of the so-called gerunds, or indeclinable 

Passive Participle in td or nd. 

952. By the accented suffix rT td - - or, in a compar- 
atively small number of verbs, *f nd is formed a verbal 



adjective which, when coming from transitive verbs, quali- 
fies anything as having endured the action expressed by 
the verb : thus, ^ff dattd, 'given'; 3WT ukta, 'spoken' . Hence 
it is usually called the passive participle; or, to distinguish 
it from the participle belonging to the passive present-sys- 
tem (771), the past passive participle. 

When made from an intransitive or neuter verb, the 
same participle, as in other languages, has no passive but 
only an indefinite past sense: thus, JH gatd, 'gone'; >7rT 
bhutd, 'been'; ^faci patitd, 'fallen'. 

953. In general, this participle is made by adding rf 
td to the bare verbal root, with observation of the ordinary 
rules of euphonic combination. 

Some roots, however, require the prefixion of the auxiliary 
vowel i to the suffix. For these, and for the verbs that add 
nd instead of td, see below, 956, 957. 

As to the accent when the root is preceded by a preposi- 
tion, see 1085 a. 

954. The root before rT td has usually its weakest form, 
if there is anywhere in the verbal system a distinction of 
weak and strong forms. Thus : 

a. A penultimate nasal is dropped: e. g. akta from y anj, 
badclM from ]/ bandh, srastd from ]/ srans or sras. 

to. Roots which in the weak forms of the perfect are abbre- 
viated (794) suffer the same abbreviation here : thus, uktd from 
y vac, udhd from y vah, istd from yyaj, suptd from y svap, 
viddhd from y vyadh, prstd from ]/ prach. 

C. Final a is weakened to i in gitd from y go, 'sing', pita 
from }/pa 'drink', dhitd from y dha 'suck', sphitd, vita from 
yvya, fita from yjya, cltd from y'fya; - and it is weakened 
to i in sthitd, hitd from y dha 'put' (with dh also changed to h : 
but dhitd is found also in compounds in V.), ditd from y da 'cut' 
and y da 'bind'; sitd, mitd from y ma 'measure', citd (or catd), 
chitt (or chatd}. 

d. A final m or n is lost after a in gatd, natd, yatd, ratd 

(from y gam etc.) ; hatd, matd, ksatd, tatd, vatd (from yhan etc.). 

e. More isolated cases are: utd from yav, utd from yva 'weave', fistd 

from y^as, murtd referred to ymurch, syuta from ysiv, dyutd from ydw 

'play', mutd from ymlv, dhautd from ydhav 'cleanse 1 (RV. has also dhutd}. 


955. Of more irregular character are the following: 

a. A number of roots ending in am retain the nasal and lengthen 
the radical vowel (as in others of their verbal forms) : thus, kantd, 
krantd, tantd, canto, , crania, from y ' kram etc.; y d/ivan 'be cov- 
ered' forms in like manner dhvanta. 

b. Three roots in an make the participle from parallel roots 
in a: thus, jatd, khatd, said, from yjan etc. 

C. The root da 'give' forms datta (from the derivative form 
dad], but data also is found in composition in V. The contracted 
tta (as if for data, with the radical vowel lost) is widely found 
in composition, especially with prepositions (1087 e), but also 
with other elements : thus, devdtta (RV.); punartta (PB. vi. 5. 12); 
and, according to the grammarians, sutfa. The root jaks (deriv- 
ative of ghas: 675) foimsjagdM, as if homjagh (once apparently 
abbreviated in composition to gdha in TS. : thus, agdhad}; y svad 
makes in Veda svattd (beside svaditd). 

956. The suffix with ^ i, or in the form ^ itd, is reg- 
ularly used with the derivative verb-stems in secondary con- 
jugation (chap. XIV.), also often with roots of a derivative 
character (as flf^f jinv, f^T hins), and not infrequently with 
original roots (as ^ pat, r^ car, rpgj manth, sft pi). 

In RV. and AV., the participles in ita from simple roots are more than 
a sixth of the whole number. Among them, uditd (}/vad) is the only case 
of abbreviation of va to u. From frath comes prthita (once). Jahita from 
yha (by substitution of the present-stem as shown in jahami) is an isolated 
irregularity, payita shows the same strengthening which appears in the 
present-system (629). 

A few roots form the participle either with or without the auxiliary i : 
thus, guptd and gupitd, drptd and drpitd, dhrstd and dhrsitd, mattd and 
inaditd, vittd (also vinnd) and viditd. 

The root grabh or grah has, as elsewhere, long I: thus, grbhitd, grhltd. 

957. The suffix ^ nd (always without auxiliary ^ i) is 
taken instead of rT td by a number of roots. Thus: 

a. Certain roots in a, and in i and w-vowels : namely, pyand or find 
from j/fya, frdnd from |/pra, jlnd (beside jitd) from yjya or jl, fund from 
ycva or fvi, hand and hind from yha, dind from yda 'bind' and 'cut', 
kslnd (beside ksitd) from yksi 'destroy', plnd from ypya or pi, vllnd from 
yvll, Una from yil, dund from ydu, dyund from ydlv or dev 'lament', lund 
from yiu and some others. 

b. The roots in variable r (so-called f-roots: 242), which before the 
suffix becomes Ir or ur : thus, klrnd, glrnd, jlrnd, tlrnd, dlrnd, plrnd, stlrnd 
(beside strtd); purnd, murnd; and jurnd. 


A few participial forms in td from such roots are met with in the older 
language: thus, gurtd, purtd, flrtd. 

c. A few roots ending in j (which becomes g before the suffix: 216. 4): 
thus, bhagnd from ybhaj, bhugnd from ybhuj, magnd from ymajj, rugnd 
from yruj. Also, one or two others that show a guttural before the na: 
thus, lagnd from yiag, vrknd from ]/urapc, aknd from j/ac. 

d. A number of roots, some of them very common ones, in d (which 
becomes n before the suffix : 161, end): thus, channd, chinnd, bhinnd, vinnd 
(beside vittd and viditd), skannd, syannd, svinnd, tunnd, pannd, sannd (beside 
sattd, which alone is found in V.); and hlannd (according to the grammarians) 
from yhlad; also trnnd from ytrd and chrnnd from ychrd, which show an 
irregular lingualization of the nasal ; and others. 

958. The grammarians reckon as participles of this forma- 
tion a few miscellaneous derivative adjectives, coming from roots 
which do not make a regular participle : such are ksama, 'burnt' , 
krqa, 'emaciated', pakvd, 'ripe', phulla, 'expanded', cuska, 'dry'. 

Past Active Participle in tavant. 

959. From the past passive participle is made, by 
adding the possessive suffix SftT vant, a secondary derivative 
having the meaning and construction of a perfect active 
participle: for example, cTrT RriIM tat krtdvan, 'having done 
that'. Its inflection is like that of other derivatives made 
with this suffix (452 if.); its feminine ends in cfcft vati; its 
accent remains on the participle. 

960. Derivative words of this formation are found in RV., but without 
anything like a participial value. The AV. has a single example, with par- 
ticipial meaning: afitdvaty dtithau, 'one's guest having eaten' (loc. abs.). 
In the Brahmanas also it is extremely rare. In the later language, however, 
it comes to be quite common. And there it is almost always used predic- 
atively, and generally without copula expressed, or with the value of a 
personal verb-form in the perfect tense (like the derivative in ta in the 
future: 942 ff.). For example: mam na kafdd drstavan, 'no one has seen 
me'; sa nakularh vyapadiiavan, 'he destroyed the ichneumon'; or, with copula, 
mahat krchram praptavaty asi, 'thou hast fallen upon great misery'. Although 
originally and properly made only from transitive verbs (with an object, to 
which the participle in ta stands in the relation of an objective or factitive 
predicate), it is finally found also from intransitives : thus, cutena sam$ritavatl 
(Qak.), 'has become united with the mango-tree'; gatavatl (ib.), 'she has gone'. 

Future Passive Participles: Gerundives. 
961. Certain derivative adjectives (for the most part 

9631 GERUNDIVES . 311 

more or less clearly secondary derivatives) have acquired in 
the language a value as qualifying something which is to, 
or which ought to, suffer the action expressed by the root 
from which they come; and they are allowed to be made 
from every verb. Hence they are, like more proper par- 
ticiples, usually treated as a part of the general verbal sys- 
tem, and called future passive participles, or gerundives 
(like the Latin forms in ndus, to which they correspond in 

962. The suffixes by which such gerundives are regu- 
larly and ordinarily made are three : namely TJ ?/, rTcET tct- 
vya, and EFffal aniya. 

Derivatives in ya having this value are made in all periods of the 
language, from the earliest down; the other two are of more modern origin, 
being entirely wanting in the oldest Veda (RV.), and hardly known in the 
later. Other derivatives of a similar character, which afterward disappear 
from use, are found in the Veda. 

963. The suffix ya in its gerundive use has nothing to 
distinguish it from the same suffix as employed to make adjec- 
tives and nouns of other character (see below, chap. XVIII.: 1213). 
And it exhibits also the same variety in the treatment of the 

The original value of the suffix is io, and as such it has to be read in 
the very great majority of its Vedic occurrences. Hence the conversion of e 
and o to ay and av before it (see below). 

Thus : a. Final a becomes e before the suffix : deya, khyeya, 
met/a (probably dk-ia etc., with euphonic y interposed) : but RV. 
has once -jnaya. b. The other vowels either remain unchanged, 
or have the guna or the vrddhi strengthening ; and e usually and 
o always are treated before the ya as they would be before a 
vowel : thus, jdyya, bhdyya, layya ; ndvya, bhdvya, hdvya, bhavyd ; 
varya: and, in the later language, riiya, Jet/a, dhuya (such cases 
are wanting earlier). In a few instances, a short vowel adds t 
before the suffix : thus, itya, mitya, cnitya, stutya, krtya (the 
only Vedic examples). c. Medial a remains unchanged or is 
lengthened: thus, ddbhya, vdndya, sddya; madya, vacya. 
d. Medial i, u, and r-vowels are unchanged or have the guna- 
strengthening : thus, L idya, guhya,, dhrsya ; dvesya, y6dhya, mdrjya. 

The RV. has about forty examples of this gerundive, and the AV. adds 
half as many more. Except in bhavid (once), the accent in RV. is always 


on the root ; AV. has several cases of accent on the i of the suffix (hence 
written adya, a$ya, -vyadhya, -dharsyh}. According to the grammarians, the 
accent is on the root or else the ending is circumflexed : always the former, 
if the ya follow a vowel. 

964. The suffix tavya is a secondary adjective derivative 
from the infinitival noun in tu (below, 972), made by adding 
the suffix ya (properly ia t whence the accent ya], before which 
the final u, as usual, has ^w^a-strengthening, and is resolved 
into av. 

Hence, both as regards the form taken by the root and 
the use or omission of an auxiliary vowel i before the tavya, 
the rules are the same as for the formation of the infinitive 
(below, 968). 

No example of this formation is found in RV., and in AV. occur only 
two, janitavya and hihsitavya. In the Brahmana language it begins to be not 
rare, and is made both from the simple root and the derived conjugational 
stems (next chapter); in the classical language it is still more frequent. 
According to the grammarians, the accent of the word is either circumflex 
on the final or acute on the penult: thus, kartavya or kartdvya; in the 
accentuated texts, it is always the former. 

As to the impersonal use of this gerundive, see below, under Passive 

965. The suffix aniya is in like manner the product of sec- 
ondary derivation, made by adding the adjective suffix lya (1215) 
to a nomen actionis formed by the common suffix ana. 

It follows, then, as regards its mode of formation, the rules 
for the suffix ana (below, chap. XVIII.: 1150). 

This derivative also is unknown in RV., and in AV. is found only in 
upajivaniya and amantraniya (in both of which, moreover, its distinct 
gerundive value admits of question). In the Brahmanas (where less than a 
dozen examples of it have been noted), and in the later language, it is much 
less common than the gerundive in tavya. Its accent, as in all the deriv- 
atives with the suffix lya, is on the penult: thus, karaniya. 

966. Other formations of kindred value are found in the Veda as follows : 

a. Gerundives in tua or tva, apparently made from the infinitival noun 
in tu with the added suffix a (1209). They are kdrtua (in two occurrences 
kartva], jdntua, jetua, ndmtua, v&ktua, s6tua, sndtua, hdntua, hetua; and, 
with auxiliary i (or 5), jdnitva, sdnitva, bhdvltva. 

b. Gerundives in enia or enya (compare 1217): they are Idenia, carenia, 
dryenia, bhusenya, yudhenia, vdrenia; with one example from an apparent 
aorist-stem, yamsenya, and three or four from secondary verb-stems (see 
below, 1038). 

C. Gerundives in dyia (once dyya: compare 1218): they are daksayia, 
pandyia, viddyia, fravdyia, hnavayia ; with a few from causative secondary con- 
jugation-stems (below, chap. XIV.) : and stuseyia is of close kindred with them. 

970] INFINITIVES. 313 

d. A few adjectives in elima, as sacelima, bhidelima (apparently not 
found in use) are reckoned as gerundives by the grammarians. 

967. The division-line between participial and ordinary 
adjectives is less strictly drawn in Sanskrit than in the other 
Indo-European languages. Thus, adjectives in u, as will be seen 
later (chap. XVII. : 1178), from secondary conjugational stems, 
have participial value ; and in the Brahmanas (with an example 
or two in AV.) is found widely and commonly used a participial 
adjective formed with the suffix uka (ibid., 1180. 


968. The later language has a single infinitive, which 
is the accusative case of a verbal noun formed by the suf- 
fix rT tu, added to the root usually directly, but often also 
with aid of the preceding auxiliary vowel ^ i. The form 
of the infinitive ending, therefore, is tj^^tum or ^s^itum. 
The root has the yema-strengthening, and is accented. Thus, 
for example, "^^etum from |/^ if SficFT kdrtum from j/sfi 
Tvr ; r\[^y\^cdritum from y^TJ" car; Hi^lrlH bhdvitum from 

The rules as to the use or omission of the auxiliary i are 
the same as those that apply to the formation of the periphras- 
tic future-noun in tr or tar (943). 

The same form, in a like use, is found also in the older language, back 
to its earliest recorded period 5 but it is there only one of a whole body of 
related formations, an account of which . is in brief as follows : 

969. In the Veda and Brahmana, a number of verbal nouns, 
nomina actionis, in various of their cases, are used in construc- 
tions which assimilate them to the infinitive of other languages 
although, were it not for these other later and more devel- 
oped and pronounced infinitives, the constructions in question 
might pass as ordinary case-constructions of a somewhat pecu- 
liar kind. 

970. The nouns thus used infinitively are the following : 

a. The root-noun, without derivative suffix, is so used in 
its accusative in am, its dative in e or (from a-roots) at, its 
genitive and ablative in as, and its locative in t. 

b. The verbal noun in tu is so used in its accusative in 
turn, its dative in tave or tavai, and its ablative and genitive 
in tos. 


Of other nouns, only single cases, generally datives, are reckoned as 
used with infinitive value; thus: 

C. From the verbal noun in as, the dative in ase ; and also, 
in an extremely small number of instances, a dative in se (or 
se\ from a noun formed with s simply. 

d. From nouns in man and van, datives in mane and vane. 

e. From nouns in ti, datives in taye, or (from one or two 
verbs) in tyai. 

f. From nouns in i, datives in dye. 

g. From nouns in dhi and si, datives in dhyai and syai. 
h. A few infinitives in sani are perhaps locatives from nouns 

in an added to a root increased by s. 

i. From a single root, dhr, are made infinitively used forms 
in tdri, of which the grammatical character is questionable. 

Among all these, the (forms which have best right to special treatment 
as infinitives, on account of being of peculiar formation, or from suffixes not 
found in other uses, or both, are those in se, sani, tari, dhyai, and tavdi. 

Except the various cases of the derivative in tu, and of the root-noun, 
these infinitives are almost wholly unknown outside the Rig- Veda. 

Other suffixes and forms than those noticed above might be added ; for 
it is impossible to draw any fixed line between the uses classed as infinitive 
and the ordinary case-uses ; and the so-called infinitives are found coordinated 
in the same sentence with common nouns, and even with compound nouns. 

More special rules as to the various formations are as follows: 

971. The root-noun used as infinitive has the same form, and the same 
accent, both when simple and when combined with prepositions, as in its 
other uses. In the very great majority of instances, it is made from roots 
ending in a consonant; but also from a few in d (khyd, da, dhd, pa?, ma, 
yd], from two or three in i and u- vowels (hi, ml, bhu), and from one or two 
in changeable r, which takes the zr-form (tir, stir). 

The roots in d form the accus. in dm (pratidhdm, AV.), the dat. in di, 
the abl. in as (understanding avasd before d as for avasds and not avasdt 
in RV. iii. 53. 20), and the locative in e (only two examples, of which one 
is better understood as dative). 

972. The infinitive noun in tu is made freely from roots of every form. 
The root takes the ywna-strengthening, if capable of it, and often adds the 
auxiliary vowel i before the suffix (according to the rule already stated, 968). 
The root is accented, unless the noun be combined with a preposition, in 
which case the latter has the accent instead : thus, kdrtum, etave, hdntos, 
but nfkartum, nfretave, nfrhantos. 

The dative in tavdi is in two respects anomalous : in having the heavy 
feminine ending di along with a strengthened u; and in taking a double 
accent, one on the root or on the prefixed preposition, and the other on the 
ending di : thus, etavdf, hdntavdf, dtyetavdi, dpabhartavdf. 

The root grah makes (as in other kindred formations) grdhltu , and long i 
is shown also by ?dritu, stdrltu, hdvltu (and compare bhdvitva, 966 a). 

980] INFINITIVES. 315 

973. The infinitive in ase is made in RV. from about twenty-five roots ; 
in AV. and later there have been noted no other examples of it. In near 
three quarters of the cases, the accent is on the suffix : thus, rnjdse, jlvdse, 
bhiydse, tujdse ; the exceptions are cdksase; dhdyase (with y inserted before 
the suffix); and dyase, bhdrase, spdrase, hdrase (with ^Una-strengthening of 
the root). Strengthening of the root is also shown by javdse, dohdse, bhojdse, 
$obhdse. In pusydse is seen, apparently, the present-stem instead of the root. 

The ending se is extremely rare, being found only in jise and perhaps 
state, and one or two still more doubtful cases. 

974. Infinitives in mane are made from only five roots: thus, trdmane, 
damane, dhdrmane, bhdrmane, and (with different accent) vidmdne. From 
yda comes davdne; turvdne may come directly from ytr, or through the 
secondary root turv ; dhurvane is rather from ydhurv than from ydhvr. 

975. The infinitives in taye are istdye (j/j's), pltdye (ypa 'drink'), vitdye, 
sdtdye. In tydi, the only examples noted are itydf (RV.) and sadhyai (AB.). 

With aye are formed tujdye, drfdye, mahdye, yudhdye, sandye. 

976. The ending dhyai is, more than any other, irregular and various 
in its treatment. It has always an a before it ; and in the majority of cases 
it is accented upon this a, and added to a weak form of root: thus, fwc- 
ddhyai, prnddhyai, dhiyddhyai, huvddhyai. But the form of root is the strong 
one in a few cases : namely, $ayddhyai, stavddhyai, tarddhyai, jarddhyai, 
mandddhydi, vandddhyai. In half-a-dozen forms, again, the root has the 
accent : namely, ksdradhyai, gdmadhyai, ydjadhyai (but once or twice also 
yajddhyai], vdhadhyai, sdhadhyai, bhdradhyai. In a single instance, ptbadhyai, 
the suffix is added distinctly to a present-stem ; and in one, vdvrdhddhydi, 
to a perfect stem. Finally, in a number of instances (ten), this infinitive 
is made from a causative stem in ay : thus, madayddhyai, risayddhyai, etc. 

This infinitive is by no means rare in RV., being made in thirty-five 
different forms (with seventy-two occurrences). But it is hardly known out- 
side of the RV.; the AV. has it but once (in a passage found also in RV.); 
and in the branches of the Yajur-Veda but two or three examples have been 
noticed (one of them TS. falsely reads gdmadhye)-, in the Brahmana language 
it appears to be entirely wanting. 

977. An example or two are met with of an infinitive in syai: thus, 
rohisyai (TS.), avyathisyai (K.). 

978. The infinitives in sani are: -bhusdni from ybhu; fusdni from y?u 
or ?va; nesdni from ynl; saksdni from ysah; parsdni from }/pr, tarlsdni 
from ytr; and grnudni and -strnlsdni from yygr and str the last con- 
taining evident present tense-signs (compare the 1st sing, grnise, 894 d). 

979. The only infinitive in tari is dhartdri (with its compound vidhartdri), 
from ydhr. 

Uses of the Infinitives. 

980. The uses of the so-called infinitives are for the most 
part closely accordant with those of the corresponding cases from, 
other abstract nouns. Thus : 


981. The accusative, which is made only from the root- 
noun and the noun in tu, is used as object of a verb. 

Especially, of forms from the roots pafc, 'be able', and arh, 'be worthy, 
have the right or the power'. Thus, fakema tvd samidham (RV.), 'may we 
accomplish thy kindling'; md pafccm pratidhdm {sum (AV.), 'may they not be 
able to fit the arrow to the string'; mdno vd imdm sadydh pdrydptum arhati 
mdnah pdribhavitum (TS.)> 'the mind, forsooth, can at once attain and surpass 
her'; ko hy etdsyd 'rhati gtihyarh ndma grdhltum (QB.), 'for who is worthy to 
take his secret name?' In the Veda, the construction with these verbs is 
only one among others ; in the Brahmana, it becomes the greatly prevalent 
one (three quarters or more of all the cases). 

Further, of verbs of motion (next most frequent case): thus, ddksindni 
h6tum eti (TS.), 'he goes to sacrifice things pertaining to sacrificial gifts'; 
fndram pratfram emy dyuh (RV.), 'I go to Indra for (i. e. beseech of him) 
the lengthening out of life'; of ydhr, 'persist in, undertake': as, sd iddrh 
jdtdh $drvam evd ddgdhum dadhre (QB.), 'he, as soon as born, began to burn 
this universe'; of verbs meaning 'desire, hope, notice, know', and the 
like: as, pdfdn vicrtarh vettha sdrvdn (AV.), 'thou knowest how to loosen 
all bonds'; tdsmdd agnfth nd "driyeta pdrihantum (QB.), 'therefore one should 
not be careful to smother the fire'; and of others, 

982. Of the infinitive datives, the fundamental and usual 
sense is that expressed by 'for, in order to, for the purpose of. 

Examples are ; : vfyvam jlvdrh cardse bodhdyanti (RV.), 'awakening every 
living creature to motion'; tdn tipa ydta ptbadhydi (RV.), 'come to drink 
them'; ndt 'tdm te devd adadur dttave (AV.), 'the gods did not give her to 
thee for eating'; prat "d yudhdye ddsyum tndrah (RV.), 'Indra went forward 
to fight the demon'; cdksur no dhehi vikhydt. (RV.), 'give us sight for looking 

Some peculiar constructions, however, grow out of this use of the in- 
finitive dative. Thus: 

a. The noun which is logically the subject or the object of the action 
expressed by the infinitive is frequently put beside it in the dative (by a 
construction which is in part a perfectly simple one, but which is stretched 
beyond its natural boundaries by a kind of attraction): thus, cakdra surydya 
pdnthdm dnvetavd u (RV.), 'he made a track for the sun to follow (made 
for the sun a track for his following)'; pfplie pfnge rdksobhyo vinfkse (RV.), 
'he whets his horns to pierce the demons'; rudrdya dhdnur d tanomi brah- 
madvtse fdrave hdntavd u (RV.), '1 stretch the bow for Rudra, that with his 
arrow he may slay the 6raftma-hater'; asmdbhyam dr$dye surydya punar 
ddtdm dsum, 'may they grant life again, that we may see the sun'. 

b. An infinitive with |/fcr, 'make', is used nearly in the sense of a 
causative verb : thus, prd 'ndhdrh frondm cdksasa etave krthah (RV.), 'ye 
make the blind and lame to see and go'; agnfih samfdhe cakdrtha (RV.), 
; thou hast made the fire to be kindled'. Of similar, character is an occasional 




construction with another verb: as, ydd im upmdsi kdrtave karat tat (RV.), 
'what we wish to be done, may he do that'. 

c. A dative infinitive is not seldom used as a predicate, sometimes with, 
but more usually without, a copula expressed : thus, agnfo iva nd pratidhfse 
bhavati (TS.), 'like fire, he is not to be resisted'; mahimd te anyena nd 
9amnd?e (VS.), 'thy greatness is not to be attained by another'; ndldm indro 
nikartave nd fakrdh pdri?aktave (RV.), 'Indra is not to be put down, the 
mighty one is not to be overpowered 1 . 

d. Sometimes an infinitive so used without a copula has pretty clearly 
the value of an imperative : thus, tyd me yafdsa . . . aufijd huvddhyai [asti] 
(RV.), 'these glorious ones shall the son of Uc.ij invoke for me'; suktebhir 
vah . . . indra nv agni dvase huvddhyai [stah] (RV.), 'with your hymns shall 
ye call now on Indra and Agni for aid'; vandddhya agnim ndmobhih [asmi] 
(RV.), 'let me greet Agni with homage'; asmdkasaf ca surdyo vfyva dfds 
tarlsdni (RV.), 'and let our sacriflcers cross all regions'. The infinitives in 
dhyai and sani (which latter is in all its uses accordant with datives) are 
those in which the imperative value is most distinctly to be recognized. 

e. In the B. (with only a sporadic case or two elsewhere) the dative 
in tavai is frequently used with a verb signifying 'speak' (bru, vac, aft), to 
express the ordering of anything to be done : thus, tdsmad 6sadhlnam evd 
mulany ucchettavaf bruydt, 'therefore let him direct the roots of the plants 
to be cut up (speak in order to cutting up)'. 

983. The ablative infinitive which, like the accusative, 
is made only from the root-noun and that in tu is found 
especially with the prepositions ft, 'until', and purh,, 'before'. 

Thus, d tdmitos (TS. etc.), 'until exhaustion'; purd vacdh prdvaditos 
(TS.), 'before utterance of the voice'. In the Brahmana language, this is 
the well-nigh exclusive construction of the ablative; in the Veda, the latter 
is used also after rte, 'without', and after several verbs, as tra and pa, 
I/M, bhi. 

In two or three instances, by an attraction similar to that illustrated 
above for the dative (982 a), a noun dependent on this infinitive is put in 
the ablative beside it : thus, pura vagbhyah sampravaditoh (PB.), 'before the 
utterance together of the voices'; trddhvam kartdd avapddah (RV.), 'save us 
from falling down into the pit'. 

984. The genitive infinitive (having the same form as the 
ablative) is in common use in the Brahmana language as depend- 
ent on icvard, lord, master', employed adjectively in the sense 
of 'capable' or 'likely' or 'exposed to'. 

Examples are : td [devdtah] ifvard enam praddhah (TS.), 'they are likely 
to burn him up'; dtha ha vd lyvarb '^nfrh citvd Mrhcid damritdm apattor vf 
va hvdlitoh (QB.), 'so in truth he is liable, after piling the fire, to meet with 
some mishap or other, or to stagger'; Ifvararh vai rathantaram udgatuq, caksuh 
pramathitoh (PB.), 'the rathantara is liable to knock out the eye of the 


The dative is once used in B. instead of the genitive (in ifvarati jd- 
nayitavaf); and, in the later language, sometimes the accusative in turn. 
Occasionally the masc. sing. nom. Ifvarah is used, without regard to the gen- 
der or number of the word which it qualifies : thus, tdsye "fvardh prajd 
pdplyasl bhdvitoh (QB.), 'his progeny is liable to deteriorate 1 . And in a few 
instances the word Ifvara is omitted, and the genitive has the same value 
without it: thus, dve madhyandinam abhi pratyetos (AB.), 'two may be added 
to the noon libation'; tdto diksitdh pamano bhdvitoh (B.), 'then the conse- 
crated is liable to get the itch'. 

This construction with ifvara, which is the only one for the genitive 
infinitive in the Brahmana, is unknown in the Veda, where the genitive is 
found in a very small number of examples with madhyd, and with the root 
If: thus, madhyd kdrtoh (RV.), 'in the midst of action'; ife ray6 ddtoh (RV.), 
'he is master of the giving of wealth'. 

985. Unless the infinitives in sani and tari are locative in form (their 
uses are those of datives), the locative infinitive is so rare, and has so little 
that is peculiar in its use, that it is hardly worth making any account of. 
An example is usdso budhf (RV.), 'at the awakening of the dawn'. 

986. In the Veda, the dative infinitive forms are very much 
more numerous than the accusative (in RV., their occurrences 
are twelve times as many; in AV., more than three times) ; and 
the accusative in turn is rare (only four forms in RV., only 
eight in AV.). In the Brahmanas, the accusative has risen to 
much greater comparative frequency (its forms are nearly twice 
as many as those of the dative) ; but the ablative-genitive, which 
is rare in the Veda, has also come to full equality with it. The 
complete disappearance in the classical language of all except- 
ing the accusative in turn is a matter for no small surprise. 

987. The later infinitive in turn is oftenest used in con- 
structions corresponding to those of the earlier accusative : thus, 
na vaspam acakat sod/ium, 'he could not restrain his tears'; tarn 
drastum arhasi, 'thou oughtest to see it'; praptum ichanti, 'they 
desire to obtain'; samkhyatum arabdham, 'having begun to count'. 
But also, not infrequently, in those of the other cases. So, 
especially, of the dative : thus, avasthatum sthanantaram cintaya, 
'devise another place to stay in'; tvam anvestum iha "gatah, 'he 
has come hither to seek for thee'; but likewise of the geni- 
tive : thus, samartho gantum, 'capable of going'; samdhatum ic- 
varah, 'able to mend'. Even a construction as nominative is 
not unknown : thus, yuktam tasya may a samaqvasayitum bha- 
ryain (MBh.), 'it is proper for me to comfort his wife'; na 
naptaram svayam nyayyam captum evam (R.), 'it is not suitable 
thus to curse one's own grandson'. 

988. In the later language, as in the earlier, the infinitive in certain 
connections has what we look upon as a passive value. Thus, kartum arabdhah, 

991] GERUNDS. 319 

'begun to be made': frotum na yujyate, 'it is not fit to be heard (for hear- 
ing)'. This is especially frequent along with the passive forms of y$ak: thus, 
tyaktum na $akyate, 'it cannot be abandoned'; fakyav iha "netum, 'they two 
can be brought hither'; na ca vibhutayah ?akyam avaptum urjitah, 'nor are 
mighty successes a thing capable of being attained'. 


989. The so-called gerund is a stereotyped case (doubt- 
less instrumental) of a verbal noun, used generally, but in 
the later language not exclusively, as logical adjunct to the 
subject of a clause, denoting an accompanying or (usually) 
a preceding action to that signified by the verb of the clause. 
It has thus the virtual value of an indeclinable participle, 
present or past, qualifying the actor whose action it de- 
scribes : 

Thus, for example : grutvai 'va ca 'bruvan, 'and hearing (or 
having heard) they spoke'; tebhyah pratijnaya 'thai 'tan paripa- 
pracha, 'having given them his promise, he then questioned 
them' . 

990. The gerund is made in the later language by one 
of the two suffixes ^T tva and T y<i. the former being used 
with a simple root, the latter with one that is compound- 
ed with a prepositional prefix or, rarely, with an ele- 
ment of another kind, as adverb or noun. 

Exceptions to this distribution of uses between the two suffixes are very 
rare : examples of simple roots with ya are arcya, grhya, usya (yvas 'dwell') ; 
of compounded roots with tva are anudhyatva, apatyaktva, pratyarpayitvd 
(AV.: only case noticed in the Veda: TA. has -rocayitva). The gerund in 
tvd, however, may have the negative particle prefixed to it: thus, akrtva, 

Of compounds of the gerund in ya with other elements than the usual 
verbal prefixes, RV. has punarddya, karnagfhya, padagfhya, hastagrhya, 
ararhkrtya, akkhallkftya, mithaspfdhya ; AV. has further namaskrtya. 

991. The suffix ^T tva has the accent. It is usually 
added directly to the root, but sometimes with interposition 
of the auxiliary vowel ^ i with regard to which, as well 
as to the form of the root before it, this formation closely 
agrees with that of the participle in rT ta (above, 952 if.). 


When i is used, the disposition to take a weak form of root is less 
marked. Roots which have na instead of ta as participial suffix usually 
reject the i. 

992. The suffix 7J ya is added directly to the root, 
which is accented, but has its weak form. A root ending 
in a short vowel takes rT tya instead of TJ ya: thus, ftfrET 
-jitya, TOT -krtya. 

Roots in am and an whose passive participle ends in ata (954 d) form 
this ! gerund also in atya: thus, -gatya, -hatya. But such am-roots are 
allowed in the later language to preserve their nasal in the gerund: thus, 
-gamya (no such form occurs in the Veda). Final changeable r becomes Ir 
or ur : thus, -glrya, -purya. Final a remains unaltered : thus, -gay a, -sthaya ; 
and mi 'establish' and mi 'diminish' take the form ma; II 'cling' is allowed 
to do the same. 

993. The older language has the same two gerund forma- 
tions, having the same distinction, and used in the same way. 

a. In RV., however, the final of ya is in the great majority of in- 
stances (fully two thirds) long (as if the instrumental ending of a derivative 
noun in i or ti). In AV., long a appears only once, in a RV. passage. 

b. Instead of tvd alone, the Veda has three forms of the suffix, namely 
tvd, tvdya, and tvi. Of these three, tvi is decidedly the commonest in RV. 
(thirty-five occurrences, against twenty-one of toa); but it is unknown in 
AV., and very rare elsewhere in the older language; tvaya is found nine 
times in RV. (only once outside the tenth Book), twice in AV., and but 
few times elsewhere. The historical relation of the three forms is obscure. 

c. Two other gerund suffixes, tvanam and tvinam, are mentioned by 
the grammarians as of Vedic use, but they have nowhere been found actually 
to occur. 

994. The use of this gerund, through not changing in its 
character, becomes much more frequent, and even excessive, in 
the later language. 

Thus, in the Nala and Bhagavad-Gita, which have only one tenth as 
many verb-forms as RV., there are more than three times as many examples 
of the gerund as in the latter. 

Early examples are : vdjrena hatvd nfr apdh sasarja (RV.), 'striking ^vith 
his thunderbolt, be poured forth the waters'; strfyam drstvdya kitavdm tatapa 
(RV.), 'the gambler is distressed when he sees a woman'; pitvi sdmasya 
vavrdhe (RV.), 'having drunk of the soma, he waxed strong'. In the older 
language almost without exception, and in the later usually, it expresses an 
action or condition belonging to the subject of the sentence; but it is in some 
texts more loosely construed : thus, tatah cabdad abhijnaya sa vyaghrena 
hatah (H.), 'thereupon he was slain by the tiger, the latter having recognized 
him by his noise'; kirn nu me sydd idam krtvct (MBh.), 'what, I wonder, 

996] GERUNDS. 321 

would happen to me, having done this?' sucintya co J ktarh suvicdrya yat krtam 
(H. , -what is said after mature thought, and done after full deliberation'. 

Adverbial Gerund in am. 

995. The accusative of a derivative nomen actionis in a, used 
adverbially, assumes sometimes a value and construction so 
accordant with that of the usual gerund that it cannot well be 
called by a different name. 

No example of a peculiar gerundial construction with such a form occurs 
either in RV. or AV., although a few adverbial accusatives are probably to 
be classed as representing this formation : thus, abhydkrdmam, pratdnkam, 
pranodam, nilayam, dbhiskdndam. The gerund is found especially in the 
Brahmanas (much oftenest in QB.), and sparingly later. In the classical 
language it is quite rare. 

A final vowel has frdd/u'-strengthening before the suffix; final a adds y; 
a medial vowel has guna; but medial a is usually lengthened. The accent 
is on the radical syllable. 

Examples are: kamam va imdny dngdni vyatyasam fete (QB.), 'he lies 
changing the position of these limbs at pleasure'; uttardm-uttardrh fdkhdrh 
samdldmbharh rohet (QB.), 'he would climb, taking hold of a higher and ever 
a higher limb'; apansu mahdnagdm iva 'bhisarhsaram didrksitdrah (QB.), 
'hereafter, running together as it were about a great snake, they will wish 
to see him'; ndmdny dsdm etdni ndmagrdham (QB.), 'with separate naming 
of these their names'; yd viparydsam avaguhati (QB.), 'whoever buries it 
upside down'. As in these examples, the form is almost always a compound 
one. In the later language, it it said to be used most often repeated : thus, 
pdyam-pdyam vrajati, 'he goes after drinking repeatedly'; prathamam bhojarh 
vrajati, 'having first eaten, he goes'; bdhutkseparh krandUum pravrttd (Qak.), 
'she proceeded to cry, throwing up her arms (with arm-tossing)'. 



996. SECONDARY conjugations are those in which a whole 
system of forms, like that already described as made from 
the simple root, is made, with greater or less completeness, 
from a derivative conjugation-stem ; and is also usually con- 
nected with a certain definite modification of the original 
radical sense. 

Whitney, Grammar. 21 


We have seen, indeed, that the tense-systems are also for the most part 
made from derivative-stems ; and even that, in some cases, such stems assume 
the appearance and value of roots, and are made the basis of a complete 
coiijugational system. Nor is there any distinct division-line to be drawn 
between tense-systems and derivative conjugations the latter are present- 
systems which have been expanded into conjugations by the addition of other 
tenses, and of participles, infinitives, and so on. In the earliest language, 
their forms outside of the present-system are still quite rare, hardly more 
than sporadic; and even later they are with the exception of one or two 
formations which attain a comparative frequency much less common than 
the corresponding forms of primary conjugation. 

997. The secondary conjugations are: I. Passive; 
II. Intensive ; III. Desiderative ; IV. Causative ; V. Denom- 

The passive is classed here rather as a matter of convenience and of 
general usage than because it is of the same kind with the others. 

I. Passive. 

998. The passive conjugation has been already in the 
main described. Thus, we have seen that: 

a. It has a special present-system, the stem of which 
is present only, and not made the basis of any of the re- 
maining forms : this stem is formed with the accented class- 
sign IT yd*-, and it takes the middle endings. It is treated 
above, 768 ff. 

b. In the other tenses, the middle forms are used also 
in a passive sense. But: 

c. There is a special passive 3d sing, of the aorist, 
ending in ^ i: it is treated above, 842 ff. And: 

d. According to the grammarians, there may be formed 
from some verbs, for passive use, a special stem for the aorist 
and the two future systems, coinciding in form with the pecu- 
liar 3d sing, aorist. , 

Thus: from yda (aor. 3d sing, adayi], beside ddasi, dasye, datdhe, also 
adayisi, dayisye, ddyitdhe. The permission to make this double formation 
extends to. all roots ending in vowels, and to grah, drp, and han. The 
duplicate forms have not been noticed in the older language, and they are, 
at the best, extremely rare in the later. 

As to the prescribed passive inflection of the periphrastic perfect, see 
below, 1072. 


e. Besides the participle from the present tense-stem 
(771. 5), the passive has a past participle in rT ta (952). or ^ 
na (957), and future participles, or gerundives, of various 
formation (961 ff.). made directly from the root. 

999. The passive construction, with the logical subject in the instru- 
mental case, is a frequent and favorite one, especially in the later language : 
thus, evam uktva tena sarvesam bandhanani chittani (H.), 'thus saying, he 
cut the bonds of them all'. And, extremely frequently, an impersonal passive 
in the third person is used ; and it may (as in other languages) be formed 
from intransitive as well as transitive verbs : thus, cruyatam, 'let it be heard' 
(i. e. hear ye!), iha "gamyatam, 'come hither'; sarvair jalam adayo 'ddiyatam 
(H.), 'let all fly up, taking the net with them'; tac chrutva jaradgaveno 
'ktam, 'hearing that, Jaradgava said'; katharh jlvitavyam, 'how is one to live?' 
yavad munina sthatavyam (H.), 'as long as that sage shall exist'. The 
predicate to the instrumental subject of such a construction is, of course, 
also in the instrumental : thus, adhuna tava J nucarena maya sarvatha bhavi- 
tavyam (H.), 'henceforth I shall always be thy companion'; tena tvaya ya- 
vajjlvarh sukhina bhavitavyam (H.), 'with that thou shalt be happy as long as 
thou livest'. The gerundive is common in this construction, and not seldom 
it has a purely future sense. 

II. Intensive. 

1000. The intensive (sometimes also called frequent- 
ative) is that one of the secondary conjugations which is 
least removed from the analogy of formations already de- 
scribed. It is, like the present-system of the second con- 
jugation-class (642 ff.), the inflection of a reduplicated stem, 
but of one that is peculiar in having a strengthened redu- 
plication. It is decidedly less extended beyond the limits 
of a present-system than any other of the derivative con- 

The intensive conjugation signifies the repetition or 
intensification of the action expressed by the primary con- 
jugation of a root. 

1001. According to the grammarians, the intensive con- 
jugation may be formed from nearly all the roots in the 
language the exceptions being: roots of more than one 



syllable, those conjugated only causatively (below, 1056\ 
and in general those beginning with a vowel. 

In fact, however, intensives in the later language are extremely rare, 
so rare that it is hard to tell precisely what value is to be given to the 
rules of the native grammar respecting them. Nor are they at all common 
earlier, except (comparatively) in the RV. , which contains about six sevenths 
of the whole number (rather over a hundred) quotable from Veda and Brah- 
mana-texts (AV. has less than half as many as RV., and many of these in 
RV. passages). 

Hence, in the description to be given below, the actual aspect of the 
formation, as exhibited in the older language, will be had primarily and 
especially in view; and the examples will be of forms found there in use. 

1002. The strong intensive reduplication is made in three 
different ways: 

a. The reduplicating syllable is, as elsewhere, composed of 
a single consonant with following vowel, and, so far as the 
consonant is concerned, follows the rules for present and per- 
fect reduplication (590) ; but the vowel is a heavy one, radic- 
al a and r (or ar] being reduplicated with a, an /-vowel by e, 
and an w-vowel by o. 

Examples are : vavad, babadh, cacvas, rarandh; dadr, dadhr ; cekit, tetij, 
neni, vevli; cocuc, popruth, cosku, johu. 

b. The reduplicating syllable has a final consonant, taken 
from the end of the root. With an exception or two, this con- 
sonant is either r (or its substitute I) or a nasal. 

Examples are: carcar, calcal, sarsr, marmrj, jarhrs ; cankram, janghan, 
tahstan, danda$ (ydah? or da(), janjabh (yjambh or jabh), tantas (ytans 
or tas), nannam (ynam). 

Only roots having a or r as vowel make this form of reduplication, but 
with such roots it is more common than either of the other forms. 

Irregular formations of this class are : with a final other than r or n in 
the reduplication, badbadh; with a final nasal in the reduplication which is 
not found in the root, jangah (RV.), janjap (B.: and the later language 
has dandah); with an anomalous initial consonant in reduplication, jarbhur 
from ybhr (compare the Vedic perfect jabhdra, 789 b) ; with various treatment 
of an r or ar-element, dardar and dardir, carkar and carkir, tartar and tartur, 
carcar and carcur, jargur and jalgul and galgul. 

The root r is the only one with vowel initial forming an intensive stem 
in the older language : it makes the irregular alar or air. 

Q. The reduplication is dissyllabic, an /-vowel being added 
after a final consonant of the reduplicating syllable. This /-vowel 
is in the older language short before a double consonant, and 
long before a single. 

Examples are: ganiyam (but ganigmatam}, varivrt, vamvah, caniskad, 

1005] INTENSIVE. 325 

sanisvan; navlnu, davidyut (and the participles ddvidhvat but tdvituat). 
A single exception as to the quantity of the i is davidhdva. 

This method of reduplication is followed in the older language by over 
twenty roots. Thus, of roots having final or penultimate n (once m], and n 
in the reduplicating syllable, pan, phan, san, suem, han; gam; krand, fcand, 
skand, syand; of roots having final or medial r, and r in the reduplicating 
syllable, kr 'make', tr, bhr, vr, mrd, vrj, vrt ; further, of roots assuming 
in the reduplication a n not found in the root, only vah (QB.: the grammarians 
allow also kas, pat, pad}-, finally, of roots having u or u as radical vowel, 
with, av before the i-vowel, tu, dhu, nu, dyut. 

In this class, the general rules as to the form of the reduplicating con- 
sonant (590) are violated in the case of ghanlghan and bharlbhr, and of 
ganlgam, karlkr (but the regular carlkr also occurs), kanikrand, and kaniskand 
(but also caniskand occurs). 

The reversion to more original guttural form after the reduplication in 
cekit, and janghan and ghanlghan, is in accordance with what takes place 
elsewhere (216.9). 

1003. The same root is allowed to form its intensive stem 
in more than one way. 

Thus, in the older language, dddr and dardr ; dadhr and dardhr ; cdcal 
and carcar (and carcur) ; tartar (and tartur] and tantr ; janghan and ghanlghan ; 
varvrt and varlvrt; jarbhur and bharibhr; dodhu and davldhu; nonu and 
navlnu; bdbadh and badbadh. 

1004. The model of normal intensive inflection is the 
present-system of the reduplicating conjugation-class (II.); 
and this is indeed to a considerable extent followed, in 
respect to endings, strengthening of stem, and accent. But 
deviations from the model are not rare ; and the forms are 
in general of too infrequent occurrence to allow of satis- 
factory classification and explanation. 

The most marked irregularity is the frequent insertion 
of an ^ * between the stem and ending. According to the 
grammarians, this is allowed in all the strong forms before 
an ending beginning with a consonant; and before the 
*i e a final vowel has t/wwa-strengthening, but a medial one 
remains unchanged. 

Present System. 

1005. We will take up the parts of the present-system in 
their order, giving first what is recognized as regular in the 


later language, and then showing how the formation appears in 
the earlier texts. As most grammarians do not allow a middle 
inflection, and middle forms are few even in the Veda, no at- 
tempt will be made to set up a paradigm for the middle voice. 

1006. As example of inflection may be taken the root 
f^ vidj of which the intensive stem is cjf^ vevid, or, in 
strong forms, ifi^ veved. 

Neither from this nor any other root are more than a few scattering 
forms actually quotable. 

1. Present Indicative. 

s. d. . 

vevedmi, vevidlmi vevidvds vevidmds 


vevetsi, v6vidlsi vevitthds vevitthd 

3 ^fo|Tfr, ^|cj<{Jiri cfNTlH il<=^irj 
vevetti, vevidlti vevittds vevidati 

From y~^ ku, the singular forms with auxiliary vowel 
would be suc^ojiJZj jdhavlmi, sii^cum johavlsi, sii^<Jiiri jo- 

1007. The forms found in the older language agree in general with 
the paradigm. Examples are: 1st sing., carkarmi, vevesmi; 2d sing., alarsi, 
ddrdarsi; 3d sing., dlarti, veveti, nenekti, janghanti, kdnikrantti, ganlgamti; 
3d du., jarbhrtds ; 1st pi., nonumas; 3d pi., ndnadati, bharibhrati, vdrvrtati, 
ddvidyutati, nenijati; and, with the auxiliary vowel, j6havimi, cdka&mi; 
cdka$iti, nonavlti, dardariti, jarbhuriti. No stem with dissyllabic reduplication 
takes the auxiliary i in any of its forms. AV. has jdgrdti, with irregular 

A single dual form with I and strong stem occurs: namely, tartarithas. 

The middle forms found to occur are: 1st sing., jdguve, nenije; 3d sing., 
nenikte, sarsrte; and, with irregular accent, tetikte, dediste; with irregular 
union-vowel, ndnnate; with ending e instead of te, jdngahe, jdguve, yoyuve, 
bdbadhe, and (with irregular accent) badbadhe; 3d du., sarsrate ; 3d pi., 

2. Present Subjunctive. 

1008. Subjunctive forms with primary endings are extremely rare: 
there have been noticed only janghdndni, jdgardsi (AV.); and, in the middle, 
tantasatte (3d du.). 

Forms with secondary endings are more frequent: thus, 2d sing., jan- 

1011] INTENSIVE. 327 

ghanas, jalgulas ; 3d sing., jdgarat, cdrkrsat, jdnghanat, bdrbrhat, mdrmrjat, 
mdrmr?at, parpharat, dardirat, caniskadat, davidyutat, sanisvanat , 1st du. ; 
janghanava; 1st pi., carkirama, vevidama; 3d pi., pdpatan, tfcucan, carkiran; 
and, with double mode-sign, cdkafdn (AV.). Besides these, rardnas and 
rardnat, cdkdnas and cakdnat and cakdnama, which their accent assimilates 
rather to perfect mode-forms with long reduplication, like mamdhas and 
sasdhat etc. (8 10 a). Of the middle are found only 3d persons plural: thus, 
jdnghananta, jarhrsanta, marmrjanta, nonuvanta, fOfucanta ; and cakdnanta 
(and cakananta once). 

3. Present Optative. 

1009. This mode would show the unstrengthened stem, 
with the usual endings (566), accented. Thus: 

d. _p. 

vevidyftm vevidy&va vevidyfima 

etc. etc. etc. 

The optative is represented by only an example or two in the older 
language: thus, active, vevisyat (AV.), jagryat (AB.); RV. has only cakanyat 
(pft.?); middle, nenijita (K.). 

4. Present Imperative. 

1010. The regular forms of the imperative, including 
the usual subjunctive first persons, would be as follows : 

vevidani vevidava vevidama 

2 5!=l!i cjftrlH^ ^f^ff 

veviddhi . vevittdm vevitta 

vevettu, veviditu vevittam vevidatu 

1011. Older imperative forms are less rare than optative. The first 
persons have been given above (janghdnani, the only accented example, does 
not correspond with the model, but is in conformity with the subjunctive of 
the reduplicating present) ; the proper imperatives are : 2d sing. , dadrhf, 
dardrhi, carkrdhi, jagrhi, nenigdhi, and raranddhf, rarandhf, cakandhi, 
vavandhf; the ending tat is found in carkrtat and jagrtat; and the latter 
(as was pointed out above, 570) is used in AV. as first person sing. ; 
barbrhi shows an elsewhere unparalleled loss of h before the ending hi; 
3d sing., vevestu, dardartu, marmarttu, and rarantu; 2d du., jagrtam; 
3d du., jagrtam; 2d pi., jagrtd, and raranta; cankramata (RV., once) has 
an anomalous union-vowel; 3d pi., only the anomalous cakantu (RV., once) r 
apparently for cakanatu. In the middle voice is found only neniksva (QB.). 


Of imperative forms with auxiliary I, RV. has none; AV. has vavaditu 
and johavltu, and such are sometimes found in the Brahmanas ; AV. has 
also, against rule, tanstanihi and janghanihi. 

5. Present Participle. 

1012. The intensive participles, both active and middle, 
are comparatively common in the older language. They are 
formed and inflected like those of the reduplicating present, 
and have the accent on the reduplicating syllable. 

Examples are: active, cdkafat, ndnadat, cekitat, memyat, ftffweat, roruvat, 
ddrdrat, mdrmrjat, jdnghanat, ndnnamat, pdriiphanat, kdnikradat, ddvidyutat ; 
middle, bdbadhana, memyana, cekitana, y6yuvana, rdrucana, jdrbhurana, 
sarsrana, ndnnamana, ddndafana. No middle participle shows the dissyllabic 

1013. On account of their accent, rarahand, raraksand, and jdhrsand 
(beside jdrhrsana) are probably to be regarded as perfect participles, although 
no other perfect forms with heavy reduplication from the same roots occur. 
The inference is, however, rendered uncertain by the unmistakably intensive 
badbadhand and marmrjand (beside mdrmrjana}. 

The RV. has once jdnghnatas, gen. sing., with root- vowel cast out; 
kdnikrat appears to be used once for kdnikradat; if cakdt is to be referred 
to yka (Grassmann), it is the only example of an intensive from a root in 
a, and its accent is anomalous. Marmrcantas (AB.) is probably a false 

6. Imperfect. 
1014. The imperfect is regularly inflected as follows : 

dvevidam dvevidva dvevidma 

2 3^rT Sl^f^te^ ^ijfori^ 5^j%fT 

dvevet, dvevidis dvevittam dvevitta 


dvevet, dvevidlt dvevittam dvevidus 

1015. The imperfect forms found in the earlier texts are not numerous. 
They are, including those from which the augment is omitted, as follows: 
in active, 1st sing., acaka^am, dedi^am; 2d sing., ajagar, adardar, ddrdar, 
cakdn; 3d sing., adardar, adardhar, avarivar, dardar, kdniskan, ddvidyot, 
ndvinot, and cakdn and rardn; 2d du., adardrtam; 1st pi., marmrjma; 
3d pi., araranws, anannamus, adardirus, acarkrsus, djohavus, anonavus : 
and, with auxiliary z, in 3d sing., dvavafit, dvavarit, dyoyavit, droravlt; 
djohavit; and, irregularly, in 3d du., avavafltam. The middle forms are 

1018] INTENSIVE. 329 

extremely few: namely, 3d sing., ddedista, dnannata (with loss of the final 
radical in a weak form of root); 3d pi. marmrjata, and avavafanta (which, 
if it belongs here, shows a transfer to an a-stem). 

1016. Derivative Middle Inflection. From every 
intensive stem. . as above described, may be formed in the 
present-system a further derivative conjugation which is 
formally identical with a passive, being made by the accented 
sign T yd, along with middle endings only. It has not, 
however, a passive value, but is in meaning and use in- 
distinguishable from the simpler conjugation. 

A final vowel before this ya is treated as before the passive- 
sign ya (770). 

The inflection is precisely like that of any other stem end- 
ing in a in the middle voice : thus, from ymrj, intensive stem 
marmrj, is made the present indicative marmrjye, marmrjydse, 
marmrjydte, etc.; optative marmrjyeya, marmrjyethas, marmrjyeta, 
etc.; imperative marmrjydsva, marmrjydtam, etc.; participle 
marmrjydmana ; imperfect dmarmrjye, dmarmrjyathas, dmarmrjyata, 
etc.; subjunctive forms do not occur. 

1017. This kind of intensive inflection is said to be 
much more usual than the other in the later language; in 
the earlier, it is comparatively rare. 

In RV., 7/6-forms are made from eight roots, five of which have also 
forms of the simpler conjugation; the AV. adds one more; the other earlier 
texts (so far as observed) only twelve more, and half of them have likewise 
forms of the simpler conjugation. Thus: from j/mr/, marmrjydte etc., and 
marimrjytta ; from ytr, tarturyante; from year, carcurydmana ; from ynl, 
nenlyeran etc.; from yvi, veviyate; from yrih, rerihydte etc.; from yvij, 
vevijydte; from ysku, coskuydse etc.; from ydif, dedicate; from j/fcaf, 
cakacydte ; from yvad, vavadydmana; from }/nam, nannamyadhvam ; from 
yvah, vanivahyeta etc. (with lengthened root- vowel, elsewhere unknown); 
from ykrand, kanikradydmana ; from yvrt, varivartydmana (QB.: should be 
varivrty-); from ymrf, amarimrfyanta (QB.? the text reads amarimrt syanta) ; 
from yyup, yoyupydnte etc.; from ynud, anonudyanta; from yvli, avevllyanta; 
from yjabh, janjabhydte etc.; from yjap, janjapydmana. 


1018. The grammarians are at variance as to whether 
a perfect may be formed directly from the intensive stem, 


or whether only a periphrastic perfect (below, 1070 if.) is to 
be admitted. 

In the older language, no example of an intensive periphrastic perfect 
has come to light. A few unmistakable perfect forms are made from the 
intensively reduplicated root in RV. : namely, davidhava and ndnava, 3d sing., 
and nonuvus, 3d pi., and TS. has once dodrdva. To these may be added 
jdgara 1st sing, and jdgara 3d sing.: but as to these, see below, 1020. 

Aorist, Future, etc. 

1019. As to the remaining parts of a full verbal con- 
jugation, also, the grammarians are not agreed (occurrences 
of such forms, apparently, being too rare to afford even 
them any basis for rules) ; in general, it is allowed to treat 
the intensive stem further as a root in filling up the scheme 
of forms, using always the auxiliary vowel ^ i where it is 
ever used in the simple conjugation. 

Thus, from y vid t intensive stem vevid, would be made the 
aorist avevidisam with precative vevidyasam, the futures vevidisyami 
and veviditasmi, the participles vevidita, veviditavya, etc., the in- 
finitive veviditum, and the gerunds vevidiiva and -vevidya. And, 
where the intensive conjugation is the derivative middle one, 
the aorist and futures would take the corresponding middle form. 
Of all this, in the ancient language, there is hardly a trace. The RV. 
has cdrkrse, 3d sing, mid., of a formation like hise and stuse (894 d), and 
the gerundives vitantasayya, and marmrjenya and vdvrdhenya , and QB. has 
the participle vanivahitd, and the infinitive dedlyitavdf. As to jdgarisydnt 
and jdgaritd, see the next paragraph. 

1020. There are systems of inflection of certain roots, the 
intensive character of which is questioned or questionable. Thus : 

The root gr (or gar] 'wake' has from the first no present-system save 
one with intensive reduplication; and its intensive stem, jdgr, begins early 
to assume the value of a root, and form a completer conjugation; while by 
the grammarians this stem is reckoned as if simple and belonging to the root- 
class (I.), and is inflected throughout accordingly. Those of its forms which 
occur in the older language have been given along with the other intensives 
above. They are, for the present-system, the same with those acknowledged 
as regular later. The older perfect is like the other intensive perfects found 
in RV.: namely, jdgara etc., with the participle jdgrvdns; and a future, 
jagarisydnt, and a passive participle jdgaritd are met with in the Brahmanas. 
The old aorist (RV.) is the usual reduplicated or so-called causative aorist: 
thus, ajigar. The grammarians give it in the later language a perfect with 

1026] INTENSIVE. 331 

additional reduplication, jajdgdra etc., an zs-aorist, ajagarisam, with precative 
jdgarydsam, and everything else that is needed to make up a complete con- 

1021. The stem irajya (active only), 'regulate', from which a number 
of forms are made in RV., has been viewed as an intensive from yraj or 
rj. It lacks, however, any analogy with the intensive formation. The same 
is true of iradh, 'propitiate' (only iradhanta and irddhyai, apparently for 

The middle stem Tj/a, not infrequent in the oldest language, is usually 
called an intensive of yi 'go', but with very doubtful propriety, as it has 
no analogy of form with any intensives. The isolated 1st pi. imafte, com- 
mon in RV., is also of questionable character. 

1022. The root U 'totter', with constant intensive reduplication, leli, 
is quite irregular in inflection and accent: thus, pres., leldyati and lelayate 
pples leldydntl and leldyatas (gen. sing.) and lelayamana, impf. alelayat, 
alelet and alellyata. 

1023. The RV. anomalous form dart (or dard), 2d and 3d sing, from 
ydv or dar, is [doubtfully referred to the intensive, as if abbreviated from 
dardar. RV. has once avarlvus (or -vur) where the sense requires a form 
from yvrt, as avarivrtus. The form rarandta (RV., once) seems corrupt. 

1024. A marked intensive or frequentative meaning is not 
always easily to be traced in the forms classed as intensive ; and 
in some of them it is quite effaced. Thus, the roots cit, mj, 
vis use their intensive present-system as if it were an ordinary 
conjugation-class ; nor is it otherwise with rand/i, ran, kan (of 
which, as noticed above, the forms admit of being referred to 
the perfect-system), and with gr (jagr). The grammarians reckon 
the inflection of nij and vis as belonging to the reduplicating 
present-system (II.), with irregularly strengthened reduplication ; 
and they treat in the same way vie and vij; jagr, as we have 
seen, they account a simple root. 

Also daridra, intensive of ydra 'run', is made by the grammarians a 
simple root, and furnished with a complete set of conjugational forms : as 
dadaridrau; adaridraslt, etc. etc. It does not occur in the older language. 
The so-called root vevi 'flutter' is a pure intensive. 

1025. It is allowed by the grammarians to make from the intensive 
stem also a passive, desiderative, causative, and so on : thus, from vevid, 
pass, vevidye; desid. vevidisdmi ; caus. veviddydmi; desid. of causative, ve- 
vidayisami. But such formations are not found in the older language, and, 
if they occur at all, are excessively rare in the later. 

III. Desideraiive. 

1026. By the desiderative conjugation is signified a de- 
sire for the action or condition denoted by the simple root : 


thus, sHM pibami, 'I drink', desid. nHlHl 1 pipasami, 'I 
wish to drink'; ^SRwfivami, 'I live', desid. ist 

visami, ( I desire to live'. Such a conjugation is allowed to 
be formed from any simple root in the language, and also 
from any causative stem. 

The desiderative conjugation, although its forms outside 
the present-system are extremely rare in the oldest language, is 
earlier and more fully expanded into a whole verbal system 
than the intensive. Its forms are also of increasing frequency: 
much fewer than the intensives in RV., more numerous in the 
Brahmanas and later ; not one third of the whole number of 
roots (about ninety) noted as having a desiderative conjugation 
in Veda and Brahmana have such in RV. 

1027. The desiderative stem is formed from the simple 
root by the addition of two characteristics: a. a reduplica- 
tion, which always has the accent; b. an appended H sa 
which, however (like the tense-signs of aorist and future), 
sometimes takes before it the auxiliary vowel S e, becoming 
^ isa. 

1028. The root in general remains unchanged; but with 
the following exceptions : 

a. A final * or u is lengthened before sa: thus, ciksisa, cikisa, 
jigisa ; cucrusa, juhusa . 

b. A final r becomes ir or ur before sa: thus, cikirsa, 
sislrsa, jihirsa; bulhursa, tustursa (the only examples noted from 
the older texts). 

c. Before isa, the same finals necessarily, and a penulti- 
mate i or u or r optionally, have the ywm-strengthening (no 
examples are quotable from the older texts). 

More special exceptions are : 

d. A few roots in a weaken this vowel to I or even i: thus, piplsa 
(beside pipasa) from ypa 'drink', jihlsa (AV.) from yha 'remove' (jihlte: 664); 
didhisa (beside dhitsa) from ydha. 

e. A few roots in an or am lengthen the vowel : thus, jigansa (beside 
jiyamisa) from ygam ; jighahsa from yhan ; mlmahsa from ]/man , and ytan 
is said to make titahsa. 

f. Reversion to guttural form of an initial after the reduplication is 
seen in cikisa from yd, cikitsa from ycit, jiglsa from yji, jighahsa from 
yhan; and yhi is said to make jighlsa. 




g. The roots van and san make vivasa and sisasa, from the root-forms 
va and sa. 

h. The root jiv forms jujyusa (B.: jijlvisa, VS.); and the other roots 
in lv (or iv : 765) are required to make the same change before sa, and to 
have guna before isa: thus, susyusa or sisevisa. Svap forms swsupser. 
Dhurv forms dudhursa. 

i. Initial s is usually left unchanged to * after the reduplication when 
the desiderative sign has s (184 c): thus, sisanksa (B.: ysanj), and susyusa 
and sisanisa, according to the grammarians. 

k. Further may be mentioned as prescribed by the grammarians : 
ninanksa (or ninafisa) from j/nap, 'perish' ; mimanksa from ymajj ; mimar- 
jisa (or mimrksa) from J/mr/. 

1029. The consonant of the reduplication follows the 
general rules (590); the vowel is ^ i if the root has an a- 
vowel, or ft r, or an i-vowel; it is 3 u if the root has an 
u- vowel. But: 

a. A few roots have a long vowel in the reduplicating syllable : thus, 
blbhatsa from ybadh or badh; mlmansa from yman; and tutursa (RV.) i'rom 

b. From j/ap is made (in B). atffisa (with a mode of reduplication 
like that followed sometimes in the reduplicating aorist: 862). The gram- 
marians give other cases of the same kind: thus, arjihisa from yarh, ici- 
ksisa from yiks, undidisa from yund, ardidhisa from y-rdh. In the older 
language, ap is the only root with initial vowel which forms a desiderative 
stem, except ap and rdh, which have abbreviated stems : see the next paragraph. 

C. RV. has the stems fnaksa and fyaksa, regarded as desideratives 
from yynaf 'attain' and yaj, with mutilated reduplication. 

1030. A number of roots, including some of very com- 
mon use, form an abbreviated stem apparently by a con- 
traction of reduplication and root together into one syllable : 

thus, 37 ~ipsa from |/^TC ap ; rTT ditsa from y^J da. 

Such abbreviated stems are found in the older language as follows : 
dhitsa (beside didhisa) from ydhd; ditsa (beside didasa) from yda ; dipsa from 
ydabh; fiksa from ]/fafc; slksa from ysah: these are found in RV.: in AV. are 
added ipsa from yap (RV. has apsa once), and Irtsa from yrdh: the other texts 
furnish lipsa (3.) or llpsa (TB.) from yiabh, ripsa (GB.) from yrabh, pitsa 
(93.) from ypad, and dhiksa (QB.) from ydih (Or, rather, dah). The gram- 
marians prescribe dhipsa or dhlpsa from ydabh, instead of dipsa; they form 
pitsa from ypat as well as pad ; and they add ritsa from yradh, jnlpsa 
(beside jijnapayisa] from the causative quasi-root jnap (below, 1042e), and 
mitsa from yyma ard ml and mi: this last could be only an anomalous 




formation, made after the analogy of the others. Also moksa is reckoned as 
a desiderative stem from )/mwc (it is denominative, rather). 

1031. The use of the auxiliary vowel ^ i is quite rare 
in the early language , but more common later ; and it is 
allowed or prescribed by the grammarians in many stems 
which have not yet been found in actual use. 

It is declared to follow in general, ithough not without 
exceptions, necessary or optional, the analogy of the futures 
(934, 943). 

No example of the use of i is found in RV., and only one each in AV. 
(pipatisa), VS. (jijlvisa), and TS. (jigomisa). The other examples noted in 
the early texts are afifisa, cikramisa, jigrahlsa (with I for i, as elsewhere 
in this root), cicarisa, jijanisa, didlksisa, bibadhisa, riradhisa, vividisa, jihih- 
sisa: most of them are found only in QB. Stems also without the auxiliary 
vowel are made from roots gram, jiv, badh, vid, 

1032. Inflection: Present-System. The desider- 
ative stem is conjugated in the present-system with per- 
fect regularity, like other a-stems, in both voices, in all the 
modes (including, in the older language, the subjunctive), 
and with participles and imperfect. It will be sufficient to 
give here the first persons only. We may take as active 
model <% ipsa, 'seek to obtain', from |/5TFJ dp; as middle 
icifrtsl titiksa, 'endure', from ]/fer tij\ 'be sharp' (see below, 

1. Present Indicative. 













2. Present Subjunctive. 

Ipsani Ipsava 



Ipsama titiksai titiksavahai titiksamahai 

etc. etc. etc. etc. 

3. Present Optative. 

ipseyam ipseva 

etc. etc. 

ipsema titikseya titiksevahi titiksemahi 





1035] DESIDERATIVE. 335 

4. Present Imperative. 
fcilritf^ J 

ipsa Ipsatam ipsata titiksasva titiksetham titiksadhvam 

etc. etc. etc. etc. etc. etc. 

5. Present Participle. 

^Hti ipsant (f. ^Htil fysanfi] (rilrl^tHIUl titiksamana. 
6. Imperfect. 

aipsam aipsava aipsama dtitikse dtitiksavahi dtitiksamahi 
etc. etc. etc. etc. etc. etc. 

There are almost no irregularities of inflection to be reported from the 
older language. No 1st pi. in most, or 2d pi. in ihana or tana, or impv. 
in tat, is met with. The quotable subjunctive forms are those in sani, sat 
and sat, san, and santa. 

But the fern, pple sisasati (instead of sisasanti) occurs once or twice 
in the older texts. 

1033. Desiderative forms outside the present-system are 
extremely rare in the oldest language. The RV. has only per- 
fect forms from a stem mimiks thus, mimiksdthus, mimiksdtus, 
mimiksus; mimikse, mimiksire along with the present forms 
mimiksati, mimiksa etc., mimiksant (pple): they show that mimiks 
or miks has taken on the character of an independent root. In 
AV. are found two aorist forms, irtsls and acikitsis, and a 
participle or two from mimansa (see below, 1037, 1039) all 
of them from stems which have lost their distinct desiderative 
meaning, and come to bear an independent value. The forms 
noted from the other earlier texts will be given in full below. 

In the later language, the complete system of verbal 
forms is allowed to be made in the desiderative conjugation, 
the desiderative stem, less its final vowel, being treated as 
a root. Thus: 

1034. Perfect. The desiderative perfect is the peri- 
phrastic (1070 if.). 

Thus, ipsam cakara etc.; titiksam cakre etc. Such forms 
are made in QB. from )/)/ kram, dhurv, badh, ruh. 

Apparent perfect forms of the ordinary kind made from mimiks in BV. 
have been noticed in the preceding paragraph. And AB. (viii. 21) has once 
didasitha, 'thou hast desired to give'. 

1035. Aorist. The aorist is of the w-form (5): thus. 
aipsisam 7 tffriffiiji^ dtitiksisi. 


The AV. has acikitsis, and irtsis (augmentless, with ma prohibitive : 
579). TB. has aipsit; and QB. afrtsit, dcikirsls and ajighahsis, and aml- 

A precative is also allowed thus, tpsyasam, titiksisiya, 
but it probably never occurs. 

1036. Futures. The futures are made with the auxil- 
iary vowel ^ i: thus, ^H^UIM Ipsisyami and 

asmi; icifrti^Ti titiksisye and (rlirlRifli^ titiksitahe. 
The QB. has titiksisyate and didrksitdras. 

1037. Verbal Nouns and Adjectives. These too 
are made with the auxiliary vowel ^ i, in all cases where 
that vowel is ever taken. 

In the older language have been rioted : participle in ta, mimahsitd 
(AV., GB.), jijyusita (AB.), fUfrusitd and dhlksita (QB.); gerundive in tavya, 
llpsitavya (AB.), didhyasitavya (QB.); gerund in tva, mimahsitva (K.). 

1038. A desiderative adjective in u for example, dipsu, bibhatsii, 
sisasu is of frequent occurrence, and has the meaning and construction of 
a present participle. An abstract noun in a for example, jiglsd is 
also a usual appendage to the desiderative conjugation. Adjectives in enya 
(having a gerundive character : 966 b) are occasionally met with from the 
earliest time : thus, didrksenya (RV.), cuprusenya (TS.), ninlsenya (PB.); also, 
with irregular reduplication (apparently) paprksenya (RV.). RV. has also 
didhisdyya (966 c). 

1039. Derivative or Tertiary Conjugations. A 
passive is allowed to be made, by adding the passive-sign 
7J yd to the desiderative root (or stem without final a) : thus, 
sCHTcT ipsydte, 'it is desired to be obtained'; and a caus- 
ative by adding in like manner the causative-sign 3BRJ dya 
(1041): thus, ^HtillH ipsdyami, 'I cause to desire obtain- 


The only trace of such formations noticed in the older language is the 
participle mimansydmana (apparently to be read instead of mimahsdmana, 
AV. ix. 6.24). 

For the desiderative conjugation formed on causative stems, 
which is found as early as the Brahmanas, see below, 1052 b. 

1040. Some stems which are desiderative in form have lost 
the peculiarity of desiderative meaning, and assumed the value 
of independent roots: examples are cikits, 'cure', juyups, 'despise', 
titties, 'endure', bibhats, 'abhor', mimans, ''ponder'. Doubtless 

10421 CAUSATIVE. 337 

some of the apparent roots in the language with sibilant final 
are akin with the desideratives in origin. 

IV. Causative. 

1041 . In the later language is allowed to be made from 
every root a complete causative conjugation. The basis of 
this is a causative stem, formed by appending the causative- 
sign 5HT dya to the, usually strengthened, root. 

But by no means all conjugation-stems formed by the 
sign 5RJ dya are of causative value; and the grammarians 
regard them as a conjugation-class, the tenth or cwr-class, 
according to which roots may be inflected as according to 
the other classes, and either alone or along with others. 

In RV., the proportion without causative value is fully one third. The 
formation is a more obviously denominative one than any of the other con- 
jugation-classes, an intermediate between them and the proper denominatives. 
A causative meaning has established itself in connection with the formation, 
and become predominant, though not exclusive. A number of roots of late 
appearance and probably derivative character are included in the class, and 
some palpable denominatives, which lack only the usual denominative accent 
(below, 1055). 

The causative formation is of much more frequent use, and more de- 
cidedly expanded into a full conjugation, than either the intensive or the 
desiderative. It is made from more than two hundred and fifty roots in the 
early language (inRV., from about one hundred and fifty); but in the oldest, 
its forms outside the present-system are (apart from the attached reduplica- 
ted aorist: 1046) exceedingly few. 

1042. The treatment of the root before the causative- 
sign 5HI ctya is as follows: 

a. Medial or initial i, u, r, I have the #ww#-strengthening 
(if capable of it) : thus, vedaya from j/Wd, codaya from ycud, 
tarpaya from ~\/trp; and kalpaya from }/klp (only example). 

But a few roots lack the strengthening: these are, in the older lang- 
uage, cit (citaya and cetaya), vip (vipaya and vepaya), is, il and , ri* 
(rimya and resaya), tuj, tur, dyut (dyutaya and dyotaya), mrd, sprit; and 
yrabli makes in RV. grbhaya. Dus and guh lengthen the vowel instead. 
Mrj sometimes has vrddhi, as in other forms : thus, marjaya (beside 

b. A final vowel has the rrdfZ/-strengthening : thus, cyavaya, 
bhavaya, dharaya, saraya. 

Whitney, Grammar. 22 


But no root in i or I has vrddhi in the older language (unless payaya 
[d, "below] comes from pi rather than pa) as, indeed, regular causatives 
from such roots are hardly quotable : only RV. has ksayaya from yksi 'dwell' ; 
for a few alternatively permitted forms, see below, e. 

A few roots have (generally in the older language only) a form also 
with puna-strengthening : thus, dru, yu 'ward off', fru, pu, jr 'decay", dr 
'burst', sr, hr; vr 'choose' makes varaya later (it is not found in V. or Br.). 

c. A medial or initial a in a light syllable is sometimes 
lengthened, and sometimes remains unchanged : thus, bhajaya, 
svapaya, adaya; janaya, crathaya, anaya. 

The roots in the older language which keep their short a are an, ;'an, 
pan, svan, dhan, ran, stan, dhvan, gam (gdmaya once in RV.), * am > dam, 
nam, prath, pratfe, fnath, vyath, svad, nad, das, dhvas, mah, nabh, tvar, 
svar. Some have both forms: namely, pat, chad, mad, ram, cam; raj has 
rajaya (AV., once) and ranjaya. The roots which lengthen the vowel are 
decidedly the more numerous. If a nasal is taken in any of the strong 
forms of a root, it usually appears in the causative stem: thus, randhaya, 
lambhaya, rambhaya, skandaya. 

d. Most roots in final a, and the root r, add p before the 
conjugation-sign: thus, dapaya, dhapaya, sthapaya; arpaya. 

Such stems are made in the older language from the roots khyd, gla, 
ghrd, jnd, da 'give', drd 'run', dha 'put' and dha 'suck', ma 'measure', mid, vd 
'blow', and va 'tire', stha, sna, ha 'remove' and ha 'leave'. From ;na and 
sna are found in AV. and later the shortened forms jnapaya and snapaya, 
and from fra only frapaya (not in RV.). Also gla forms in the later language 

Stems from a-roots showing no p are, earlier, payaya from ypa 'drink' 
{or pi), pyayaya from ypya or pyay ; sayaya from ysa (or si); also, later, 
payaya from }/f, hvayaya from yhva; and further, from roots cha, va 
'weave', and vya, according to the grammarians. 

e. The same p is taken also by a few i and I-roots, with various ac- 
companying irregularities: thus, ksepaya from yksi 'dwell' (RV., beside 
ksayaya} ; ksapaya ( AV. ) and ksapaya and ksayaya from yksi 'destroy' ; 
japaya (VS. and later) from yji; lapaya (TB. and later) from yil; adhya- 
paya from adhi-\-yi; smapaya (beside smayaya, which does not occur) 
from ysmi; hrepaya from yhri; and, according to the grammarians. 
repaya from yrl, vlepaya from yvll, krapaya from ykri, bhapaya (beside 
bhayaya and bhlsaya) from ybhl, and cdpaya (beside cdyaya) from yd. 
Moreover, yruh forms later ropaya (earlier rohaya), and yknii or knuy is 
said to form knopaya. 

f. More anomalous cases, in which the so-called causative is palpably 
the denominative of a derived noun, are: pdlaya from ypa 'protect'; prlnaya 
from yprl; llnaya (according to grammarians) from yil; dhunaya (not cau- 
sative in sense) from ydhu; bhlsaya from ybhl; ghdtaya from yhan; sphdvaya 
{according to grammarians) from ysphd or sphdy. 

1043] CAUSATIVE. 339 

1043. Inflection: Present-System. The causa- 
tive stem is inflected in the present-system precisely like 
other stems in f a : it will be sufficient to give here in gen- 
eral the first persons of the different formations, taking as 
model the stem UT^I dhardya, from j/J dhr. Thus : 

1. Present Indicative. 

active. middle. 

d. p. s. d. 

dhardyami dhardyavas dhardyamas dhardye dhardyavahe dhardyamahe 

etc. etc. etc. etc. etc. etc. 

The 1st pll. act. in masi greatly outnumber (as 10 to 1) those in mas 
in both RV. and AV. No example occurs of 2d pi. act. in thana, nor of 
3d sing. mid. in e for ate. 

2. Present Subjunctive. 

For the subjunctive may be instanced all the forms noted 
as occurring in the older language : 

1 dhdrdyani dhdrdydva dhdrdydma dhdrdydi dhdrayavahdi 

(dhdrdydsi (dhdrdyddhve 

2 I _ , _ dharayathas dharayatha dharayase 

(dharayas (dhdrdyadhvdi 

\ dhdrdydti _ ( dhdrdydte 

3 { dharayatas dharayan ],,_ , .. dharayaite 
( dharayat ( dharayatai 

Only one dual mid. form in aite occurs : maddyaite (RV.). The only 
RV. mid. form in ai, except in 1st du., is madayadhvai. The primary end- 
ings in 2d and 3d sing. act. are more common than the secondary. 

3. Present Optative. 

dhardyeyam dhardyeva dhardyema dhardyeya dhardyevahi dhurdyemahi 
etc. etc. etc. etc. etc. etc. 

Optative forms are very rare in the oldest language (four in RV., two 

in AV.); they become more common in the Brahmanas. AB. has once kamayita. 

4. Present Imperative. 

dhardya dhardyatam dhurdyata dhardyasva dhardyetham dhardyadhvam 

etc. etc. etc. etc. etc. etc. 

Imperative persons with the ending tat occur: dharayatat (AV.) is 2d 
sing.; gamayatat and cyavayatat (K. etc.), and varayatat (TB.) are used as 
2d pi.; varayadhvcit (K. etc.) is 2d pi., and the only noted example (see 
above, 570). 




5. Present Participle. 
dhardyant (f. Ttft -yanfi) ^|(t(HIUI dhardyamana. 

6. Imperfect. 

ddharayam ddharayava ddhurayama ddharaye ddharayavahi ddharayamahi 

etc. etc. etc. etc. etc. etc. 

For a few forms in Is and It which perhaps belong to the imperfect, 
see below. 

1044. As was above pointed out, the formations from the causative 
stem in aya outside the present-system are in the oldest language very 
limited. In RV. are found two forms of the future in syami, and ten in- 
finitives in dhydi ; also one or two derivative nouns in tr (bodhayitr, coday- 
itri), five in isnu, seven in itnu, and a few in a (atiparayd, nidharayd, 
vacaminkhayd, vi$vamejaya). In AV., also two s-future forms and four ger- 
unds in tva; and a few derivative noun-stems, from one of which is made 
a periphrastic perfect (gamayam cakdra}. In the Brahmanas, verbal deriva- 
tive forms become more numerous and various, as will be noted in detail 

1045. Perfect. The accepted causative perfect is the 
periphrastic (1070), the derivative noun in 5TF a, in accusa- 
tive form, to which the auxiliary is added, being formed 
from the causative stem: thus, 

TT5[Ert cjsfil^ dharayam cakara STTfETt tT^IT dharayhm cakre. 
Of this perfect no example occurs in RV. or SV. or VS., only one 
gamaydm cakdra in AV., and but two or three in all the various texts 
of the Black Yajur-Veda, and these not in the mantra-parts of the text. 
They are also by no means frequent in the Brahmanas, except in Qli- 
(where they abound : chiefly, perhaps, for the reason that this work uses 
in considerable part the perfect instead of imperfect as its narrative tense). 

1046. Aorist. The aorist of the causative conjugation 
is the reduplicated, which in general has nothing to do with 
the causative stem, but is made directly from the root. 

It has been already fully described (above, 856 ff.). 

Its association with the causative is doubtless founded on 
an original intensive character belonging to it as a reduplicated 
form, and is a matter of gradual growth : in the Veda it is 
made from a considerable number of roots (in RV., more than 
a third of its instances; in AV., about a fifth) which have no 
causative stem in aya. 

1050] CAUSATIVE. 341 

The causative aorist of y J dhr. then, is as follows : 

ddidharam ddidharava ddldharama ddidhare ddldharavahi adidharamahi 

etc. etc. etc. etc. etc. etc. 

1047. In a few cases, where the root has assumed a pecu- 
liar form before the causative sign as by the addition of a p 
or s (above, 1042 d if.) the reduplicated aorist is made from 
this form instead of from the simple root : thus, atisthipam from 
sthap (stem sthapaya] for ystha. Aorist-forms of this character 
from quasi-roots in ap are made from sthap, jnap, hap, jap, and 
crap (above, 861) ; the only other example from the older language 
is bibhisas etc. from bhis for ~\/bhi. 

1048. A small number of Vedic forms having Is and It after y of the 
causal stem are apparently sporadic attempts at making an fc-aorist : thus, 
vyathayls (AV.), dhvanayit (RV.; TS. has instead the anomalous dhvanayit}, 
and ailayit (AV.). The two former are augmentless forms, used with ma 
prohibitive (compare the denom. unayls, RV., also with ma). 

1049. A precative is of course allowed by the gram- 
marians to be made for the causative conjugation: in the 
middle, from the causative stem with the auxiliary S i sub- 
stituted for its final ^ a; in the active, from the form of 
the root as strengthened in the causative stem, but without 
the causative sign: thus, 

LntllHH dharyasam etc. E-H^lUcfltl dharayisiya etc. 
This formation is doubtless to be regarded as purely fictitious. 

1050. Futures. Both futures, with the conditional, 
are made from the causative stem, with the auxiliary ^ t, 
which takes the place of its final f a. Thus: 

dharayisyhmi etc. ^|^|6jo.6| dharayisye etc. 


M adharayisyam etc. fc||^|kj adharayisye etc. 

Periphrastic Future. 
dharayit&smi etc. M(KJcn< dharayithhe etc. 

The 5-future participles are made regularly : thus, dharayi- 
syant (fern, -ydnfi or -yatl), dharayisydmana. 


It has been mentioned above that RV. and AV. contain only two 
examples each of the s-future, and none of the periphrastic. The former 
begin to appear in the Brahmanas more numerously, but still sparingly, 
with participles, and conditional (only adharayisyat, QB.); of the latter, QB. 
affords two examples (parayitdsmi and janayitdsi). 

1051. Verbal Nouns and Adjectives. These are 
in part formed from the causative stem in the same manner 
as the futures ; in part namely, the passive participle in 
cT ta and the gerundive and gerund in U ya (and the root- 
infinitive) from the causatively strengthened root- form. 
The auxiliary ^ i is taken in every formation which ever 
admits that vowel. 

Thus, of formations permitted in the later language (but 
the examples taken from the earlier): 

participle in ta : irita, vasitd, cravitd; 

gerundive in tavya: tarpayitavya, kalpayitavya, gamayitavya, 
lhaksavitavya ; 

<7 e7 f 

gerundive in ya : sthapya, yajya ; 

infinitive in turn: josayitum, dharayitum, janayitum, parayittim; 
gerund in tva: kalpayitva, sadayitvn, -arpayitvti (AV.: see 
990), -rocayitva (TA.), crapayitva (AB.); 

gerund in ya : -gharya, -padya, -vasya, -sadya, -sthapya ; 
gerund in am: -sthapam (QB.). 

Further, of formations found only in the older language : 
root-infinitive, accusative: -sthapam (PB.); 

infinitive in tavai : jdnayitavaf, tdrpayitavat, pdyayitavai, -kalpayitavaf, 
-Ccotayitavaf (all $B.); 

infinitive in dhyai: isayddhyai, irayddhyai, tahsayddhyai, na$ayddhyai, 
mandayddhyai, madayddhyai. risayddhyai, vartayddhyai, vajayddhyai, 
syandayddhyai (all RV.); 

gerundive in ayya : panaydyya, sprhaydyya, trayaydyya (ytra : ?). 
All these, it will be noticed, follow the same rule as to accent with 
the similar formations from the simple root, showing no trace of the special 
accent of the causative stem. 

1052. Derivative or Tertiary Conjugations. 
From the causative stem may be made a passive and a de- 
siderative conjugation. Thus : 

a. The passive-stem is formed by adding the usual pas- 
sive-sign T yd to the causatively strengthened root, the caus- 
ative-sign being dropped: thus, STRTrT dharydte. 


Such passives are not found In the Veda, but they are met with here 
and there throughout the Brahmana language : examples are jnapyd- (TS.), 
sadya- (K.), padya- (AB.), vadya- (TB.), sthapya- (GB.), and so on. 

b. The desiderative stem is made by reduplication and 
addition of the sign ^ isa, of which the initial vowel replaces 
the final of the causative stem : thus, (^MI^fllMTH didharayisati . 

These, too, are found here and there in the Brahmanas and later: 
examples are pipayayisa (K.), bibhavayisa and cikalpayisa and lulobhayisa 
(AB.), didrapayisa and riradhayisa and apipayisa (QB.), and so on. 

As to causatives made from the desiderative stem, see above, 1039. 

V. Denominative. 

1053. A denominative conjugation is one that has for 
its basis a noun-stem. 

It is a view now prevailingly held that most of the present- 
systems of the Sanskrit verb, along with other formations anal- 
ogous with a present-system, are in their ultimate origin denom- 
inative ; and that many apparent roots are of the same character. 
The denominatives which are so called differ from these only in 
that their origin is recent and undisguised. 

1054. The grammarians teach that any noun-stem in 
the language may be converted, without other addition than 
that of an f a (as union-vowel enabling it to be inflected 
according to the second general conjugation) into a present- 
stem, and conjugated as such. 

As examples of what is allowed in this way, are given 
kr swati, 'acts like Krishna'; malati, 'is like a garland (mala)'; 
kavayati, "plays the poet (kam)'; bhavati, 'is like the earth (bhuf; 
crayati, 'resembles Qri (goddess)'; pitarati, 'acts the father'; 
rajanati, 'is kingly'. But such formations are at the best of 
extreme rarity in actual use. The RV. has a few isolated and 
doubtful examples, the clearest of which is bMsd&ti, 'he heals', 
from bhisaj, 'physician' ; it is made like a form of the root-class ; 
abhisnak seems to be its imperfect according to the nasal class. 
And patyate, 'he rules', appears to be a denominative of pdti, 
'master'. Other possible cases are (Delbriick) isanas etc., krpd- 
nanta, tarusema etc., vanusanta, bhurafanta, vdnanvati. None of 
the other Veda or Brahmana texts has anything additional of the 
same character. 

1055. In general, the base of denominative conjugation 



is made from the noun-stem by means of the conjugation- 
sign ET ya- which has the accent. 

The identity of this ya with the ya of the so-called causative conjugation, 
as making with the final a of the noun-stem the causative-sign aya, is 
altogether probable. What relation it sustains to the ya of the t/a-class 
(VIII.), of the passive, and of the derivative intensive stem, is more open 
to question. 

1056. Intermediate between the denominative and causative 
conjugations stands a class of verbs, plainly denominative in 
origin, but having the causative accent. Examples, beginning 
to appear at the earliest period of the language, are mantrdyate 
(from mantra, y man -\-tra] and kirtdyati (from kirti, ykr 'praise'). 
These, along with like forms from roots which have no other 
present-system (though they may make scattering forms outside 
that system from the root directly), or which have this beside 
other present-systems without causative meaning, are reckoned 
by the grammarians as a separate conjugation-class (above, 607). 

1057. Denominatives are formed at every period in the 
history of the language, from the earliest down. 

They are most frequent in RV., which contains over a 
hundred, of all varieties ; AV. has only half as many (and per- 
sonal forms from hardly a third as many : from the rest, present 
participles, or derivative nouns); AB., less than twenty; QB., 
hardly more than a dozen; and so on. In the later language 
in general, they are far from numerous ; and most of those 
which occur are "cwr-class" verbs. 

1058. The denominative meaning is, as in other lang- 
uages, of the greatest variety ; some of the most frequent 
forms of it are: 'be like, act as, play the part of, 'regard 
or treat as', 'cause to be. make into', 'use. make applica- 
tion of, 'desire, wish for, crave' that which is signi- 
fied by the noun-stem. 

The modes of treatment of the stem-final are also various ; 
and the grammarians make a certain more or less definite as- 
signment of the varieties of meaning to the varieties of form; 
but this allotment finds only a dubious support in the usages 
of the words as met with even in the later language, and still 
less in the earlier. Hence the formal classification, according 
to the final of the noun-stem, and the way in which this is 
treated before the denominative sign yd, will be the best one 
to follow. 




1059. From stems in a. a. The final a of a noun- 
stem oftenest remains unchanged : thus, amitraydti, 'plays the 
enemy, is hostile'; devaydti, 'cultivates the gods, is pious'. 

b. Final a is lengthened: thus, aghaydti, 'plans mischief; 
priyaydte, 'holds dear'; acvaydti, 'seeks for horses'; acanaydti, 

'desires food'. 

In the Veda, forms of the same verb with short and long a before ya 
sometimes exchange with one another. 

c. It is changed to ~i, or rarely ; thus, adhvariydti, 'per- 
forms the sacrifice'; tavislydti, 'is mighty'; piitriydti or putriy&ti, 
'desires a son'; mahsiydti, 'craves flesh'. 

Denominatives of this form show a special proclivity toward the mean- 
ing 'desire'. 

d. It is dropped (after n or r}: thus, turanydti, 'is rapid'; 
adhvarydti, 'performs the sacrifice'. 

e. Other modes of treatment are sporadic : thus, the addition 
of s, as in stanasyati, 'seeks the breast'; the change of a to e, 
as in varcydti, 'plays the wooer' . 

1060. From stems in a. Final a usually remains, as 
in gopaydti, 'plays the herdsman, protects'; prtanayati, 'fights'; 
but it is sometimes treated in the other methods of an a-stem : 
thus, prtanyati, 'fights'. 

1061. From stems in e, i, and tt, u. Such stems are 
(especially those in u, u] much less common. They show reg- 
ularly i and u before ya: thus, aratiydti (also -tiy-}, 'plots in- 
jury'; jamydti (also -my-), 'seeks a wife'; sakhiydti, 'desires 
friendship'; catruydti, 'acts the foe'; rjuydti, 'is straight'; 
vasuydti, ' desires wealth'; asuydti, 'grumbles, is discontent': with 
short u, gatuydti, 'sets in motion'. 

More rarely, i is treated as a (or else is gunated, with loss of a y}: 
thus, dhunaydti, 'comes snorting'. Sometimes, as to a (above, e), a sibilant 
is added: thus, avisydti, 'is vehement'; urusydti, 'saves'. From dhl, RV. 
makes dhiyaydte. 

1062. From other vowel-stems, a. Final r is said 
to be changed to ri: thus, pitriydti, 'is fatherly': no example in 
use has been noted. 

b. The diphthongs, in the few cases that occur, have their 
final element changed to a semivowel : thus, gavydti, 'seeks cattle, 
goes a-raiding'. 

1063. From consonant-stems. A final consonant 
usually remains before ya : thus, bhisajydti, 'plays the physician, 
cures'; uksanydti, 'acts like a bull'; apasydti, 'is active'; namasydti, 
'pays reverence'; sumanasydte, 'is favorably disposed'; tarmydti, 


But a final n is said to be sometimes dropped, and the preceding vowel 
treated as a final : thus, rajaydte or rajlydti, 'is kingly', fromro/an: vrsaydte 
from vrsan is the only example quotable from the older language. Sporadic 
cases occur of other final consonants similarly treated: thus, ojaydte from 
ojas ; while, on the other hand, an a-vowel is occasionally added to such 
a consonant before ya : thus, isaydti from 10, satvanayati from satvan. 

1064. By far the largest class of consonantal stems are 
those showing a s before the ya; and, as has been seen above, 
a sibilant is sometimes, by analogy, added to a final vowel, 
making the denominative-sign virtually sya or even, with a 
also added after an i or w-vowel, asya; and this comes to be 
recognised in the later language as an independent sign, forming 
denominatives that express desire : thus, madhusyati or ma- 
dhvasyati, longs for honey'; ksirasyati, 'craves milk'. 

1065. The grammarians reckon as a special class of denominatives in 
kamya what are really only ordinary ones made from a compound noun-stem 
having kama as its final member : thus, rathakamyati, 'longs for the chariot' 
(K.: only example noted from the older language); putrakamyati, 'desires a 
son' coming from the possessive compounds rathakama, putrakama. And 
satyapayati, 'declares true' (from satya], is an example of yet another form- 
ation declared to occur. 

1066. a. A number of denominative stems occur in the Veda for which 
no corresponding noun-stems are found, although for all or nearly all of them 
related words appear: thus, ankuya, stabhuya, isudhya; dhisanya, risanya, 
ruvanya, huvanya, isanya; ratharya, f ratharya, saparya; irasya, dapasya, 
makhasya, panasya, sacasya. Those in anya, especially, look like the begin- 
nings of a new conjugation-class. 

b. Having still more that aspect, however, are a Vedic group of stems 
in aya, which in general have allied themselves to present-systems of the 
na-class (V.), and are found alongside the forms of that class: thus, grbhaydti 
beside grbhndti. Of such. RV. has grbhaya, mathaya, prusaya, musaya, 
$rathaya, skabhaya, stabhaya. A few others have no na-class companions: 
thus, damaya, camaya, tudaya (AV.); and panaya, nafaya, vrsaya (]/ur 
'rain'), vasaya (yvas 'clothe'), and perhaps afaya (j/ap 'attain'). 

1067. The denominative stems in RV. and AV. with causative accent- 
uation are : RV. ankhdya, arthdya, isdya (also isayd)^ urjdya, rtdya, krpdya, 
mantrdya, mrgdya, vavrdya, vajdya (also vajayd), vlldya, susvdya (also 
susvayd); AV. adds kirtdya, dhupdya, paldya, virdya, sabhagdya. 

The accent of dnniya and hdstaya (RV.) is wholly anomalous. 

1068. Inflection. The denominative stems are in- 
flected with regularity like the other stems ending in 51 a 
throughout the present-system. Forms outside of that sys- 

1070] DENOMINATIVE. 347 

tern except from the stems which are reckoned to the 
causative or cwr-class, and which follow in all respects the 
rules for that class are of the utmost rarity. 

In RV. occurs no form not belonging to the present-system, unless (as 
seems most likely) unayls (with md prohibitive) is to be regarded as 2d 
sing, of the to-aorist. Unquestionable examples of this aorist are asuyit (B.), 
papayista (TS. iii. 2. 8 3 : pi., with md prohibitive) and avrsayisata (VS.). 
The form dsaparyait (AV. xiv. 2. 20), with ai for z (555 b), might be aorist; 
but, as the metre shows, is probably a corrupt reading; amanasyait, certainly 
imperfect, appears to occur in TB. (ii. 3. S 3 ). B. has the future gopayisyati, 
and TS. the participles kanduyisydnt and kanduyitd. From roots assimilated 
to the causatives occur in the older language mantrayam asa (AB., GB.), 
mantraydm cakratus and cakre (QB.), mantritd (QB., TA.), -mantrya (TB.), 
and one or two other like forms. The gerundival adjectives saparyenya and 
atasayya also are met with. 



1069. ONE periphrastic formation, the periphrastic 
future, has been already described (942 fT.), since it has 
become in the later language a necessary part of every verb- 
al conjugation, and since, though still remaining essen- 
tially periphrastic, it has been so fused in its parts and al- 
tered in construction as to assume in considerable measure 
the semblance of an integral tense-formation. 

By far the most important other formation of the 
class is 

The Periphrastic Perfect. 

1070. This (though almost unknown in the Veda, and 
coming only gradually into use in the Brahmai as) is a 
tense widely made and frequently used in the classical 


It is made by prefixing the accusative of a derivative 
noun-stem in TT a (accented) to the perfect tense of an 
auxiliary verb : namely, of ]/5fj kr 'make', more often of 
y^fc^as 'be 1 , and very rarely of ]/>T bhu, 'be'. 

In the older language, as is pointed out below, kr is used as auxiliary 
almost alone, and bhu not at all. Even in MBh., bhu hardly ever occurs 

1071. The periphrastic perfect occurs as follows: 
a. It is the perfect of the derivative conjugations : in- 
tensive, desiderative, causative, and denominative ; the noun 
in ^T a being made from the present-stem which is the 
general basis of each conjugation : thus, from j/SfT budh, 
intensive emsthiiH bobudhdm. desiderative SfarHM bubhutsdm. 

o *s oo *s 

causative SHMUIH bodhayam; denominative H^IUIH mantr ay- 


The formation from causative stems, and from those denominatives which 
are assimilated to causatives, is by far the most frequent. 

b. Most roots beginning with a vowel in a heavy syl- 
lable (long by nature or long by position) make this perfect 
only, and not the simple one : thus, ^IHH asam from yETITT 
as, f^Fl^iksam from yjjfi^iks; ~S^\^ubjam from y3&!R ubj. 

Excepted are the roots ap and anch, and those beginning 
with a before two consonants (and taking an as reduplication: 

c. The roots (that is, stems reckoned by the grammarians 
as roots) of more than one syllable have their perfect of this 
formation : thus, T^CJU^J^ cakasam. 

But urnu (712) is said to form urnonava only, and jagr and daridra 
(1020, 1024) to have a perfect of either formation. 

d. A few other scattering roots : namely, ay, day, and kas , 
and optionally vid and us, and a few roots of the reduplicating 
class, bhi, bhr, hu, and hri. All these make the derivative noun 
from their present-stem : thus, dayUm, vidhm, os&m, bibhayhm, 
Juhav&m, bibhartim, jihrayhm (these with guna of the final vowel 
before the a). 

An occasional example is met with from other roots : thus, naydm from 
nl (pres.-stem naj/a); hvayam from yhvd (pres.-stem hvaya). 

1072. The periphrastic perfect of the middle voice is 


made only with the middle inflection of y^\ kr ; that of 
the active, with any one of the three auxiliaries. For pas- 
sive use, the auxiliaries 3R7 as and ^ bhu are also allowed 

x c\ 

to take a middle inflection. 

It is unnecessary to give a paradigm of this formation, as 
the inflection of the auxiliaries is the same as in their independ- 
ent use (for that of ykr, see 800 f) ; of ybhu, see 800 b ; of y'cw, 
see 636). 

The connection of the noun and auxiliary is not so close that other 
words are not sometimes allowed to come between them : thus, tdm patayam 
prathamam asa, 'him he first made to fall'; prabhrancaydm yo naghusam 
cakarn, 'who made Naghusha fall headlong' (both Raghuvanc.a). 

1073. The above is an account of the periphrastic forma- 
tion with a derivative noun in am as it appears in the later 
language ; earlier, its aspect is quite different : namely, as that 
of a more general, but quite infrequent, combination of such a 
noun with various forms of the root kr. Thus : 

a. Of forms with the perfect of the auxiliary occurs only a single 
example in the whole body of Vedic texts (metrical): namely, gamaydm 
cakara (AV. xviii.). In the brahmana parts of the Black Yajus texts are 
found vidam cakara (TS., K., MS.) and viddm cakrma (K.), and ydjaydm 
cakara (K.). In the Brahmanas, examples from causative etc. stems in aya 
begin to prevail over others, and in B. they are rather frequent. Examples 
from desiderative stems have been noted only from QB.: they are cikramisdm, 
ruruksam, dudhursdm, blbhatsdm. From simple roots having the same form- 
ation in the later language, occur viddm (TB., QB., GB.), asdm (QB., GB.), 
iksdm (B., GB.), edhdm (8.), juhavdm (AB., TB., B.), bibhaydm ($B.); 
and also laydm (nilaydm) from yil (CB.). 

b. Forms with the aorist of the auxiliary are in the oldest Brahmanas 
as numerous as those with the perfect. Thus, with akar occur ramayam 
(K.), janayam and sadayam and svadayam and sthapayam (MS.); and with 
akran, viddm (TS., MS., TB.). With the aorist optative or precative has 
been noted only pavaydrh kriyat (MS.). 

C. Like combinations with other tenses are excessively rare, but not 
entirely unknown: so, juhavam karoti (Qankh. Qr. Su.). 

d. With any other auxiliary than ykr appears only mantraydm asa 
(AB., GB.; in ^!B. the same noun is combined with ykr in mantrayarh 
cakratus and mantraydm cakre). 

As the examples show, the noun (as in the case of the 
periphrastic future : 945) has its independent accent. 

Participial Periphrastic Phrases. 

1074. Combinations of participles with auxiliary verbs, 
of condition or motion, forming phrases which have an office 


analogous with that of verb-tenses, are not unknown in any 
period of the language. 

They occur even in the Veda, but are far more common 
and conspicuous in the Brahmanas, and become again of little 
account in the later language. 

1075. Examples of the various formations are as follows . 

a. A (usually present) participle with the tenses of the verb i. 'go'. 
This is the combination, on the whole, of widest and most frequent occur- 
rence. Thus : yatha sucya vasah sarhdadhad iyad evam evai 'tabhir yajnasya 
chidram sarhdadhad eti (AB.), 'just as one would mend [habitually] a garment 
with a needle, so with these one mends any defect of the sacrifice'; agnir va 
idarh vdifvdnaro dahann ait (PB.), 'Agni Vaicvanara kept burning this 
creation'; te 'surah pdrajitd ydnto dydvaprthivi updfrayan (TB.), 'those 
Asuras, getting beaten, took refuge with heaven and earth'; te ( sya grhdh 
pafdva upamurydmana lyuh (B.), 'the animals, his family, would be con- 
tinually destroyed'. 

b. The same with the verb car, 'go (continually or habitually)', signifying 
still more distinctly than the preceding a continued or habitual action. Thus : 
agndv agnfy carati prdvistah (AV.), 'Agni is constantly present in the fire'; 
adandyam dandena ghnantaf caranti (PB.), 'they make a practice of beating 
with a rod what is undeserving of punishment'. 

c. The" same with the verbs as, 'sit', and stha, 'stand', with a like 
meaning. Thus, juhvata asate (K.), 'they continue sacrificing'; te l pakramya 
prativavadato 'tisthan (AB.), 'they, having gone off, kept vehemently refusing'. 

In the later language, stha is the verb oftenest used, with predicates, of 
various kind, to make a verbal phrase of continuance. 

d. A participle with as and bhu, 'be'. The participle is oftenest a 
future one; as only is used in the optative, bhu usually in other forms. 
Thus: yah purvam ariijanah syat (AB.), 'whoever may not have made sacri- 
fice before'; samavad eva yajne kurvana asan (GB.), 'they did the same thing 
at the sacrifice'; parikridanta asan (MS.), 'they were playing about'; itara me 
kena devatd upaptd bhavisyanti (AB.), 'wherewith shall the other deities be 
won by me?' ydtra suptvd punar nh 'vadrasydn bhdvati (QB.), 'when, after 
sleeping, he is not going to fall asleep again'; havyam hi vaksyan bhavati 
(AB.), 'for he is intending to cany the sacrifice'; ddsyant syat (K.), 'may be 
going to give'; ytna vdhanena syantsydnt sydt (QB.), 'with what vehicle he 
may be about to drive'. 

Composition with Prepositional Prefixes. 

1076. All the forms, personal and other, of verbal con- 
jugation of both primary and secondary conjugation, 
and even to some extent of denominative (so far as the 


denominative stems have become assimilated in value to 
simple roots) - - occur very frequently in combination with 
certain words of direction, elements of an adverbial character 
(see the next chapter), the so-called prepositions, according 
to the original use of that term, or the verbal prefixes. 

Practically, in the later language, it is as if a compounded root were 
formed, out of root and prefix, from which then the whole conjugation (with 
many derivatives : below, chap. XVII.) is made, just as from the simple 
root. Yet, even there (and still more in the older language: 1081), the 
combination is so loose, and the members retain so much of their independent 
value, that in most dictionaries (that of Monier Williams is an exception) 
the conjugation of each root with prefixes is treated under the simple root, 
and not in the alphabetic order of the prefix. Derivative words, however, 
are by universal agreement given in their independent alphabetic place, like 
simple words. 

1077. Those verbal prefixes which have value as such 
throughout the whole history of the language are given 
below, in alphabetic order, with their fundamental meanings : 
dtij 'across, beyond, past, over, to excess'; 
ddhi, 'above, over, on, on to'; 
ami, 'after, along, toward'; 

ctntdr, 'between, among, within'; 
dpa, 'away, forth, off'; 
dpi, 'unto, close upon or on'; 

abhi, 'to, unto, against' (often with implied vio- 
lence) ; 

R dva, 'down, off'; 

5TT , 'to, unto, at'; 

3^" ud, 'up, up forth or out'; 

3q upa, 'to, unto, toward'; 

ft ni, 'down; in, into'; 

FKT m's, 'out, forth'; 

^^\ pdra, 'to a distance, away, forth'; 

crf^ pdri, 'round about, around'; 

"3 prdj 'forward, onward, forth, fore'; 


prdti, 'in reversed direction, back to or against, 
against, in return'; 

|cf m', -apart, asunder, away, out'; 
?PT sdm, 'along, with, together'. 

a. Some of these, of course, are used much more widely and frequently 
than others. In order of frequency in the older language (as estimated by 
the number of roots with which they are found used in RV. and AV.), they 
stand as follows: pra, a, vi, sam, abhi, ni, ltd, pari, arm, upa, prati, ava, 
nis, ati, apa, para, adhi, api, antar. Apt is of very limited use as prefix in 
the later language, having become a conjunction, 'too, also'. 

b. The meanings given are only the leading ones. In combination with 
the roots they undergo much modification, both literal and figurative yet 
seldom in such a way that the steps of transition from the fundamental 
sense are not easy to trace. Sometimes, indeed, the value of a root is hardly 
perceptibly modified by the addition of the prefix. An intensive force is not 
infrequently given by pari, vi, and sam. 

1078. Prefixes essentially akin with the above, but more 
distinctly adverbial, and of more restricted use, are these: 

dcha (or acha\ 'to, unto': tolerably frequent in RV. (used 
with over twenty roots), but already very rare in AV. (only two 
roots), and entirely lost in the later language ; 

avis, 'forth to sight, in view': used only with the roots 
bhu, as, and kr ; 

tiros, 'through, crossways ; out of sight' : hardly used except 
with 'kr, dha, bhu (in RV., with three or four others); 

purds, 'in front, forward': used with only half-a-dozen roots, 
especially kr, dha, i; 

pradus, 'forth to view': only with bhu, as, kr. 
A few others, as bahia, 'outside', vina, 'without', saksat, 'in view', are 
still less removed from ordinary adverbs. 

1079. Of still more limited use, and of noun rather than 
adverb-value are : 

trad (or prat/i?), only with dha (in RV., once also with kr}: ?raddha, 
'believe, credit'; 

hin, only with kr (and obsolete in the classical language): hinkr, 'make 
the sound hing, low, murmur'. 

And beside these stand yet more fortuitous combinations: see below, 

1080. More than one prefix may be set before the same 
root. Combinations of two are quite usual ; of three, much 
less common; of more than three, rare. Their order is in 
general determined only by the requirements of the meaning, 

1083] VERBAL ^PREFIXES. 353 

each added prefix bringing a further modification to the 
combination before which it is set. But 5TT a is never allowed 
in the later language, and only extremely rarely in the older, 
to be put in front of any of the others. 

1081. In classical Sanskrit, the prefix always stands 
immediately before the verbal form. 

In the older language, however, of both Veda and Brah- 
mana, its position is quite free : it may be separated from the 
verb by another word or words, and may even (much less often) 
come after the form to which it belongs ; it may also stand 
alone, qualifying a verb that is understood, or conjointly with 
another prefix one that is expressed. 

Thus, sd devdn e 'hd vaksyati (RV.) f 'he shall bring the gods hither'; 
pro, no, dyunsi tdrisat (AV.), 'may he lengthen out our lives'; tdv d yatam 
tipa dravdt (RV.), 'do ye two come hither quickly 1 ; gdmad vdjebhir d sd nah 
(RV.), 'may he come with gifts hither to us'; pdri mdm pdri me prajdm pdri 
nah pdhi ydd dhdnam (AV.), 'protect me, my progeny, and what wealth we 
own'; ydtah sadyd d ca para ca ydnti (AV.), 'from whence every day they 
advance and retire'; vy ahdrh sdrvena pdpmdnd [avrtam] vf ydksmena sdm 
dyusd (AV. ), 'I have separated from all ill-luck, from disease, [I have joined 
myself] with life'. 

The separation of the prefix from a verbal noun or adjective is very 
much more difficult, and of quite rare occurrence. 

1082. As regards the accent of verb-forms compounded 
with prefixes, only the case needs to be considered in which 
the prefix stands (as always in the later language) immediately 
before the verb ; otherwise, verb and prefix are treated as two 
entirely independent words. 

1083. A personal verbal form, as has been seen above 
(592), is ordinarily unaccented : before such a form, the prefix 
has its own accent ; or, if two or more precede the same form, 
the one nearest the latter is so accented, and the others lose 
their accent. 

If, however, the verb-form is accented, the prefix or pre- 
fixes lose their accent. 

That is, in every case, the verb along with its normally 
situated prefix or prefixes so far constitutes a unity that the 
whole combination is allowed to take but a single accent. 

Examples are: pare J hi nari ptinar e 'hi ksiprdm (AV.), 'go away, 

woman ; come again quickly'; dthd J stam vipdretana (RV.), 'then scatter ye 

away to your home'; samdcinusvd 'nusamprdyahi (AV.), 'gather together, go 

forth together after'; ydd grhdn upodatti (AV.), when he goes up to the 

Whitney, Grammar. 23 


house 1 ; evd ca tvdrh sarama ajagdntha (RV.), 'now that you, Saraina, have 
thus come hither 1 ; yend "vistitah pravivefithd 'pah (RV.), 'enveloped in which 
thou didst enter the waters'. 

1084. A prefix, however, not seldom has a more inde- 
pendent value, as a general adverb of direction, or as a preposition 
(in the usual modern sense of that term), belonging to and 
governing a noun ; in such case, it is not drawn in to form 
part of a verbal compound, but has its own accent. The two 
kinds of use shade into one another, and are not divisible by 
any distinct and fixed line. 

1085. In combination with the non-personal parts of the 
verb-system, with participles, infinitives, and gerunds, the general 
rule is that the prefix loses its accent, f in favor of the other 
member of the compound. But the prefix instead has sometimes 
the accent : namely, when combined 

a. with the passive participle in ta or na : thus, pdreta, 
'gone forth'; antdrhita, 'concealed'; dvapanna, 'fallen'; sdmpurna, 

b. with the infinitive in tu (972), in all its cases : thus., 
sdmhartum, 'collect'; dpidhatave, "to cover up'; dvagantos, 'of 
descending'; the doubly accented dative in tavai retains its final 
accent, but throws the other back upon the prefix : thus, dnvetavai, 
'for following'; dpabhartavai, 'for carrying off. 

1086. The closeness of combination between the root and 
the prefix is indicated not only by their unity of accent, but 
also by the euphonic rules (e. g. 185, 192), which allow the 
mutual adaptations of the two to be made to some extent as if 
they were parts of a unitary word. 

1087 . A few special irregularities call for notice : 

a. In the later language, opi, adhi, and ava, in connection with certain 
roots, sometimes lose their initial vowel: namely, api with nah and dha, 
adhi with stha, ava with gah. In the Veda, on the other hand, is is in a 
few cases found instead of nis with ykr, 

b. The final vowel of a prefix, especially an i, is (oftenest in the older 
language) sometimes lengthened, especially in derivative words : thus, pratlkara, 
nivrt, parihara, virtidh, adhlvasd, dpivrta, abhlvartd anurudh ; pravrs, upavasu. 

In the Veda, the initial of arm is sometimes lengthened after negative 
an : thus, ananuda. 

C. In combination with yi 'go', the prefixes para, part, and pra sometimes 
change their r to i. 

In this way is formed a kind of derivative stem palay, 'flee', inflected 
according to the a -class, in middle voice, which is not uncommon from the 
Brahmanas down, and has so lost the consciousness of its origin that it takes 


the augment prefixed : thus, apalayisthas ; it makes the periphrastic perfect 
palayam cakre. The stem palyay, similarly inflected, seems to occur only in 
B.; and play has been found nowhere except in MS. 

d. The root kr 'make 1 sometimes assumes (or retains from a more original 
condition) an initial s after the prefixes sam, pari, and upa: thus, samskurute, 
samaskurvan, sarhskrta, pariskrnvanti, pariskrta, upaskrta. And ]/fcr 'scatter' 
is said by the grammarians to add s in the same manner, under certain 
circumstances, after upa and prati. 

e. The passive participle of the roots da 'give' and da 'cut' has often 
the abbreviated form tta after a prefix of which the -final vowel, if i, is 
lengthened (compare the similar contraction with other elements, above, 955 c, 
and that of the derivative in ti, below, chap. XVII.). Thus, in AV. are found 
from da 'give', vydtta, pdritta, dpratitta : in Brahmanas, from the same, atta, 
pratta, apatta ; from da 'cut', dvatta, nirdvatta, samdvatta. 

f. The AB. has once niniyoja (instead of ni-yuyoja] from ni -\~yyuj, 
and udaprapatat; and in MBh. are found a few cases like vivyasa from 
vi -\~yas (where, however, an intentional play on the word may be assumed: 
Delbriick); and anvasarhcarat (instead of anusamacarat) . Such unifications of 
prefix and root, with treatment of the result after the manner of a simple 
root, are extremely rare. 

Some hold, however, that certain of the apparent roots of the language 
are results of this unification: thus, dp from a-\-ap, vyac from in-j-ac, tyaj 
from ati-(-aj, etc. (see Weber, Ind. Stud., xiii. 61). 

g. The loss of the initial s of stha and stambh after the prefix ud has 
been noticed above (233 a). 

Also (137 a, b), certain peculiarities of combination of a prefix with the 
initial vowel of a root. 

1088. As to the more general adverbial uses of the pre- 
fixes, and their prepositional uses, see the next chapter. 

1089. The adverbial prefixes sw, 'well', and dus, 'ill', are said to be 
sometimes combined with verbal forms ; but no examples of such combination 
are quotable from accentuated texts. 

As to the addition of the comparative and superlative suffixes taram and 
tamam to verbs, see above, 473. 

Other Verbal Compounds. 

1090. It has been seen above that some of the preposi- 
tional prefixes are employed in combination with only very small 
classes of roots, namely those whose meaning makes them best 
fitted for auxiliary and periphrastic uses such as kr, 'make', 
bhu and as, 'be', dha, 'put', i, 'go' and that the first three 
of these are widely used in combination with a derivative in 
am to make a periphrastic conjugation. Such roots have also 
been, from the earliest period of the language, but with increas- 
ing frequency, used in somewhat analogous combinations with 



other elements, substantive and adjective as well as adverbial ; 
and this has become, in part, developed finally into a regular 
and indefinitely extensible method of increasing the resources of 
verbal expression. 

1091. Most analogous with hin -\~ykr (1079) are a few other onomato- 
poetic compounds in the Veda: akkhallkftya (RV.), 'making a crackling sound', 
janjanabhdvant (RV.), 'flimmering', and alalabhdvant (RV.), 'making merry'; 
and (in AB.) bababakurvan, 'crackling'. And AV. has masmasd with ykr 
(TS. and VS., masmasd), 'crush'. 

Further, combinations of ykr with terms used at the sacrifice, as vdsat, 
svdha, svadhd, svaga: at first phrases only, and noun-compounds, but 
becoming -verbal combinations in which the prefixed word is treated like a 
prepositional prefix: thus, svagakardti (B.: but svadhd kartiti, TA.); and 
other prefixes are set before them, as anuvasatkuryat. 

1092. a. The noun namas, 'obeisance, homage', in a still more purely 
noun-value, becomes combined with ykr: in the Veda, only with the gerund, 
in namaskrtya (beside hastagfhya and karnagfhya: above, 990.). 

b. A solitary combination with yi, 'go', is shown by the accusative 
dstam, 'home'; which, appearing only in ordinary phrases in RV., is in AV. 
compounded with the participles in astamydnt, astamesydnt, dstamita 
(with accent like that of ordinary compounds with a prefix) and in the 
Brahmanas and the later language is treated quite like a prefix : thus, 
astameti (QB.). 

c. Other ordinary accusative forms of adjectives in combination with 
verbal derivatives of kr and bhu are found here and there in the older lan- 
guage : thus, prtamkftya and nagnamkrtya (TS.); nagnambhdvuka, pama- 
nambhdvuka, etc. (TS. et al.). 

1093. In the early but not in the earliest language, a noun- 
stem thus compounded with kr or bhu, in verbal nouns and 
ordinary derivatives, and then also in verbal forms, begins to 
assume a constant ending i (of doubtful origin). 

There is no instance of this in RV., unless the I of akkhallkftya (above, 
1091) is to be so explained. In AV., besides the obscure vatikrta and 
vaiikard, is found only phallkdrana. In the Brahmana language, examples 
begin to occur occasionally cyetl and mithum in TS.; these and phali, 
kruri, udvasl in TB.; the first three, with suphali, ekl, sin, brahmam, and 
daridri in QB.; and so on. The accent of the combination is in general 
accordance with the accent of compounds with the usual prefixes; and if the 
prefixed stem takes the tone, this rests upon the final i. Sometimes a mere 
collocation takes place: thus, mithum bhdvantis (TS.), phali kriydmananam 
(TB.). vajri bhutvd (TA.). The I is variously treated : now as an uncombinable 
final, as in fyeti akuruta and mithuni abhavan (TS.); now as liable to the 
ordinary conversions, as in mithuny enaya syam and svyakurvata (QB.). 

The examples in accentuated texts, and especially those in which the 


verb is entitled to the tone, are too few to furnish more than a fragmentary 
illustration of the formation. 

Out of such beginnings has grown in the later language the following 
rule : 

1094. Any noun or adjective stem is liable to be com- 
pounded with verbal forms or derivatives of the roots ]/5fi 
kr and ^ bhu (and, it is said, of ETH as also ; but such ca- 
ses, if they occur, are at least extremely rare), in the man- 
ner of a verbal prefix. If the final of the stem be an a or 
i-vowel, it is changed to ^ l\ if an w-vowel, it is changed 
to ^37 u. 

It is prescribed also that a final r become n, and that as 
and an be changed to t; but no genuine examples appear to be 

Examples are : stambhibhavati, 'becomes a post'; ekacittlbhuya, 'becoming 
of one mind'; upaharikarosi, 'thou makest an offering'; nakhapraharajarjankrta, 
'torn to pieces with blows of the claws'; fithilibhavanti, 'become loose'; 
kundallkrta, 'ring-shaped.' 

1095. Of all the forms which constitute or are attached to 
the verbal system, the passive participle is the one most closely 
assimilated in its treatment as a combinable element to an or- 
dinary adjective. Next to it come the gerund and the gerund- 
ives. Combinations of the kind here treated of are especially 
common with passive participles and gerunds. 



1096. THE indeclinable words are less distinctly divi- 
ded into separate parts of speech in Sanskrit than is usual 
elsewhere in Indo-European language especially owing 
to the fact that the class of prepositions hardly has a real 
existence, but is represented by certain adverbial words 
which are to a greater or less extent used prepositionally. 
They will, however, be briefly described here under the 
usual heads. 



1097. Adverbs by Suffix. Classes of adverbs, some- 
times of considerable' extent, are formed by the addition 
of adverb-making suffixes especially to pronominal roots or 
stems, but also to noun and adjective stems. 

There is no ultimate difference between these suffixes and the case- 
endings in declension ; and the adverbs of this division sometimes are used 
in the manner of cases. 

1098. With the suffix tas are made adverbs having an 
ablative sense, and not rarely also an ablative construction. Such 
are made : 

a. From pronominal roots, in dtas, itds, tdtas, ydtas, ktitas, amutas, 
svatas (not found earlier); and from the pronominal stems in t or d (494) of 
the personal pronouns ; thus, mattds (only example in V.), tvattas, asmattas, 

b. From noun and adjective stems of every class, since the earliest 
period, but more freely later : thus, mukhatds, agratds, rbhutds, hrttas, 
Cirsatds, nastds, yajustas, pardtas, anydtas, sarvdtas, daksinatds, abhipatds 
(once, in RV., from a case-form: patsutds}. 

c. From a few prepositions : thus, abhttas, parttas, dntitas. 
Examples of ablative construction are : tdtah sasthdt (AV.), 'from that 

sixth'; tdto jydyan (AV.), 'older than they'; kuta? cid defad agatya (H.), 
'arriving from some region or other'. 

But the distinctive ablative meaning is not infrequently effaced, and the 
adverb has a more locative value : thus, agratds, 'in front' ; asmatsamlpatas, 
'in our presence'; dharmatas, 'in accordance with duty'; chagatas (H.), 'with 
reference to the goat 1 . 

1099. With the suffix tra (in V. often tra) are made 
adverbs having a locative sense, and occasionally also a locative 

These adverbs are formed from pronominal roots, namely dtra, tdtra, 
ydtra, ktitra, amtitra, asmatrd, satrd; and also from noun and adjective 
stems, as anydtra, vi$vdtra, samandtra, martyatrd, daksinatrd, devatrd, 
purutrd, bahutrd. 

The words in (accented) trd are Vedic only, except satrd, 'altogether' (of 
which satram is also given as an alternative form). 

Examples of quasi-locative or locative construction are : hdsta d daksinatrd 
(RV.), 'in the right hand'; tatra 'ntare (H.), 'in that interval'; prabhutvam tatra 
yujyate (H.), 'sovereignty is suited to him'; ekatrapuruse (MBh.), 'in a single man'. 

As the locative case is used also to express the goal of motion (304), 
so the adverbs in tra have sometimes an accusative as well as a locative 
value : thus, tatra gacha, 'go there or thither'; pathd devatrd ydnan (RV.), 
'roads that go to the gods'. 


1100. One or two other suffixes of locality are : 

a. ha, in zM, 'here', k-'tha, 'where? 1 and the Vedic vipvdlia (also vifvdhd, 
vifvcihd), 'always' (compare below,",! 104, end). 

b. tat, which is added to words having already a local or directive 
value: thus, to adverbial accusatives, as prdktdt, tidaktdt; to adverbial abla- 
tives, as drattdt, uttarattdt, pardkattdt; and to prepositional adverbs, as 
adhdstdt, avdstdt, pardstdt, purdstdt, bahistdt. Apparently by analogy with 
these last, the suffix has occasionally the form stdt : thus, updristat. 

C. hi, in uttarahi (QB.) and daksinahi (no occurrence). 

1101. By the suffix tha are made adverbs of manner, es- 
pecially from pronominal roots or stems. 

Thus, tdthd, ydthd; katha and ittha (by the side of which stand kathdm 
and itthdm}', and the rare imdthd and amuthd. And dtha (V. often dthd], 'so 
then', doubtless belongs with them. Further, from a few adjective and noun 
stems : as, vifvdthd, sarvdthd, anydthd, ubhaydthd, itardtha, yatamdthd, 
urdhvdthd, rtutha, ndmdthd (once, AV.). 

Yatha becomes usually toneless in V., when used in the sense of iva 
after a noun forming the subject of comparison: thus, tdydvo yathd (RV.), 
'like thieves'. 

1102. One or two other suffixes of manner are: 

a. ti, in iti, 'thus', very commonly used, from the earliest 
period, as particle of quotation, following the words quoted. 

Examples are : brahmajdye ''yam iti ced dvocan (RV.), 'if they have said 
"this is a Brahman's wife"'; tdrh devd abruvan vrdtya kim nu tisthasi 'ti 
(AV.), 'the gods said to him: "Vratya, why do you stand?" 1 Often, the iti 
is used more pregnantly : thus, yah fraddddhdti sdnti devd iti (AV.), 'whoever 
has faith that the gods exist'; tarn vydghram munir musiko 'yam iti pa?yati 
(H.), 'the sage looks upon that tiger as being really a mouse'; yuyarh kim 
iti sidatha (H.), 'why (lit'ly, alleging what reason) do you sit?' Or the iti 
marks an onomatopoeia, or indicates a gesture : thus, bahfs te astu bal iti 
(AV.), 'let it come out of you with a "splash"'; ity dgre krsaty dthe 'ti (^B.), 
'he ploughs first this way, then this way'. A word made by iti logically 
predicate to an object is usually nominative: thus, svargo lokd iti yarn vddanti 
(AV.), 'what they call "the heavenly world"'; vidarbharajatanaydm damayanti 
'ti viddhi mam (MBh.), 'know me for the Vidarbha-king's daughter, Damayanti 
by name'; but ajnam bdlam ity dhuh (M.), 'they call an ignorant man a child'. 

With the suffix of iti is to be compared that of tdti etc. (519). 

b. va in iva (toneless), 'like, as', and evd (in V. often evd\ earlier 
'thus', later a particle emphasizing the preceding word; for 'thus' is used 
later the related evam, which hardly occurs in RV., and in AV. only with 
yvid : as, evdm vidvdn, 'knowing thus'. 

In later Vedic (AV. etc.) iva more often counts for only a single syllable. 

1103. a. By the suffix da are made adverbs of time, but 
almost only from pronominal roots. 


Thus, tadd, yadd, kadd, idd (only in V.); and sdda, beside which is 
found earlier sddam. Besides these, in the older language, only sarvadd ; 
later a few others, as anyada, ekada, nityadd. 

b. By the perhaps related danlm are made iddnlm, taddnlm, vifvaddriim. 

C. With rhi are made, from pronominal roots, tdrhi, etdrhi, ydrhi, kdrhi, 

d. The suffix di, found only in yddi, 'if, is perhaps related with da, 
in form as in meaning. 

1104. By the suffix dha are formed adverbs especially from 
numerals, signifying '-fold, times, ways', etc. 

Thus, ekadhd, dvidhd (also dvfdha and dvedhd), trfdhd (in the old lan- 
guage usually tredhd), saddhd (also sodhd and saddhd), dvada^adhd, sahasradhd, 
and so on. Also, naturally, from words having a quasi-numeral character : 
thus, tatidha, bahudha, purudha, vifvddhd, facvadha, etdvaddha. In a very 
few cases, also from general noun and adjective stems : thus, mitradhd (AV.), 
priyadhd (TS.), rjudhd (TB.), paristubdhd (PB.). 

The particle ddha or ddha, a Vedic equivalent of dtha, probably belongs 
here (purudhd and vifvddha, with shortened final, occur a few times in RV.); 
also addhd, 'in truth'; and perhaps sahd, 'with', which has an equivalent 
sadha- in several Vedic compounds. And the other adverbs in ha (1100 a) 
may be of like origin. 

1105. From a few numerals are made multiplicative adverbs with s: 
namely, dvfs, frfs, and cattir (probably, for caturs). 

The corresponding word for 'once', safcft, is a compound rather than a 
derivative; and the same character belongs still more evidently to pancakftvas, 
navakftvas, aparimitakrtvas, etc., though krt and krtvas are regarded by the 
native grammarians as suffixes (AV. has dd$a krtvas and sapid krtvas). 

110 6. By the suffix cds are made, especially from numeral 
or quantitative stems, adverbs of quantity or measure or manner, 
generally used distributively. 

Thus, efcafas, 'one by one', patafds, 'by hundreds', riufds, 'season by 
season', pacchas, 'foot by foot', aksara^ds, 'syllable by syllable', ganafds, 'in 
crowds', stambafds, 'by bunches', pamppds, 'limb by limb', tdvacchds, 'in 
such and such number or quantity': and, in a more general way, sarva$dsi 
'wholly', mukhyafas, 'principally', krchra$as, 'stingily'. 

1107. By the suffix vat are made with great freedom, in 
every period of the language, adverbs signifying 'after the man- 
ner of, like', etc. 

Thus, angirasvdt, 'like Angiras', manusvdt (RV.), 'as Manu did', jamad- 
agnivdt, 'after the manner of Jamadagni'; purvavdt or pratnavdt or purdnavdt, 
'as of old', kakatallyavat, 'after the fashion of the crow and the palm-fruit'. 

This is really the adverbially used accusative (with adverbial shift of 
accent : below, 1 1 1 1 e) of the suffix vant (next chapter), which in the Veda 
makes certain adjective compounds of a similar meaning : thus, tvavant, 'like 
thee', mdvant, 'of my sort', etc. 


1108. By the suffix sat are made adverbs signifying 'into the condition 
of, which are used along with verbs of becoming and of making. 

Thus, agnisat -f- )//cr, 'reduce to fire, burn up'; bhasmasat -f- ybhu, 'turn 
to ashes'; atmasatkrta, 'made one's own'. 

These derivatives are unknown in the earlier language, and not common 
in the later. The s of sat is not liable to conversion into s. The connection 
with the following verb is not so close as to require the use of the gerund 
in ya: thus, bhasmasatkrtva (not -krtya: above, 990). 

1109. Suffixes, not of noun-derivation or inflection, may be traced with 
more or less plausibility in a few other adverbs. Thus, for example, in 
pratdr, 'early', and sanutdr, 'away'; in daksintt, 'with right hand', and cikitvtt, 
'with consideration'; in nwndm, 'now', and nandndm, 'variously'. But the 
cases are in the main too rare and doubtful to be worth notice here. 

The adverbs of this division are almost never used prepo- 
sitionally. Those of the next division, however, are in many 
instances so used. 

1110. Case-forms used as Adverbs. A large num- 
ber of adverbs are more or less evidently cases in form, 

made from stems which are not otherwise in use. Also 


many cases of known stems, pronominal or noun or adject- 
ive, are used with an adverbial value,, being distinguished 
from proper cases by some difference of application, which 
is sometimes accompanied by an irregularity of form. 

1111. The accusative is the case most frequently and 
widely used adverbially. Thus: 

a. Of pronominal stems: as, t/ad, 'if, when, that', etc.; tad, 'then' etc.; 
fcfm, 'why, whether', etc.; iddm, 'now, here'; adds, 'yonder'; and so on. 
Of like value, apparently, are the (mostly Vedic) particles kdd, kdm and 
kam(?), id, cid (common at every period), smdd and sumdd, im and slm 
(by some regarded as still possessing pronoun-value), -kirn. 

Compounds with fd are ced, 'if, ncd, 'lest', svid, kuvfd ; with cid, 
kucid ; with -kim, ndklm and makim, and aklm. 

b. Of noun-stems : as, noffto, 'by name'; stikham, 'happily'; kdmam, 
'at will, if you please'; n'iktam, 'by night'; raftas, 'secretly'; osdm, 'quickly' 
(V.); and so on. 

C. Of adjective stems, in great numbers : as, satydm, 'truly'; cirdm, 
'long'; nttyam, 'constantly'; bh&yas, 'more, again'; and so on. 

1. The neuter singular is the case commonly employed in this way; 
and it is formed and used adverbially from a large class of compound stems 
which do not occur in adjective use (the so-called avt/at/i&/iaua-compounds : 
below, chap. XVIII.). 


2. But the feminine singular also is sometimes used; especially in the 
so-called adverbial endings of comparison, taram and tamam, which are 
attached to particles, and even, as it is claimed (473), to verb-forms : thus, 
pratardm, pratamdm, uccaistaram, fanaistaram, jyoktamdm. 

In the oldest language (RV. and AV.), the neuter instead of the feminine 
form of these suffixes is almost alone in use: see 1119. 

d. Many adverts of obscure form or connection are to be explained with 
probability as accusatives of obsolete noun or adjective stems: examples are 
tusnim, 'in silence'; saydm, 'at evening'; dlakam, 'vainly'; sakdm, 'together, 
with (prep.)'; dram or atom, 'sufficient' (in the later language used with ykr 
in the manner of a prefix); prayas, 'usually'; Isdt, 'somewhat'; amnds, 'un- 
expectedly'; bahfs, 'outside'; mtthu and mithds, mtihu and muhus, jdtu, 
and so on. Madrtk etc., and ninfk (in RV.), are perhaps contracted forms 
of adjectives having j/oe or anc as their final (407 ff.); and visundk and 
vfthak, with prthak and fdhak, may be of the same character. The presence 
of other roots as final members is also probable for u^ddhak, anusdk and 
dyusdk, anusthu and susthu, yugapdt, etc. Compare also the forms in am 
beside those in 5, above, 1099, 1101, 11 03 a. 

e. In (Vedic) dravdt, 'quickly', is to be seen a change of accent for 
the adverbial use (pple drdvant, 'running'); and drahydt, 'stoutly' (RV., once), 
may be another example. The comparative and superlative suffixes (above, c) 
show a like change ; and it is also to be recognized in the derivatives with 
vdt (1107). 

1112. The instrumental is also very often used with 
adverbial value : generally in the singular, but sometimes also 
in the plural. Thus : 

a. Of pronominal stems : as, end and ayd, kdyd, and, sdna, amd, 

b. Of noun-stems : as, ksanena, 'instantly'; afesena, 'completely'; vifesena, 
'especially'; dtva, 'by day'; distyd, 'fortunately'; sdhasd, 'suddenly'; aktubhis, 
'by night'; and so on. 

C. Of adjectives, both neuter (not distinguishable from masculine) and 
feminine : as, ddksinena, 'to the south'; uttarena, 'to the north'; antarena, 
'within'; cirena, 'long'; fdnais and fdnakais, 'slowly'; uccafs^ 'on high'; 
paracafs, 'afar'; tdvisibhis, 'mightily'; and so on. 

d. More doubtful cases, mostly from the older language, may be in- 
stanced as follows : tirafcdta, devdta, bahtita, and sasvdrta (all RV.), hom- 
onymous instrumental from nouns in ta; dvitd, tadttna, irmd, mrsd, vftha, 
sdca, asthd(?), mudha (not V.), adhund (Br. and later). 

e. Adverbially used instrumental are (in the older language), oftener 
than any other case, distinguished from normal instrumentals by differences 
of form: thus, especially, by an irregular accent: as, amd and dfva (given 
above); perhaps guha; apakd, asayd, kuhaydf?); naktayd, svapnayd, samand 
adatrayd, rtayd, ubhayd, sumnaydf?); daksind, madhyd ; nicd~, pracd, uccd, 
paced, tirafcd; in a few w-stems, by a y inserted before the ending, 




which is accented : thus, amuyd (given above), acuyd, sadhuyd, raghuyd, 
dhrsnuya, anusthuyd, mithuyd ; and urviyd (for urvyd) and vfyvya (properly 
vfcvayd] are more slightly irregular. 

1113. The dative has only very seldom an adverbial use. 
Examples are (from the later language only): ciraya, 'long'; arthaya, 

'for the sake of; ahnaya, 'presently'. 

1114. The ablative is not infrequently used adverbially. 
Thus : 

a. Of pronominal stems: as, kdsmat, 'why?' akasmat, 'casually, un- 
expectedly'; dt, tdt, ydt (V.: normal forms, instead of the pronominal 
asmat etc.). 

b. Of noun-stems: as, asdt, 'near'; ardt, 'afar'; balat, 'forcibly'; kutu- 
halat, 'emulously'; sakafat, 'on the part of. 

C. Oftenest, of adjective stems : as, durdt, 'afar'; nlcdt, 'below'; pafcdt, 
'behind'; saksdt, 'plainly, actually' ; samantat, 'completely'; acirat, 'not long'. 

d. In a few instances, adverbially used ablatives likewise show a changed 
accent in the early language : thus, apakdt, 'from afar'; amdt, 'from near 
by 1 ; sandt, 'from of old' (but instr. sdna); uttardt, 'from the north'; adhardt, 

1115. The genitive is almost never used adverbially. 

In the older language occur akttis, 'by night', and vdstos, 'by day'; 
later, cirasya, 'long'. 

1116. The locative is sometimes used with adverbial 
value . Thus : 

From noun and adjective stems: tike, 'near'; are and dure, 'afar'; 
abhisvare, 'behind'; astamlke, 'at home'; rt", 'without' (prep.); dgre, 'in 
front'; sthane, 'suitably'; sapadi, 'immediately'; -arthe and -krte (common in 
composition), 'for the sake of"; aparisu, 'in after time'. 

1117. Even a nominative form appears to be stereotyped into an ad- 
verbial value in (Vedic) fefs, interrogative particle, and its compounds ndkis 
and mdkis, negative particles. 

1118. Verbal Prefixes and kindred words. The 
verbal prefixes, described in the preceding chapter (1076 ff.), 
are properly adverbs, having a special office and mode of 
use in connection with verbal roots and their more imme- 
diate derivatives. 

Their occasional looser connection with the verb has been 
noticed above (1084). In the value of general adverbs, how- 
ever, they only very rarely occur (except as dpi has mainly 
changed its office from prefix to adverb or conjunction in the 
later language) ; but their prepositional uses are much more fre- 
quent and important: see below, 1125. 


In composition with nouns, they (like other adverbial elements) not in- 
frequently have an adjective value: see below, chap. XVIII. 

1110. Several of the prefixes (as noticed above, 473 4) form com- 
parative and superlative adjectives, by the suffixes tara and tama, or ra and 
ma : thus, uttara and uttamd ; ddhara and adhamd, dpara and apamd, dvara 
and avamd, tipara and upamd ; and prathamd is doubtless of the same char- 
acter; also, dntara and dntama. And accusatives of such derivative adjectives 
(for the most part not otherwise found in use) have the value of comparatives, 
and rarely superlatives, to the prefixes themselves: thus, sdrhfitam cit 
sarhtardm sdrh fifadhi (AV.), 'whatever is quickened, do thou still further 
quicken'; vitardrh v{ kramasva (RV.), 'stride out yet more widely'; prd tdrh 
naya pratardm vdsyo dcha (RV.), 'lead him forward still further toward ad- 
vantage'; ud enam uttardrh naya (AV.), 'lead him up still higher'. 

Besides those instanced, are found also nitardm, avatardm, paratardm, 
parastardm. In the Brahmanas and later (above, 11 lie), the feminine 
accusative is used instead: thus, pratitardm, sarhtardm, nitardm, uttardm, 
pratardm and pratamdm (and sarhtardm, RV., once). 

1120. Kindred in origin and character with the verbal pre- 
fixes, and used like them except in composition with verbs, are 
a few other adverbs: thus, avds, 'down'; adhds, 'below'; paras, 
'far off"; pura, 'before'; antara (apparently, antdr-\-a), 'among'; 
dnti, 'near'; updri, 'above': and sahd (already mentioned, 1104), 
'along, with', and sdca, 'together, with', may be noticed with them. 
Vina, 'without', and visu-, 'apart', appear to be related with vi. 

1121. Inseparable Prefixes. A small number of 
adverbial prefixes are found only in combination with other 
elements. Thus : 

a. The negative prefix a or an an before vowels, a before 

It is combined especially with innumerable nouns and adjectives ; much 
more rarely, with adverbs, as aktitra and dpunar (RV.), dnadhas (TB.), akas- 
mat, asakrt; and, according to the grammarians, sometimes also with pronouns 
(asas, anesas), and with verbs (apacati, 'does not cook'), but no such com- 
binations appear to be quotable. 

The independent negative adverbs, nd and md, are only in rare and 
exceptional instances used in composition: see below, 1122b. 

b. The comitative prefix sa, used instead of the preposition 
sdm, and exchangeably with sahd, before nouns and adjectives. 

c. The prefix of dispraise dus, 'ill, badly' (identical with 
ydus: 225). 

It is combined in the same manner as a or an. Of combinations with 
a verbal form, at least a single example appears to be quotable^ dupcarati 
(B.), 'behaves ill' (BR.). 

1122] ADVERBS. 365 

d. The corresponding laudatory prefix *M, 'well', is in gen- 
eral so closely accordant in its use with the preceding that it is 
best mentioned here, through it occurs not rarely as an inde- 
pendent particle in the oldest language (in RV. , more than two 
hundred times ; in the peculiar parts of AV. , only fourteen 
times], and even occasionally later. 

No combination of su with a verbal form appears to be quotable from 
any accentuated text (though the worthless pada-text of AV. xix. reads 
su-dpdyati at 49. 10). K. has na su vijnayete and na vai su viduh (or su- 
vijnayete ?). 

e. The exclamatory and usually depreciative prefixed forms of the inter- 
rogative pronoun (506) are most analogous with the inseparable prefixes. 

1122. Miscellaneous Adverbs. Other words of 
adverbial character and office, not clearly referable to any 
of the classes hitherto treated, may be mentioned as follows: 

a. Asseverative particles (in part, Vedic only): thus, angd, 
hdnta, kila, khdlu, tu (rare in older language), vai, vavd (in 
Brahmana language only), hi, hind, u, aha, ha, aha, samaha, sma, 

Of these, hdnta is a word of assent ; hi has won also an illative meaning, 
and accents the verb with which it stands in connection (595 d); sma some- 
times gives a past meaning to a present tense (778 b); u is often combined 
with the final a of other particles: thus, dtho, no, md, uto, upo, pro; but 
also with that of verb-forms, as datto, vidmti. The final o thus produced 
is pragrhya or uncombinable (138c). 

Particles of kindred value, already mentioned above, are fd, kdm or kam, 
cid, jdtu, evd. 

Some of the asseverative particles are much used in the later artificial 
poetry with a purely expletive value, as devices to help make out the metre 
(padapurana, 'verse-fillers'); so especially ha, hi, tu, sma. 

b . Negative particles are : nd, signifying simple negation ; 
ma, signifying prohibition. 

As to the construction of the verb with md, see above, 579 80. 

In the Veda, nu (or nu : 248 a) has also sometimes a negative meaning. 
For the Vedic nd of comparison, see below, d. 

In nahf, nd is combined with M, both elements retaining their full 
meaning; also with fd in ned. 'lest'. It is perhaps present in nanti and 
cand, but not in hind (RV., once). In general, neither nd nor md is used 
in composition to make negative compounds, but, instead, the inseparable 
negative prefix a or an (1121 a): exceptions are the Vedic particles ndkis 
and mdkis, ndkim and makim ; also naciram and maciram, and a few others. 

c. Interrogative particles are only those already given : kdd, 
Urn, kuvid, svid, nanu, of which the last introduces an objection 
or expostulation. 


d. Of particles of comparison have been mentioned the 
toneless iva, and yatha (also toneless when used in the same 
way). Of frequent occurrence in the oldest language is also nd, 
having (without loss of accent) the same position and value as 
the preceding. 

Examples are : rsidvfsa {sum nd srjata dvfsum (RV.), 'let loose your 
enmity like an arrow at the enemy of the singer'; vdyo nd vrksdm (AY.), 
<as birds to the tree'; gaur6 nd trsitdh piba (RV.), 'drink like a thirsty 
buffalo'. This use is generally explained as being a modification or adaptation 
of the negative one: thus, '[although, to be sure] not [precisely] a thirsty 
buffalo'; and so on. 

e. Of particles of place, besides those already mentioned, 
may be noticed kva (in V., always to be read Ma). 

f . Particles of time are : nu, 'now' (also nu : nundm was 
mentioned above, 1109) adyd and sadyds and sadivas (RV., once), 
'today, at once' (all held to contain the element div or dyu), 
hyds, 'yesterday', qvds, 'tomorrow', jyok (also related with dyu), 
'long'; punar, 'again'. 

g. Of particles of manner, besides those already mentioned, 
may be noticed nuna, 'variously' (for nanandm, its derivative, 
see 1109); sasvdr (RV.), 'secretly'. 

In the above classifications are included all the Vedic adverbial words, 
and most of those of the later language : for the rest, see the dictionaries. 


1123. There is, as already stated, no proper class of 
prepositions (in the modern sense of that term), no body of 
words having for their exclusive office the "government" of 
nouns. But many of the adverbial words indicated above 
are used with nouns in a way which approximates them 
to the more fully developed prepositions of other languages. 

If one and another of such words as vina, rte occurs almost solely 
in prepositional use, this is merely fortuitous, and of no consequence. 

1124. Words are used prepositionally along with all the 
noun-cases excepting the dative. But in general their office is 
directive only, determining more definitely, or strengthening, the 
proper case-use of the noun. Sometimes, however, the case- 
use is not easy to trace, and the noun then seems to be more 
immediately "governed" by the preposition that is, to have 
its case-form more arbitrarily determined by its association with 
the latter. This is oftenest true of the accusative ; and of the 

1128] PREPOSITIONS. 367 

genitive, which, has, here as elsewhere (294), suffered an exten- 
sion of its normal sphere of use. 

1125. The adverbs by derivative form (1097 ff.) have least 
of a prepositional value (exceptions are especially a few made 
with the suffix tas : 1098). 

Most of the verbal prefixes (exceptions are ud, ni, para, 
pra: and ava and ni are almost such) have their prepositional 
or quasi-prepositional uses with cases ; but much more widely in 
the older time than in the later : in the classical language the 
usage is mainly restricted to prati, anu, and a. 

Most of the directive words akin with the more proper pre- 
fixes are used prepositionally : some of them as saha, vina, 
upari, antara, pur a freely, earlier and later. 

The case-forms used adverbially are in many instances used 
prepositionally also : oftenest, as was to be expected, with the 
genitive; but frequently, and from an early time, with the accus- 
ative ; more rarely with other cases. 

We will take up now the cases for a brief exposition, beginning with 
those that are least freely used. 

1126. The Locative. This case is least of all used with words that 
can claim the name of preposition. Of directives, antdr and its later deriv- 
ative antard, meaning 'within, in', are oftenest added to it, and in the 
classical language as well as earlier. Of frequent Vedic use with it are a 
and adhi (illustrated above, 305); apt and upa are much rarer: thus, 
yd apdm dpi vrate [sdnti] (RV.), 'who are in the domain of the waters'; 
amur yd Upa surye [sdnti] (RV.), 'who are up yonder in the sun'; sdca, 
'along with', is not rare in RV., but almost entirely unknown later: thus, 
pitrdh sdca sail, 'staying with her parents'. 

1127. The Instrumental. The directives used with this case are 
almost only those which contain the associative pronominal root sa : as saha 
(most frequent), sakam, sardham, samam, samaya, saratham- and, in the 
Veda, the prefix sam : as, te sumatlbhih sdm pdtnlbhir nd vrsano nasimahi 
(RV.), ''may we be united with thy favors as men with their spouses'. By 
substitution of the instrumental for the ablative of separation (283), vina, 
'without' (not Vedic), takes sometimes the instrumental ; and so, in the Veda, 
avas, 'down', and paras, 'beyond', with which the ablative is also, and much 
more normally, construed. And adhi, in RV., is used with the instrumental 
snuna and snubhis, where the locative would be expected. 

1128. The Ablative. In the prepositional constructions of the ab- 
lative (as was pointed out and partly illustrated above, 293), the ablative 
value of the case, and the merely directive value of the added particle, are 
for the most part clearly to be traced. Many of the verbal prefixes are more 
or less frequently joined in the older language with this case: oftenest, adhi 
and pari; more sporadically, anu, apa, ava, prati, and the separatives nis 
and vi. The change of meaning of the ablative with a, 'hither', by which 


it comes to fill the office of its opposite, the accusative, was sufficiently ex- 
plained above (293 c). Of directive words akin with the prefixes, many 
as bahis, puras, avas, adhas, paras, pwra, -uma, and tzras, 'out of knowledge 
of accompany this case by a perfectly regular construction. Also the 
case-forms arvak, prak, pafcdt, urdhvam, purvam, param, parena, prabhrti; 
and rte, 'without', of which the natural construction with an ablative is- 
predominant earlier. Antikam, 'near', is said to take the ablative as well as 
its more normal companion the genitive. 

1129. The Accusative. Many of the verbal prefixes and related words 
take an accompanying accusative. Most naturally (since the accusative i& 
essentially the 'to'-case), those that express a motion or action toward any- 
thing : as abhi, prati, arm, upa, a, ati and adhi in the sense of 'over on to* 
or 'across, beyond', tiras, 'through', antar and antard when meaning 'between', 
pan, 'around'. Examples are: ydh pradfyo abhf sUryo vicdste (AV.), 'what 
quarters the sun looks abroad unto'; dbodhy agnfli prdty dyatim usdsam (RV.), 
'Agni has been awakened to meet the advancing dawn'; gached kaddcit 
svajanam prati (MBh.), 'she might go somewhither to her own people'; imam 
praksydmi nrpatim prati (MBh.), 'him I will ask with reference to the king'; 
mama cittdm anu cittebhir e '<a (AV.), 'follow after my mind with your 
minds'; e J hy d nah (AV.), 'come hither to us'; tipa na e 'hy arvdn (RV.), 
'come hither unto us'; yd devomdrtydn ati (AV.), 'the god who is beyond mortals'; 
adhisthdya vdrcasd 'dhy anydn (AV.), 'excelling above others in glory'. Also 
abhitas and paritas, which have a like value with the simple abhi and pari ,- 
and upari : 'above' (oftener with genitive). Less accordant with ordinary- 
accusative constructions is the use of this case with adhas, paras, puras, 
vind, beside other cases which seem more suited to the meaning of those 
particles. And the same may be said of most of the adverbial case-forms 
with which the accusative is used. Thus, a number of instrumentals of 
situation or direction: as ye i varena "ditydm, ye pdrend JJ ditydm (TB.), 
'those who are below the sun, those who are beyond the sun'; dntarena 
ytinim (QB.), 'within the womb'; te hi J dam antarena sarvam (AB.), 'for all 
this universe is between them'; fittarena gdrhapatyam (B.), 'to the north of 
the householder's fire'; ddksinena vedim (^)B.), to the south of the sacrificial 
hearth'; daksinena vrksavdtikdm (Qak.), 'to the right of the orchard'; nikasd, 
'near to'. Similarly, urdhvam and purvam have an accusative object as well 
as an ablative; and the same is true later of rte. Abhimukham, 'toward', 
has a more natural right to construction with this case; and samdyd (later 
samayd], 'through between', is analogous with antard and tiras. 

1130. The Genitive. The words which are accompanied by the genitive 
are mostly case-forms of nouns, or of adjectives used substantively, retaining 
enough of the noun-character to take this case as their natural adjunct. 
Such are the locatives agre, 'in front of, abhydce, 'near', arthe and krte, 
'for the sake of, nimitte and hetdu, 'by reason of, madhye, 'in the midst 
of; and other cases, as artham and arthdya, antikam and abhimukham (which 
have also other constructions), kdrandt, sakdcdt, hetos. And really, although 



less directly and obviously, of the same character are other adjective cases 
(some of them showing other constructions, already noticed): as adharena, 
uttarena and uttarat, daksinena and daksinat, pafcat, urdhvam, anantaram, 
samaksam, saksat. More questionable, and illustrations rather of the general 
looseness of use of the genitive, are its constructions (almost wholly unknown 
in the oldest language) with more proper words of direction : thus, with the 
derivative paritas, paratas, and antitas, and parastat and purastat (these 
found in the Brahmana language : as, sarhvatsarasya parastat, 'after a year'; 
suktasya purastat, 'before the hymn' [AB.]); with anti, adhas, auas, puras ; 
with upari, 'above' (common later); and with antar. 


1131. The conjunctions, also, as a distinct class of 
words, are almost wanting. 

The combination of clauses is in Sanskrit in general of a 
very simple character ; much of what in other Indo-European 
languages is effected by subordinating conjunctions is here man- 
aged by means of composition of words, by the use of the 
gerunds (994), and of iti (1102 a), and so on. 

1132. The relative derivative adverbs, already given 
(1098 if.), may properly be regarded as conjunctions; and 
a few other particles of kindred value, as ced and ned (1111 a. 

1133. Purely of conjunctive value are tf c#, 'and', and 
SfT vcij 'or' (both toneless, and never having the first place 
in a sentence or clause). 

Of copulative value, along with ca, is in the older lan- 
guage especially uta (later it becomes a particle of more inde- 
finite use); and api, tatas, tatha, him ca, with other particles 
and combinations of particles, are used often as connectives of 

Adversative is tu, 'but' (rare in the older language); also r 
less strongly, u (toneless). 

Of illative value is hi, 'for' (originally, and in great part at 
every period, asseverative only): compare above, 1122 a. 

To ca (as well as to its compound ced} belongs occasionally the meaning 'ifV 
It is needless to enter into detail with regard to those uses which may 
be not less properly, or more properly, called conjunctive than adverbial, of 
the particles already given, under the head of Adverbs. 


1134. The utterances which may be classed as inter- 
jections are, as in other languages, in part voice-gestures, 

Whitney, Grammar. 24 


in part onomatopoeias, and in part mutilations and corrup- 
tions of other parts of speech. 

1135. a. Of the class of voice-gestures are, for example : 
a, ha, haha, ahaha, he, hai (AV.), ayi, aye, haye (RV.), aho, 
bat (RV.), bata (RV.) or vata, and (probably) hiruk and huruk (RV.). 

b. Onomatopoetic or imitative utterances are, for example 
(in the older language); cicca, 'whiz' (of an arrow : RV.); kikira 
(palpitation: RV.); bal and phat (phds ?) OT pMl, 'splash' (AV.); 
bhuk, 'bow-wow' (AV.); cdl, 'pat' (AV.); as, Ms, as, and has 
(PB.); and see the words already quoted in composition with 
the roots kr and bhu, above, 1091. 

c. Nouns and adjectives which have assumed an inter] ectional 
character are, for example : bhos (for the vocative bhavas, 456) ; 
are or re (voc. of ari, 'enemy'); dhik, 'alas!' (may be mere 
voice-gesture, but perhaps related with ydih); kastam, 'woe is 
me!' distya, 'thank heaven!' svasti, 'hail!' susthu, sadhu, 'good, 
excellent!' None of these are Vedic in interj ectional use. 



1136. THE formation from roots of conjugable stems 
namely, tense-stems, mode-stems, and stems of secondary con- 
jugation (not essentially different from one another, nor, it is 
believed, ultimately from the formation of declined stems) 
was most conveniently treated above, in the chapters devoted to 
the verb. Likewise the formation of adverbs by derivation (not 
essentially different from case-formation), in the chapter devoted 
to particles. And the formation of those declinable stems 
namely, of comparison, and of infinitives and participles which 
attach themselves most closely to the systems of inflection, has 
also been more or less fully exhibited. But the extensive and 
intricate subject of the formation of the great body of declin- 
able stems was best reserved for a special chapter. 

Of course, only a brief and compendious exhibition of the subject can 
be attempted within the here necessary limits: no exhaustive tracing out of 
the formative elements of every period; still less, a complete statement of 
the varied uses of each element; least of all, a discussion of origins: but 
enough to help the student in that analysis of words which must form a part 


of his labor from the outset, giving a general outline of the field, and pre- 
paring for more penetrating investigation. 

The material from accented texts, and especially the Vedic material, 
will be had especially in view (nothing that is Vedic being intentionally left 
unconsidered); and the examples given will be, so far as is possible, words 
found in such texts with their accent marked. No word not thus vouched 
for will be accented unless the fact is specifically pointed out. 

1137. The roots themselves, both verbal and pronom- 
inal, are used in their bare form, or without any added 
suffix, as declinable stems. 

As to this use of verbal roots, see below, 1147. 

The pronominal roots, so-called, are essentially declinable ; 
and hence, in their further treatment in derivation, they are in 
accordance with other declinable stems, and not with verbal 

1138. Apart from this, every such stem is made by a 
suffix. And these suffixes fall into two general classes: 

A. Primary suffixes, or those which are' added directly 
to roots; 

B. Secondary suffixes, or those which are added to de- 
rivative stems (also to pronominal roots, as just pointed out, 
and sometimes to particles). 

The. division of primary suffixes nearly corresponds to the krt (more 
regular) and unadi (less regular) suffixes of the Hindu grammarians ; the 
secondary, to their taddhita-snfftxes. 

1139. But this distinction, though one of high value, 
theoretically and practically, is not absolute. Thus: 

a. Suffixes come to have the aspect and the use of primary 
which really contain a secondary element that is to say, the 
earliest words exhibiting them were made by addition of second- 
ary suffixes to words already derivative. 

Sundry examples of this will be pointed out below: thus, the gerundival 
suffixes, tavya, anlya, etc., the suffixes uka and afca, tra, and others. This 
origin is probable for more cases than admit of demonstration ; and it is 
possible for others which show no distinct signs of composition. 

b. Less often, a suffix of primary use passes over in part 
into secondary, through the medium of use with denominative 
" roots" or otherwise: examples are yu, iman, lyas and istha. 

1140. Moreover, primary suffixes are added not only 


372 XVII. DERIVATION. [1140 

to original roots, but, generally with equal freedom, to ele- 
ments which have come to wear in the language the aspect 
of such, hy being made the basis of primary conjugation 
and even, to a certain extent, to the bases of secondary 
conjugation, the conjugation-stems, and the bases of tense- 
inflection, the tense-stems. 

a. The most conspicuous examples of this are the participles, present 
and future and perfect, which are made alike from tense and conjugation- 
stems of every form. The infinitives (968 ff.) attach themselves only in sporadic 
instances to tense-stems, and even from conjugation-stems are made but 
sparingly earlier; and the same is true of the gerundives. 

b. General adjectives and nouns are somewhat widely made from con- 
jugation-stems, especially from the base of causative conjugation : see below 
the suffixes a, a, as, anz, u, ti, tr, tnu, snu. 

c. From tense-stems, the examples are far fewer, but not unknown: 
thus, from present-stems, occasional derivatives in a (1148. 3b), a (1149), 
u (1178e,g,h), ta (1176e), tu (1161c), tra (1185d), ru (1192), vin (or 
in: 1232, end); from stems in a s apparently of aoristic character (besides 
infinitives and gerundives), occasional derivatives in a (1148. 3 b), ana 
(1150.2b), am (1159b), an (1160a), ana (1175), as (1151. 1 b), istha 
(1184), u (1178e), us (1154), tr (1182c), in (1183). 

1141. The primary suffixes are added also to roots as 
compounded with the verbal prefixes. 

Whatever, namely, may have been originally and strictly 
the mode of production of the derivatives with prefixes, it is 
throughout the recorded life of the language as if the root and 
its prefix or prefixes constituted a unity, from which a deriva- 
tive is formed in the same manner as from the simple root, with, 
that modification of the radical meaning which appears also in 
the proper verbal forms as compounded with the same prefixes. 

Not derivatives of every kind are thus made; but, in the 
main, those classes which have most of the verbal force, or 
which are most akin in value with infinitives and participles. 

The occurrence of such derivatives with prefixes, and their accent, will 
be noted under each suffix below. They are chiefly (in nearly the order of 
their comparative frequency), besides root-stems, those in a, in ana, in to', 
in tar and tra, and in in, ya, van and man, i and u, as, and a few others. 

1142. The suffixes of both classes are sometimes joined 
to their primitives by a preceding union-vowel that is to 
say, by one which wears that aspect, and, in our uncertainty as 
to its real origin, may most conveniently and safely be called 
by that name. The line between these vowels and such as de- 


serve to be ranked as of organic suffixal character cannot be 
sharply drawn. 

Each of the two great classes will now be taken up by 
itself, for more particular consideration. 

A. Primary Derivatives. 

1143. Form of root. The form of root to which a 
primary suffix is added is liable to more or less variation. 

a. By far the most frequent is a strengthening change, by 
guna or vnfr/^z-increment. The former may occur under all 
circumstances (except, of course, where ywmz-change is in general 
forbidden: 235. 240): thus, veda from yvid, moda from y mud, 
vdrdha from ]/ vrdh : ay ana from y i, sdvana from y su, sdrana 
from y sr ; and so on. But the latter is only allowed under 
such circumstances as leave long a as the resulting vowel : that 
is, with non-final #, and with a final i or w-vowel and r, before 
a vowel: thus, nadd from ynad, grabhd from y grbh or grabh, 
vahd from yvah, nay a from y m, bhavd from ybhu, hard from 
ykr; such strengthening as would make vaida and mauda is 
entirely unknown in primary derivation. 

No general rules can be given as to the occurrence of strengthening in 
derivation : it must be left to be pointed out for each suffix. 

b. Other vowel changes are more sporadic, and will be noticed in detail 
below: thus, occasional abbreviation of a root, as in ukti from yvac ; change 
of final r to ir and ur ; the loss of final a, or its conversion to an i-vowel ; 
and so on. 

c. The reversion of a final palatal or h to a guttural has been already 
noticed (216). A final n or m is occasionally lost, as in formations already 

d. After a short final vowel is sometimes added a t : namely, where a 
root is used as stem without suffix (1147), and before a following y or v 
of van (1169), vara and van (1171), yu once (1165 a), and ya (1213). 

The presence of t before these suffixes indicates an original secondary 
derivation from derivatives in ti and tu. 

e. The root is sometimes reduplicated : rarely in the use without suffix : 
oftenest before z, a, u; in only a few cases each, also before other suffixes 
(ana, vana, van and varl, vani, vi, vit, ani, in, aw, tnw, ia, ti, tr, tra, 
uka and Ifca, ma, ya}. 

1144. Accent. No general laws governing the place of 
the accent are to be recognized ; each suffix must in this 
respect be considered by itself. 

In connection with a very few suffixes is to be recognized a certain 


degree of tendency to accent the root in case of a nomen actionis or infini- 
tival derivative, and the ending in the case of a nomen agentis or participial 
derivative: see the suffixes a, ana, as, an, and man, below, where the ex- 
amples are considered. Differences of accent in words made by the same 
suffix are also occasionally connected with differences of gender : see the 
suffixes as and man. 

1145. Meaning. As regards their signification, the pri- 
mary derivatives fall into two great classes, the one indicating 
the action expressed by the verbal root, the other the person or 
thing in which the action appears, the agent or actor the 
latter, either substantively or adjectively. The one class is ab- 
stract, infinitival; the other is concrete, participial. All other 
meanings may be viewed as modifications or specializations of 
these two. 

Even the words indicating recipience of action, the passive participles, 
are, as their use also as neuter or reflexive shows, only notably modified 
words of agency. The gerundives are, as was pointed out above (961), sec- 
ondary derivatives, originally indicating only 'concerned with the action'. 

1146. But these two classes, in the processes of formation, 
are not held sharply apart. There is hardly a suffix by which 
action-nouns are formed which does not also make agent-nouns 
or adjectives; although there are not a few by which are made 
only the latter. In treating them in detail below, we will first 
take up the suffixes by which derivatives of both classes are 
made, and then those forming only agent-nouns. 

To follow an alphabetical arrangement, although in some respects more 
practically convenient, seems too artificial, and destructive of all natural 
connection. The Index will give what help is needed toward finding any 
particular suffix which is sought. 

1147. Stems without suffix: Root-words. These 
words and their uses have been already pretty fully consid- 
ered above (323, 348 ff., 383 ff.. 400, 401). 

a. They are used especially (in the later language, almost 
solely) as finals of compounds, and have both fundamental val- 
ues, as action-nouns (frequently as infinitives : 971), and as 
agent-nouns and adjectives (often governing an accusative : 27 le). 
As action-nouns, they are chiefly feminines (384 : in many in- 
stances, however, they do not occur in situations that determine 
the gender). 

b. In a small number of words, mostly of rare occurrence, 
the reduplicated root is used without suffix. 

The Vedic cases are: with simple reduplication, sasydd, dfe/t, sasrut 
(irregular, for susrut], didyd and didytit, juhu, and perhaps pffM; with in- 

1148] STEMS IN a. 375 

tensive reduplication, -nenf, jogU, yamytidh, vdnlvan (with the intensive 
instead of the usual radical accent). In ddridra is seen a transfer to the 
a-declension. AsusQ is probably to be understood as a compound, asu-sfi. 

c. If the root end in a short vowel, a t is regularly and 
usually added (383 b). 

Examples, and the exceptions found to occur, have been given at the 
place just quoted. 

d. Words of this form in combination with verbal prefixes 
are very numerous. The accent rests (as in combination of the 
same with other preceding elements) on the root-stem. 

e. In a single instance, crtitkarna (RV.), 'of listening ears', a stem of 
this class occurs as prior member of a compound. 

1148. 51 a. With the suffix ^ a is made a very large 
and heterogeneous body of derivatives, of various meaning 
and showing various treatment of the root : ^^-strengthen- 
ing, vrddM-strengthening, retention unchanged, and redu- 

In good part, they are classifiable under the two usual gen- 
eral heads ; but in part they have been individualized into more 
special senses. 

1 . With ^wmj-strengthening of the root (where that is pos- 
sible : 235, 240). These are the great majority, being more 
than twice as numerous as all others together. 

a. Many nomina actionis : as, prdma, 'weariness', grdha, 'seizure', aya, 
'movement', veda, 'knowledge', hdva, 'cair, krodha, 'wrath', jo'sa, 'enjoyment*, 
tdra, 'crossing', sdrga, 'emission'. 

b. Many nomina agentis : as, ksamd, 'patient', svajd, 'constrictor', hinvd, 
'impelling', jivd, 'living', meghd, 'cloud', coda, 'inciting', plavd, 'boat', sard, 
'brook', sarpd, 'serpent', area, 'radiant'. 

c. Of the examples here given, those under a accent the radical syllable, 
and those under b the ending. And this is in a majority of cases the fact 
as regards the two classes of derivatives ; so that, taken in connection with 
kindred facts as to other suffixes, it indicates such a. difference of accent as 
a general tendency of the language. A few sporadic instances are met with 
of the same form having the one or the other value according to its accent: 
thus, esa, 'haste', esd, 'hasting'; fdsa, 'order', fasd, 'orderef (other examples 
are feama, fafea, pofea): compare a similar difference with other derivatives 
as, ana, an, man;. But exceptions are numerous thus, for example, 

jayd, javd, smara, action-nouns; frdva, mdgha, Mnsa, agent-nouns and 
the subject calls for a much wider and deeper investigation than it has yet 
received, before the accentuation referred to can be set up as a law of the 
language in derivation. 

2. With ^^'-strengthening of the root but only where 


a is the resulting radical vowel : that is, of medial a, and of 
final r (most often), u or u, i or i (rare'. 

a. Examples of action-nouns are: kama, 'love', bhagd, 'share', nada, 
'noise', davd, 'fire', tard, 'crossing'. Hardly any forms of clear derivatior 
and meaning are quotable with accent on the root-syllable. 

b. Examples of agent-nouns are: grabhd, 'seizing', vahd, 'carrying', 
nayd, 'leading', jard, 'lover 1 . 

3. With unstrengthened root, the examples are quite few: thus, fcrpd, 
'lean', yuga, 'yoke', sruvd, 'spoon', priyd, 'dear', vrd, 'troop', cued, 'bright'. 

a. A number of words of this class, especially as occurring in com- 
position, are probably results of the transfer of root-stems to the a-declension : 
thus, -raja, -ghusa, -sphura. 

b. A few Vedic examples are made from conjugational stems : thus, 
from the causative stem, kupaya and tdnaya and mrgraj/a(?), -Inkhaya, 
-ejaya, -dhdrayd, -parayd; from present-stems of the nit-class (IV.), or sec- 
ondary roots made from them (716), hinvd, -mud, -jinvd, -pinvd, -sinvd, 
-sunvd-j from others, -prnd, -mrnd, -pacyd, -mam/a, -dasya, -jurya, -ksudhya, 
-trsya, -jighrd, -piba; from aorist-stems(?), yesd, nesa (in nesatama, RV., once). 

4. With reduplicated root, the derivatives in the older language are a 
class of some extent. They are made in a few instances with the simple 
reduplication: thus, cacard, 'movable' (?), cicayd, 'inciting', sasrd, 'flowing', 
uaura, 'hiding'; much more often, with the intensive reduplication : thus, 
reriM, 'licking', vevijd, 'hasting', -roruda, 'weeping', -canfcrama, 'striding', 
caracard and calacald, 'movable', kanikradd, 'noisy', panispadd, 'quivering', 
varlvrtd, 'rolling'. 

Only a single example of an action-noun has been noticed, namely 
fipndtha, 'attack', with accent on the root. Other examples of accent else- 
where than on the suffix are only the stems (of doubtful meaning or derivation) 
car car a and gar gar a. 

5. Derivatives with this suffix from roots as compounded 
with the verbal prefixes are quite common, in all the modes of 
formation (in each, in proportion to the frequency of independ- 
ent words) : constituting, in fact, considerably the largest body 
of derivative stems with prefixes. They are of both classes as 
to meaning. The accent is, with few exceptions, on the end- 
ing and that, witlftmt any reference to the value of the stem 
as action-noun or agent-noun. 

Examples are : sarhgamd, 'assembly', nimesd, 'wink', abhidrohd, 'enmity', 
anukard, 'assistance', wdand, 'inspiration', pratyacravd, 'response'; paricara, 
'wandering', sarhjayd, 'victorious', vibodhd, 'wakeful', atiyajd, 'over-pious 1 , 
udard, 'inciting, elevated', uttudd, 'rousing', samgird, 'swallowing', adardird, 
'crushing', adhicankramd, 'climbing'. 

The only definable class of exceptions in regard to accent appears to 
be that of the adverbial gerunds in am (above, 995). which are accented 
on the root-syllable. A very few other stems have the same tone: for ex- 


STEMS IN a, a. ana. 


ample, utpata, 'portent'; apresa, 'plague'. A few others, mostly agent-nouns, 
have the accent on the prefix: for example, vybsa (i. e. vf-osa), 'burning', 
prdtivefa, 'neighbor', dbhaga, 'sharing'; but also sdrhkufa, 'appearance'. 

For the remaining compounds of these derivatives, with the inseparable 
prefixes and with other elements, see the next chapter. It may be merely- 
mentioned here that such compounds are numerous, and that the a-derivative 
has often an active participial value, and is frequently preceded by a case- 
form, oftenest the accusative. 

Many words in the language appear to end with a suffix a, while yet 
they are referable to no root which can be otherwise demonstrated as such. 

1149. 5TT a. This suffix has. for the most part, the 
subordinate office of making corresponding feminines (espe- 
cially adjective] to masculines in % a in which use, of 
course, it might be open to question whether we should 
not regard it rather as an element of secondary character. 
It is comparatively little employed independently. 

There are, however, a considerable number of feminine 
nouns in a in the language to which no corresponding mascu- 
lines exist ; and many of them are distinctly traceable to roots. 
The latter are especially action-nouns, accented on the suffix : 
and such derivatives in u form a distinct class, being made some- 
times from simple roots, but especially, and at will, from 
conjugational stems, desiderative, causative, and denominative. 

Examples from simple roots are : ifa, 'lordship', Ifcndd, 'play', jara, 
'old age : , nindd, 'reproach'. 

Examples from desiderative stems are very few in the oldest language, 
but common later: thus, jigisd and bhiksd (RV.); the latter and virtsd (AV.); 
blbhatsd (VS.); in Brahmana language, pipasa, bubhuksa, mimahsa, $iksa or 
Ciksd, etc. 

From causative stems, the only Vedic example is gamayd (compare the 
derivatives in -at/a, above, 1148. 3 b); but the formation becomes common 
later, in the making-up of the periphrastic perfect (above, 1070ff.). The 
derivatives from present-stems thus, vidd, edhd, juhavd, etc. made 
for a similar office, have been mentioned in connection with this perfect 

From denominative stems, the older instances are more numerous : for 
example, afvayd, sukratuyd, apasyd, urusyd, asuyd, afanayd, jwanasya. 
Their y is not resolved into i. 

1150. R ana. With this suffix (as with 5[ a) are form- 
ed many derivatives, of both the principal classes of mean- 
ing, and with not infrequent specializations. The root has 


usually <7^#-strengthening, but sometimes vrddhi instead; 
and in a few cases it remains unstrengthened. Derivatives 
of this formation are frequent from roots with prefixes, and 
also in composition with other elements. 

The normal and greatly prevalent accent is upon the root- 
syllable, without regard to the difference of meaning : but cases 
occur of accented final, and a few of accented penult (which last 
are possibly to be referred to another suffix of the same form 
but of a different origin). 

The action-nouns are in general of the neuter gender. The 
feminine of adjectives is made either in a or in 1 (for details, 
see below). And a few feminine action-nouns in ana and am 
occur, which are to be ranked as belonging to this suffix. 

1. With strengthened and accented root-syllable. Under 
this head fall, as above indicated, the great mass of forms. 

a. With ^Una-strengthening : examples of action-nouns are sddana, 'seat', 
rdksana, 'protection', ddna, 'giving', cdyana, 'collection', vedana, 'property', 
hdvana, 'call', bhtijana, 'enjoyment', kdrana, 'deed', vdrdhana, 'increase'; 
of agent-nouns, tdpana, 'burning', cetana, 'visible', co'dana, 'impelling'. 

The derivatives dUsana and -mdrjana have the same irregular strength- 
ening which appears in their present-systems (627, 1042 a), and other 
formations from the same roots. 

b. With weZd/u'-strengthening (only in such circumstances that a remains 
as vowel of the radical syllable): examples (all that have been noted in the 
older language) are -cdtana, ndfana, mddana, -vdcana, -vdsana, -vdhana, sdd- 
ana, -spd?ana, svddana, -dyana, -ydvana, -srdvana, -pdrana, vdrana. 

c. From roots with prefixes, the derivatives of this formation are very 
numerous, being exceeded in frequency only by those made with the suffix 
a (above, 1148.5).' A few examples are: akrdmana, 'striding on', udydna, 
'upgoirig', nidhdna, 'receptacle', prdnana, 'expiration', vimocana, 'release' and 
'releasing', samgrdmana, 'assembly' and 'assembler', adhivikdrtana, 'cutting 
off', avaprabhrdh$ana, 'falling away down'. 

For other compounds of these derivatives, showing the same accent (and 
the same feminine stem), see the next chapter (below, 1272). 

A few exceptions occur: vicaksand, upariyayand, and the feminines 
pramandani and nirdahani. 

d. The adjectives of this formation, simple or compound, make their 
feminine usually in 1: thus, cddanl, pfyani, spdranl, jdmbham ; prajndnl, 
prtiksani, sarhgrdham, abhisdvanl, vidhdrani (cetarii is of doubtful meaning : 
below, 2 a). 

An adjective compound, however, having a noun in ana as final 
member, makes its feminine in a: thus, supasarpand, 'of easy approach', 
sddvidhana, 'of sextuple order', anapavacand, 'not to be ordered away'. 

2 . The more irregular formations may be classed as follows : 

1151] STEMS IN ana, as. 379 

a. With accent on the final: a small number of agent-nouns and ad- 
jectives, as karand, 'active' (against kdrana, {'act'), tvarand, 'hasting', 
rocand, 'shining', fcrofemd, 'yelling', svapand, 'sleepy', ksayand, 'habitable'. 

These, unlike the more regular class, make their feminines in a: thus, 
tvarana, spandana, etc. And a very few feminine action-nouns have the same 
form: jarand, dyotand (? RV. i. 123. 4), fvetand (and compare kapand, rafand). 

Beside these may be mentioned a few feminines in ani, of more or less 
doubtful character: arsant, cetarii (to cetana, taparil (to tdpana), vrjaril (with 
vrjdna), hdyani (with hdyand), rajanl, tedaril. 

b. With accent on the penult : a very small number of adjectives : as 
dohdna, 'milking', mandna, 'considerate', manddna, 'rejoicing', saksdna, 
'overcoming', and perhaps vaksdnd, 'carrying' (with aoristic s); and a still 
smaller number of neuter action-nouns: dansdna, 'great deed', vrjdna, 'en- 
closure, town', krpdna, 'misery' (against krpand, 'miserable'), with the 
masculine kirdna, 'dust'. 

The only noticed example of a feminine is in a : turdna. And a few 
feminine nouns have the same form : as jardnd, bhanddnd, kanydna, 
vaksdna, etc. (and compare the anomalous masc. name ufdna: 355 a). 

c. Without strengthening of the root are made a very small number of 
derivatives : namely (besides those already noted, krpdna and krpand, vrjdna 
and vrjaril, kirdna, turdna), bhurana, -stivana and bhuvana (compare the 
perfects of the same roots, sasuva and babhuva, without the usual strength- 
ening : 789 a) : kfyana, pffana, vrsana are doubtful. 

d. A few isolated cases may be here mentioned : jagarand and pdlpulana 
from reduplicated root-forms ; sdmana apparently from the prefix sam; 
anhurand apparently from ahhurd; yosana, 'woman' (beside ydsan, y6sa, etc.), 
and pftana, 'fight', are the only feminines with accented root-syllable. 

1151. EfH^s. By this suffix are made (usually with 
<7W^a-strengthening of the root-vowel) especially a large class 
of neuter nouns, mostly abstract (action-nouns), but some- 
times assuming a concrete value; and also, in the older 
language, a few agent-nouns and adjectives, and a consid- 
erable number of infinitives. 

The accent in words of the first class is on the root, and 
in the second on the ending ; and in a few instances words of 
the two classes having the same form are distinguished by their 
accent; the infinitives have for the most part the accent on the 
suffix . 

1. a. Examples of the first and principal class are: avas, 

'aid, favor', tdpas, 'warmth, prdyas, 'pleasure', tejas, 'splendor', 

grdvas, 'fame', ddhas, 'milking', karas, 'deed', prathas, 'breadth', 

cetas and mdnas, 'mind', cdkms, 'eye', sdras, 'pond', vdcas, 'speech'. 

b. A few words of this class are of irregular formation : thus, without 


strengthening of the root, juvas, 'quickness' (beside jdvas], uras, 'breast\ 
mfdhas, 'contempt'; and pfras, 'head', is to be compared; with vrddhi- 
strengthening, -uacas, vdsas, vdhas; -svadas, -hayas; perhaps with an 
aoristic s, hesas, 'missile'. 

c. After final a of a root is usually inserted y before the 
suffix (258) : thus, dhhyas, ghyas. 

But there are in the oldest language apparent remains of a formation 
in which as was added directly to radical a: thus, bhas and -das (often to 
be pronounced as two syllables), jrias, mas; and -jas and -dhas and -das, 
from the roots ja and dha and da (Benfey, Abh. Gott. Ges. xxiii., 1878). 

2. a. The instances in which an agent-noun is differentiated by its 
accent from an action-noun are : dpas, 'work', and apds, 'active'; ydfas, 
'beauty', and j/apds, 'beauteous'; fdras, 'quickness', and tards (SV., once), 
'quick'; dtivas, 'worship', and duvds, 'lively' (?); mdhas, 'greatness', and 
mahds, 'great'; between rdksas, n., and raksds, m., both meaning 'demon', 
and between tydjas, n., 'abandonment' (?), and tyajds, m., 'descendant' (?), 
the antithesis is much less clear. 

b. Adjectives in as without corresponding abstracts are: tavds, 'strong', 
vedhds, 'pious', probably ahands, 'heady'; and a few other words of isolated 
occurence, as uepas, dhvards. From a denominative stem is made mrgayds, 
'wild animal' (RV. } once). 

c. But there are also a very few cases of abstract nouns, not neuter, 
accented on the ending: thus, jards, 'old age', bhiyds, 'fear'; and doubtless 
also havds, 'call', and tvesds. 'impulse'. The feminine usds, 'dawn', might 
belong either here or under the last preceding head. 

d. Apparently containing a suffix as are the noun upas, 'lap', and certain 
proper names : dngiras, nodhds, bhalands, arcanands. The feminine apsards, 
'nymph', is of doubtful derivation. 

The irregular formation of some of the words of this division will be 
noticed, without special remark. 

3. The infinitives made by the suffix as have been explain- 
ed above (973) : they show various treatment of the root, and 
various accent (which last may perhaps mark a difference of 
gender, like that between saJias and jards}. 

4. The formation of derivatives in as from roots compounded with pre- 
fixes is very restricted if, indeed, it is to be admitted at all. No infini- 
tive in as occurs with a prefix; nor any action-noun; and the adjective 
combinations are in some instances evidently, and in most others apparently, 
possessive compounds of the noun with the prefix used adjectively: the most 
probable exceptions are -nybkas and wfspardhas. As in these examples, the 
accent is always on the prefix. 

In connection with this, the most common and important 
suffix ending in s, may be best treated the others, kindred in 
office and possibly also in origin, which end in the same sib- 


STEMS IN as, tas ETC., is, us, 


1152. cTR tas, ^H nas, W$ sas. With these suffixes are 
made an extremely small number of action-nouns. Thus : 

a. With tas are made retas, 'seed', and srotas, 'stream'. 

b. With nas are made dpnas, 'acquisition', amas, 'wave', -bharnas, 
'offering', reknas, 'riches'; and in drdvinas, 'wealth', and pdrlnas, 'fulness', 
is apparently to be seen the same suffix, with prefixed elements having the 
present value of union-vowels. Probably the same is true of ddmunas, 

C. With sas is perhaps made vdpsas, 'beauty' (?); and tdrusas may be 
mentioned with it (rather tarus-a?}. 

1153. ^FT is. With the suffix is is formed a small num- 
ber (about a dozen) of nouns. 

They are in part nouns of action, but most are used con- 
cretely. The radical syllable has the $wm-strengthening, and 
the accent is on the suffix (except in jyotis, 'light', vyathis (?), 
and amis, 'raw meat'). Examples are : arcis, rods, and cods, 
'light', hams, 'oblation', vartis, /track'. 

The forms tuvis- and surabhis- appear inorganically for tuvi and surablii 
in a few compounds or derivatives. 

1154. 3H us. With this suffix are made a few words, 


of various meaning, root-form, and accent. 

They are words signifying both action and agent. A few have both 
meanings, without difference of accent: thus, tapus, 'heat' and 'hot'; drus, 
'wound' and 'sore'; cdksus, 'brightness' and 'seeing, eye'; vdpus, 'wonderful' 
and 'wonder'. The nouns are mostly neuter, and accented on the root- 
syllable; exceptions are: in regard to accent, jantis, 'birth'; in regard to 
gender, mdnus, 'man', and nahus, proper name. Of adjectives, are accented 
on the ending jaytis, vanus, vidus (which alone shows an unstrengthened 
root), and daksus, 'burning' (which appears to attach itself to the aorist- 
stem) . 

1155. ^ *. With this suffix are formed a considerable 
body of derivatives, of all genders: adjectives and mascu- 
line agent-nouns, feminine abstracts, and a few neuters. 
They show a various form of the root: strong, weak, and 
reduplicated. Their accent is also various. 

Many words in i have meanings much specialized ; and many 
(including most of the neuters) are hardly to be connected with 
any root elsewhere demonstrable. 

1. The feminine action-nouns are of very various form: thus, with 
weak root-form, ruci, 'brightness', tvtii, 'sheen'; krst, 'ploughing', rtrtf, 


'dance', yudhf, 'fighting; with ptma-strengthening (where possible), ropi, 
'pain', rdndhi, 'subjection', poof, 'heat : , vanfand sanf, 'gain'; with vrddhi- 
strengthening, grdhi, 'seizure', dkraji, 'course', ajf, 'race'; from ydus comes 
dusi (compare dusayati, 1042 a). The variety of accent, which seems reducible 
to no rule, is illustrated by the examples given. 

The few infinitively used words of this formation (above, 975) have a 
weak root-form, with accent on the ending. 

2. The adjectives and masculine agent-nouns exhibit the same variety. Thus : 

a. With unstrengthened root: fuci, 'bright', bhrmi, 'lively' (ybhram], 
grbhi, 'container'. 

b. With strengthened root (or root incapable of gruna-change) : dri, 
'enemy', mdhi, 'great', arcf, 'beam', praniK, 'knot', kridi, 'playing'; and 
some words, of obscure connections, show an apparent vrdd/u-increment : 
thus, drapf, 'mantle', raft' 'heap', pani, 'hand', etc. 

c. With reduplicated root. This is in the older language a considerable 
class, of quite various form. Thus: with weak or abbreviated root, cdfen, 
jdghri (yghar), pdpri, sdsri, -mamrz, babhrf, vawrf, jdgmi, jdghni, sdsni, 
susvi, -fifvi; and, with displacement of final a (or its weakening to the 
semblance of the suffix), dadf, papf, yayf (with a case or two from yayi], 
-jey'fii, dddhi ; from the wr-form of roots in changeable r, jdguri, tdturi, 
pdpuri; with simple reduplication, yuyudhi, vfvici; with strengthened 
reduplication, -cacali, tdtrpi, dddhrsi, vdvahi, sasaM, tutuji and tutujt, yUyuvi, 
yuyudhi; and jarbhdri. And karkart. 'lute', and dundubhi, 'drum', have the 
aspect of belonging to the same class, but are probably onomatopoetic. The 
accent, it will be noticed, is most often on the reduplication, but not seldom 
elsewhere (only once on the root). 

It was noticed above (271 f) that these reduplicated derivatives in i 
especially often take an object in the accusative, like a present participle. 

d. Formations in i from the root compounded with prefixes are not at 
all numerous. They are accented usually on the suffix. Examples are: 
abhidrohf, ayaji, vyanavt, nijaghnf, paradadf, visasaht; but also ajdni, amtiri, 

As compounded with other preceding words, the adjectives or agent- 
nouns in i are not rare, and are regularly accented on the root : see the next 
chapter, 1276. 

e. From ydha comes a derivative -dhi, forming many masculine com- 
pounds, with the value both of an abstract and a concrete: thus, with pre- 
fixes, antardhf, uddhi, nidhi, paridhf, etc. Opinions are divided as to whether 
it is to be regarded as formed with the suffix i, displacing the radical a, or 
with weakening of a to i. 

3. Neuter nouns in i are few, and of obscure derivation: examples are 
afcsi, 'eye', dsthi, 'bone', dddhi, 'curds', etc. 

1156. i. The suffix t, like ETT (above, 1149), has 
in general the office of making a feminine from a masculine 
stem. Thus : 


STEMS IN ', i ti. 


Especially, from consonant-stems in general : from stems in r (or 
ar): 376 a; from some stems in u: 344; from many stems in 
(for details, see the various suffixes ending in a): stems in i very rarely 
have corresponding feminines in I : an example is krmi (masc. kfmi), 

In the oldest language, a few masculines also are made in final I: they 
have been noticed, and their inflection described, above (355 ff.). 

1157. fH ti. This suffix forms a large class of fre- 
quently used feminine nouns of action; and also a few 
agent-nouns (masculine) and adjectives. The root has in 
general the same form as before the suffix rT to, of the pass- 
ive participle (952 ff.) that is to say, a weak, and often 
a weakened or abbreviated, form. 

The accent ought, in analogy with that of the participle, 
to rest always upon the suffix ; and it does so in the (small) 
majority of words quotable from accented texts ; but the ex- 
ceptions are numerous. In the few words used as infinitives 
(975), the accent is on the suffix only. 

1. a. Examples of the normal formation are: rait, 'gift', itf : 'progress', 
ritf, 'flow', stuff, 'praise', bhutf, 'prosperity', bhrtf, 'bearing', pezfctf, 'power', 
kirtf, 'fame', purtf, 'bestowal', maff, 'thought', pitf, 'drink' (|/pa; pple pita], 
dhautf, 'stream' (ydhav ; pple dhauta); and, with accented root, gdti, 
'motion', fdnti, 'repose', dfti, 'division' (yda; pple ditd], trpti, 'satisfaction', 
tsti, 'offering' (yyaj : pple zsfd), -ukti, 'speech' (yvac : pple wfefd), vfddhi, 

b. The roots which form their participle in ita (956) do not have tho 
i also before ti : thus, only gupt i, -drpti (but AV. has once tiditi, beside 
uditd, from j/uad). A few roots having their participle in na instead of ta 
(957), form the abstract noun also in ni (below, 1158). And from the roots 
tan and ran occur tdnti and rdnfi, beside the more regular tati and rati; 
also dhanti (once, VS.) beside dhati. From roots having the form da, the 
derivative in composition is sometimes -tti (for ddti with loss of radical 
vowel: compare the participle-form -tfa, above, 955c): thus, niravatti (K.), 
vdsutti, bhdyatti, maghdtti (all RV.). 

c. A few derivatives are made from reduplicated roots ; their accent is 
various : thus, carfcrff, didhiti and -diditi. jfyarti, and perhaps the proper 
name yaydti. 

d. Derivatives from roots with prefixes are numerous, and have (as in 
the case of the participles in fa, and the action-nouns in tu) the accent on 
the prefix : examples are dnumati, abhlti, dhuti, nfrrti, vykpti, sdmgati. The 
only exceptions noticed are dsaktt and asuti. 

In other combinations than with prefixes, the accentuation is in general 
the same : see the next chapter. 

2. The adjectives and agent-nouns which, as masculines, are to be 


connected with these instead of \\ith the feminine abstracts are very few : 
thus, pUti, 'putrid', vdsti, 'eager', dhuti, 'shaker', jnatf, 'relative', pattf, 
'footman'; and a few others, of more or less dubious character. The accent 
is various, as in the other class. 

3. A few words show the same suffix ti preceded by a vowel, in which 
no organic character seems recognizable, and which may therefore pass for a 
"union-vowel". One, uditi, has been quoted above ; another with i is saniti 
(RV., once); and snihitl and snehili, notwithstanding their long final, may be 
mentioned with them. With ati are made a few, variously accented : thus, 
the action-nouns anhatf, drfatf, vasatf 'nest', ramdti, vratdti, amdti and 
dmati, -dhrajati; and the agent-words aratf, khalatf, vrkdti, rdmati. With 
Iti, rjiti and ddbhiti. In the Brahmanas appear occasional derivatives from 
conjugational stems, like jdnayati (TS.) and agamayiti (K. xxviii. 6). The 
feminine yuvati, 'young (adj.), maiden', is of isolated character. 

4. In some of the words instanced in the last paragraph, ti is perhaps 
applied as a secondary suffix. A kindred character belongs to it in the 
numeral derivatives from pronominal roots, kdti, tdti, ydti, and from numerals, 
as vihcatf, sastf, etc.; with panktf (from pdfica); and in addhatf, from the 
particle addha. 

1158. ft ni. This suffix agrees in general in its uses 
and in the form of its derivatives with the preceding; but 
it makes a much smaller number of words, among which 
the feminine abstracts are a minority. 

1. As was noticed above (1157. Ib), a few verbs (ending in vowels) 
making their passive participle in na instead of ta make their action-noun 
in ni instead of ti. From the Veda are quotable only -jyani, 'injury', and 
jurnf, 'heat'; later occur glani, jirni, and others. Certain other feminine 
nouns of concrete meaning occur: thus, jdni, 'woman', etc. 

2. Examples of words of the other class are: vdhni, 'carrying', fwrni, 
'hasty', bhurni, 'excited', prenf, 'loving', vrsnf and vfmi, 'virile'. 

In prem, y6ni, meni, freni, from is seen a strengthening of the root, 
such as does not appear among the derivatives in ti. 

Derivatives in ni from roots with prefixes do not appear to occur. 

In the words ending in ani, the a has probably the same value with 
that of ati (above, 1157.3); but ani has gained a more independent status, 
and may be best treated as a separate suffix. 

1159. ofa ani. The words made by this suffix have 
the same double value with those made by the preceding 
ones. Their accent is various. Thus: 

a. Feminine action-nouns, sometimes with concreted meaning : as, isani, 
'impulse', pardm, 'injury', dyotani, 'brightness', ksipanf, 'blow', apam, 
'missile', variant, 'track', djani (a-djani : the only example with prefix), 'goad'. 

1161] STEMS IN ti, ni, ani, an, tu. 385 

b. Adjectives and other agent-words are such as cardni, 'movable', 
caksani, 'enlightener', vaksdni, ; strerigthener'. From a reduplicated root- 
form comes -paptani. From desiderative stems are made ruruksdni, sisasani, 
and (with prefix) a-pufuksdni. And a small number of words appear to 
attach themselves to an s-aorist stem: thus, parsdni, saksdni, carsant. 

3. It is questionable whether the infinitives in sdni (978) are to be 
put here, as accusatives of a formation in am, or under the next suffix, as 
locatives of a formation in an, from roots and stems increased by an aoristic s. 

1160. ER an. Not many words appear to be made with 
a suffix of this form, and of these still fewer are plainly to 
be connected with roots. Certain rare neuters (along with 
the doubtful infinitives) are nouns of action; the rest are 
masculine and neuter agent-nouns. The accent is various. 

a. The infinitives which admit of being referred to this suffix, as locative 
cases, are those in am, of which the sibilant appears to be the final of a 
tense-stem. They are all given above (978). 

b. The other action-nouns in an are mahdn, 'greatness', mahhdn (?), 
'liberality', rajdn, 'authority' (RV., once: compare rdjan; the accent-relation 
is the reverse of the usual one), and gdmbhan, 'depth' (VS., once). 

c. Examples of agent-nouns are : masc. tdksan, 'carpenter', r<f?an, 'king', 
vfsan, virile, bull', wfesdn, 'ox'; neut. wddn, 'water 1 , cdksan, 'eye': with 
prefixes, pratidivan, 'antagonist at play' (dtidivan, AV., is perhaps a bad 
reading), vibhvdn, 'superior'. 

d. A few stems in an, running parallel with those in other suffixes 
and filling out their declension, were mentioned above (429 ff.). 

1161. rT tu. The great mass of the words of this form- 
ation are the infinitives accusatives in the later lan- 
guage, in the earlier likewise datives and ablative-genitives : 
see above, 970 b, 972. But a few are also used independ- 
ently, as action-nouns or with concreted meaning; and an 
extremely small number, of somewhat questionable charac- 
ter, appear to have the value of agent- words. They are of 
all genders, but chiefly masculine. The root has the guna- 

The infinitive words are accented on the radical syllable 
when simple, and most of the others have the same accent ; but 
a few have the tone on the ending. 

a. Examples are: of the regular formation, masc., gdntu, 'way', dhdtu, 
'element', rndntu, 'counsel'; fern, vdrtu, 'morning'; neut. vdstu, 'abode'; 
Whitney, Grammar. 25 


with accent on the ending, jantu, 'being', gatti, 'way' and 'song', hetit, 
'cause', fcetoJ, 'banner' (all masc.); with unstrengthened root, rtu, 'season', 
pita, 'drink', sfttu, 'birth'; with vrddhi- strengthening, vdstu (above). 

b. The infinitives in tu have (972) often the union-vowel i before the 
suffix, and this in a few cases is lengthened to I. In other use occur also 
-tdritu and -dhdrltu (both with dus); turphdritu seems of the same formation, 
but is obscure. 

The infinitives, when made from roots with prefixes, have (as was pointed 
out above, 972) the accent on the prefix. But the same words, when used 
(not infinitivally) in further combination (with su and dws), retain the radical 
accent which belongs to the simple word : thus, duratyetu, durniyGmtu, 
dusparihdntu, suprattu. 

c. In a few instances, the suffix tu appears to be added to a tense- or 
conjugation-stem in a : thus, edhatti and vahatu ; tanyatu and tapyatu ,- and 
sisasdtu. The accent of the last is paralleled only by that of jivdtu, 'life', 
which is further exceptional in showing a long a; it is used sometimes in 
the manner of an infinitive. 

1162. ^ nu. This suffix forms a comparatively small 
body of words , generally masculine, and having both the 
abstract and the concrete value. 

The accent is usually on the ending, and the root un- 

Examples are : bhanu, 'light' (later 'sun'), vagnti, 'sound', sunu, 'son', 
ddnu (with irregular accent), m. 'demon', n. 'drop, dew'; dhenii, f., 'cow'; 
grdhnu, 'hasty', dhrsnti, 'bold'. 

This also (like <u), appears sometimes with a prefixed a : thus, krandanu 
and nadanti, 'roaring', nabhanu (and -nti, f.), 'fountain', vibhanjanti (only 
instance with prefix), 'breaking to pieces'; and perhaps krfdnu belongs here. 

1163. 51 ilia. The words made with this suffix are al- 
most without exception action-nouns (though some have 
assumed a concrete value). They are of all genders. The 
root is of a weak (or even weakened) form, and the accent 
usually on the suffix. 

a. Examples are : masc. bhrthd, 'offering', -ferffta, 'making', -itha, 
'going'; neut. ukthd, 'saying', nithd, 'song', tlrthd, 'ford'; fern, (with a) gatha, 
'song', nitha, 'way'. Radical a is weakened to i in -githd and -plthd. Final 
m or n is lost in -gathd and hdtha (as sometimes in the verbal inflection of 
the same roots: 637, 834 b). 

A few examples of combination with prefixes occur, with accent on the 
final: thus, nirrthd, 'destruction', sarhgathd, 'union', etc. 

b. Still more common in the older language is a form of this suffix to 
which has become prefixed an a, which is probably of thematic origin, though 

1168] STEMS IN nu. tha. thu, yu 1 ma, mi, man. 387 

become a union-vowel. Thus : masc. cardtha, 'mobility', yajdtha, 'offering', 
ravdtha, 'cry 1 , capdtha, 'curse', stavdtha, 'praise'; neut. ucatha, 'speech', 
viddtha, 'ordering'. Before this, a root has sometimes guna: thus, faydtha, 
'couch', tvesdtha, 'vehemence'. With a prefix, the accent is thrown forward 
upon the final : thus, avasathd, 'abode', pravasathd, 'absence'; prandtha, 
'breath', is treated as if pran were an integral root. 

c. Isolated combinations of tha with other preceding vowels occur: thus, 
vdrutha, 'protection', with another doubtful case or two ; and matutha (yman ?). 

1164. 51 thu. This suffix has an [ a attached to it (like 

O ' V 

2T tha, above), and. in the very few derivatives which it 
makes, appears only as 5fST dthu. 

The only Vedic examples are ejdthu, 'quaking', vepdthu, 'trembling', 
standthu, 'roaring'. Later cases are nanddthu (TS.), vamathu, cvayathu, etc. 

1165. T yu. With this suffix are made a very few nouns, 
both of agent and of action, with unstrengthened root and 
various accent. Thus : 

a. Abstracts (masc.) are manyv, 'wrath', mrtyti, 'death' (with t added 
to the short final of the root). 

b. Adjectives etc. are bhujyu, 'pliable', yundhyti, 'pure'; ydjyu, 'pious', 
sdhyu, 'strong', ddsyu, 'enemy', and one or two more. 

For other derivatives ending in yu, see the suffix , below, 1178g,h. 

1166. 13 ma. The action-nouns made by this suffix are 
almost all masculine ; and they are of various root-form and 
accent, as are also the agent-nouns and adjectives. 

a. Examples of the former class are : ajmd, 'course', gharmd, 'heat'; 
ema, 'progress', bhdma, 'brightness', sdrma, 'flow', s(dma, 'song of praise'. 

b. Examples of the latter class are : tigmd, 'sharp', bhimd, 'terrible', 
fagmd, 'mighty'; idhmd, 'fuel', yudhmd, 'warrior'. A single instance from a 
reduplicated root is tutumd, 'powerful'. 

1167. FT mi. A very small number of nouns, masculine 
and feminine, formed with mi, may be conveniently noticed here. 

Thus, from r-roots, urmf, 'wave', -kurmi, 'action', surrm, f., 'tube'; 
from others, jamf (?), 'relation', bhUmi or bh&mi, f., 'earth', laksmi, 'sign'; 
also probably rafmf, 'line, ray'. 

1168. R man. The derivatives made with this suffix 


are almost only action-nouns . The great majority of them 
are neuter, and accented on the root-syllable; a much small- 
er number are masculine, and accented on the suffix. The 
few agent-words are, if nouns, masculine, and have the 



latter accent : in several instances, a neuter and a masculine, 
of the one and the other value and accent, stand side by 
side. The root has in general the ^^-strengthening. 

1. a. Examples of regularly formed neuters are: kdrman, 'action', 
jdnman, 'birth', ndman, 'name', vdrtman, 'track', vefman, 'dwelling', homan, 
'sacrifice', -dytitman, 'splendor'. 

b. Examples of masculine abstracts are: ojmdn, 'strength', jemdn, 
'conquest', svddmdn, 'sweetness'. 

c. Corresponding neuter action-nouns and masculine agent-nouns are : 
brahman, 'worship', and brahman, 'priest'; ddman, 'gift', and daman, 'giver'; 
dhdrman, 'rule', and dharmdn, 'orderer'; sddman, 'seat', and sadmdn, 'sitter'. 
Very few other agent-nouns occur; and all, except brahman, are of rare 

On the other hand, varsman and svadman (and variman) have the dif- 
ference of gender and accent without a corresponding difference of meaning. 

The noun dfman, 'stone', though masculine, is accented on the radical 
syllable; and two or three other questionable cases of the same kind occur. 

The derivatives in man used as infinitives (974) have for the most part 
the accent of neuters: the only exception is vidmdne. 

d. A few words, of either class, have an irregular root-form : thus, 
bhuman 'earth', syuman; bhumdn 'abundance', slmdn, bhujmdn, vidmdn; and 
kdrsman, bhdrman, fdkman. 

e. Derivatives in man from roots with prefixes are not numerous. They 
are usually accented on the prefix, whether action-nouns or adjectives : thus, 
prdbharman, 'forthbringing', prdydman, 'departure'; dnuvartman, 'following 
after' : the exceptions, vijaman, prativartmdn, visarmdn, are perhaps of pos- 
sessive formation. 

2. The same suffix, though only with its abstract-making 
value, has in a number of cases before it a union-vowel, i or 
I; and imdn comes to be used as a secondary suffix, forming 
abstract nouns (masculine) from a certain number of adjectives. 

a. The neuters in iman and Iman are all primary formations, belonging 
to the older language: thus, jdniman and variman (beside variman, as no- 
ticed above); and ddrlman, dhdrlman, pdriman (and pdreman, SV., once), 
bhdrlman, variman, sdrlman, stdrlman, sdvlman, and hdvlman. Those in 
Iman are hardly met with outside the Rig- Veda. 

b. The masculines in imdn are in the oldest language less frequent 
than the neuters just described : they are jarimdn, prathimdn, mahimdn, 
variman (beside the equivalent variman and variman), varsimdn (beside the 
equivalent varsman and varsman), harimdn, and draghimdn (VS.) beside dragh- 
mdn (RV.). Some of these, as well as of the derivatives in simple man, 
attach themselves in meaning, or in form also, to adjectives, to which they 
seem the accompanying abstracts : compare the similar treatment of the primary 
comparatives and superlatives (above, 468): such are papmdn (to papa, 

1170] STEMS IN man. van. vana, vani, vanu. 389 

pdpiyas, etc.); drdghmdn etc. (to dirghd, drdghlyas, etc.); vdriman etc. (to 
urn, vdrlyas, etc.); prdthiman (to prthu, prdthistha) ; harimdn (to Mrz or ftan'ta); 
vdrsman etc. (to vdrslyas etc.); svddman etc. (to suadw, svddlyas, etc.). 
Then in the Brahmana language are found further examples: thus, dhumri- 
man (TS., K.), dradhiman (K.: to drdhd, drddhlyas, etc.), iaruniman (K.), 
lohitiman (KB.); and still later such as laghiman; while ksepiman (to ksiprd, 
kseptyas, etc.) and krsniman, and so on, are allowed by the grammarians. 

1169. 5R tfaft. By this suffix are made almost only 
agent-words, adjectives and nouns, the latter mostly mas- 
culines. The root is unstrengthened, and to a short final 
Yowel is added a rT t before the suffix. The accent is al- 
most always on the root, both in the simple words and in 
their compounds. 

The insertion of t is an indication that the words of this form are orig- 
inally made by the addition of an to derivatives in u and tu; yet van has 
the present value of an integral suffix in the language, and must be treated 
as such. 

1. a. Examples of the usual formation are: masc. ydjvan, 'offering', 
drtihvan, 'harming', fdkvan, 'capable', -rikvan, 'leaving', -jttvan, 'conquering', 
sfitvan, 'pressing', kftvan, 'active', -gdtvan (like -gat, -gatya), 'going', sdtvan 
(j/san), 'warrior', drvan (only example with strengthened root), 'courser'; 
neut. pdrvan, 'joint', dhdnvan, 'bow'. 

b. Examples from roots with prefixes (which are not rare) are: atitvan, 
'excelling', upahdsvan, 'reviler', sambhrtvan, 'collecting'; and probably vivas- 
van, 'shining': abhfsatvan is a compound with governing preposition (1310). 

For the compounds with other elements, which, except in special cases, 
have the same accent, see below, 1277. 

C. The stem musivdn, 'robber' (RV., once), is the only one with a 
union-vowel, and is perhaps better regarded as a secondary derivative of 
which a few are made with this suffix: see below, 1234. 

d. From a reduplicated root are made rdravan and cikitvdn (and possibly 

2. The number of action-nouns made, with the suffix van is extremely 
small : namely, davdn, 'giving', and turvdn, 'overcoming', both used as in- 
finitives (974), and bhurvdn, 'unrest' (?): likewise dhurvan, 'injury', also used 
as infinitive (unless this is rather dhurv-an). 

The feminines corresponding to adjectives in van are not 
made (apparently) directly from this suffix, but from vara, and 
end in vari; see below, 1171 b. 

1170. SR vana. 3H vani. cR vanu. The very few words 
made with these suffixes may best be noticed here, in con- 


nection with 3R van (of which the others are probably sec- 
ondary extensions). 

a. With vana are made vagvand, 'talkative', satvand, 'warrior' (beside 
sdtvan, above); and, from a reduplicated root, wukvand, 'shining'. 

b. With vani are made from simple roots turvdni, 'excelling', and 
bhurvdni, 'restless' (compare turvdn and bhurvdn, just above); and, from 
reduplicated roots, fUfukvdni, 'shining', dadhrsvdni, 'daring', tuturvdni, 
'striving after', and jugurvdni, 'praising': arharisvdni is obscure. 

c. With vanu is made only vagvanti, 'tone, noise'. 

1171. cTf vara. With this suffix are made a few deriv- 
atives, of all genders, having for the most part the value 
of agent-nouns and adjectives. 

Much more common are the feminine stems in of^t 
vari, which, from the earliest period, serve as the corres- 
ponding feminines to the masculine stems in 3Ff van. 

a. A few masculine adjectives in vara occur, formally accordant (except 
in accent) with the feminines: thus, itvard, 'going', -advard, 'eating', 
-sadvard, 'sitting', ifvard, 'ruler, lord'; and with them doubtless belongs 
vidvald, 'knowing' (with I for r). 

b. The feminines in van accord in treatment of the root and in accent 
with the masculines in van to which they correspond : thus, ydjvari, -jttvari, 
sftvari, -$ivari, -yavari, and so on (about twenty-five such formations in 
RV.); from a reduplicated root, -f if van. 

6. A very small number of neuters occur, with accent on the root: 
thus, kdrvara, 'deed', gdhvara, 'thicket'; and a feminine or two, with accent 
on the penult : urvdra, 'field', and urvdrl, 'tow' (both of doubtful etymology). 

We take up now the suffixes by which are made only stems 
having the value of agent-nouns and adjectives ; beginning with 
a brief mention of the participial endings, which in general have 
been already sufficiently treated. 

1172. %R[^ant (or ETcT^). The office of this suffix, in 
making present and future participles active, has been fully 
explained above, in connection with the various tense-stems 
and conjugation-stems (chaps. VIII. XIV.), in combination 
with which alone it is employed (not directly with the root, 
unless this is also used as tense-stem). 

With the same or a formally identical suffix are made from 
pronominal roots lyant and kiyant (517). And ddvayant, 'not 
double-tongued' (RV. , once), appears to contain a similar form- 

1176] STEMS IN vara, ant, vans, mana, ana, ta. 391 

ation from the numeral dm unless we are to assume a de- 
nominative verb-stem as intermediate. 

Here may also be best mentioned the words made with 
the so-called suffix anta (fern, anta or anti], being evident trans- 
fers (Prakritic) of stems in ant to the a-declension. 

Extremely few such words occur in the oldest language : namely, pdnta, 
'draught' (RV.), vasantd, 'spring' (RV.: beside it also hemantd, 'winter'); 
vefantd or ve$anti, 'tank' (AV.; QB. ve$dnta], jivanti, a certain healing 
plant (RV.j; and probably the proper name tarantd (RV.). A few others are 
instanced as admitted later : thus, nandanta and nandayanta, gadayanta, 
jayanta, jaranta, prananta, etc.: all are said to be accented on the final. 

A different extension of the same suffix is exhibited in the proper 
names dhvasdnti and pucantf (RV.), with which may be mentioned purusdnti. 

1173. 3"fa vans (or cfH vas). For the (perfect active) par- 
ticiples made with this suffix, see above, chaps. X. and 
XIV., and 458 ff. 

A few words of irregular and questionable formation were noticed at 
462, above. Also, apparent transfers to a form us or usa. 

The oldest language (RV.) has a very few words in vas, of doubtful 
relations: fbhvas, 'seizing', and ffkvas, 'skilful' (beside words in va and 
van], and perhaps khidvas (ykhad). The neuter abstract vdrivas, 'breadth, 
room' (belonging to urti, 'broad', in the same manner with variyas and varimdn], 
is quite isolated. 

The unique tatan'usti (RV., once) is possibly to be divided tatanus-ti, 
and connected with this suffix. 

1174. JTH mana. The participles having this ending 
are, as has been seen, present and future only, and have 
the middle,, or the derived passive, value belonging in gen- 
eral to the stems to which the suffix is attached. 

1175. ^H ana. The participles ending in TR ana are 
of middle and passive value, like those just noticed, and 
either present, perfect, or (partly with the form $TH sana: 
above, 897) aorist. 

A few other words ending in the same manner in the old language may 
be mentioned here. The RV. has the adjectives vdsavana, 'well-endowed' r 
and urdhvasand, 'uplifted', evidently made on the model of participial stems. 
Also the proper names dpnavana, pfthavana, and cydvana and cydvatana. 
Pdr$ana, 'abyss', is doubtful. 

1176. rT ta. The use of this suffix in forming parti- 
ciples directly from the root, or from a conjugational (not 


a tense) stem, was explained above, chap. XIII. The par- 
ticiples thus made are in part intransitive, but in great part 
passive in value (like those made by the two preceding suf- 
fixes, but in much larger measure, and more decidedly:. 

For the frequent use of the connecting vowel with the 
suffix, making its form, ita, see also chap. XIII. 

a. A few general adjectives, or nouns with concrete meaning, are 
adaptations of this participle. Examples are: trsta, 'rough', citd, 'cold 1 , 
drdhd (for drdhd: 224 a), 'firm'; dutd, 'messenger', sutd, 'charioteer'; rtd, 
'right', ghrtd, 'ghee', jatd, 'kind', dyutd, 'gambling', nrttd, 'dance', fivitd, 
'life', caritd, 'behavior.'. The adjective tigitd (RV.), 'sharp , shows anomalous 
reversion of palatal to guttural before the i (216). Vavdta, 'dear', is a single 
example from a reduplicated root. 

b. Doubtless after the example and model of participles from denomin- 
ative stems (of which, however, no instances are quotable from the Yeda), 
derivatives in ita are in the later language made directly from noun and 
adjective-stems, having the meaning of 'endowed with, affected by, made to 
be', and the like (compare the similar English formation in ed, as horned, 
barefooted, bluecoated). Examples are rathita, 'furnished with a chariot', etc. 

c. A few words ending in fa are accented on the radical syllable, and 
their relation to the participial derivatives is very doubtful: such are dsta, 
'home', mdrta, 'mortal', vdta, 'wind'; and with them may be mentioned 
gdrta (?), nafcta, 'night', hdsta, 'hand'. 

d. Several adjectives denoting color end in ita, but are hardly connect- 
ible with roots of kindred meaning: thus, palitd, 'gray', dsita, 'black', rtihita 
and lohita, 'red', hdrita, 'green'; akin with them are eta, 'variegated', cyetd, 

The feminines of these stems are in part irregular : thus, enl and cyenl ; 
rohinl and lohirii, and hdrinl (but the corresponding masc. hdrina also occurs); 
and dsikni, pdliknl, and hdriknl, 

e. A small number of adjectives in the older language ending in ata 
are not to be separated from the participial words in fa, although their 
specific meaning is in part gerundive. They are: pacatd, 'cooked', dar$atd 
and pafyata, 'seen, to be seen, worth seeing'; and so yajatd, haryatd, 
bharatd. The y of pafyata and haryatd indicates pretty plainly that the a 
also is that of a present tense-stem. Rajatd, 'silvery', is of more obscure 
relation to yraj 'color'. 

1177. ^ na (and ^T ina, 3^ una}. The use of the suffix 
^ na in forming from certain roots participles equivalent to 
those in cT ta, either alongside the latter or instead of them, 
was explained above, chap. XIII. (957). 

a. With the same suffix are made a number of general adjectives, and 

1178] STEMS IN ta, na, u. 393 

of nouns of various gender (fern, in na). The accent is on the suffix or on 
the root. A few examples are : umd, 'hot', puna, 'fortunate', dpna, 'ravenous', 
fvitna, 'white'; masc. prapnd, 'question', yajna, 'offering', ghrnd, 'heat', 
uarna, 'color', svdpna, 'sleep'; neut. parnd, 'wing', rdtna, 'jewel' (?); fern, 
trena, 'thirst', yacnd, 'supplication'. But many of the stems ending in na 
are not readily connectible with roots. An antithesis of accent is seen in 
fedrna, 'ear', and fcarnd, 'eared'. 

b. The very few words ending in ina are perhaps related with these 
(perhaps rather with those in ana): thus, amind, 'violent', vrjind, 'crooked', 
ddksina, 'right', and one or two others of questionable etymology. 

c. Of the words in wna, few are clearly referable to roots: thus, fcart/na, 
'action', dharuna, 'bearing', -cetuna, 'showing'; drjwna, 'white', tdruna, 
'young', vdruna, 'Varuna'. In meaning and in accent they vary like the de- 
rivatives in ana. 

These are all the proper participial endings of the language. 
The gerundives, later and earlier, are in so great part evident 
secondary formations, that they will be noticed farther on, un- 
der the head of secondary derivation. 

We will take up now the other suffixes forming agent-nouns 
and adjectives, beginning with those which have more or less 
a participial value. 

1178. 3 u. With this suffix are made a considerable 
body of derivatives, of very various character adjectives, 
and agent-nouns of all genders, with different treatment of 
the root, and with different accent. It is especially used 
with certain conjugational stems, desiderative (particularly 
later) and denominative (mainly earlier), making adjectives 
with the value of present participles; and in such use it 
wins in part the aspect of a secondary suffix. 

The root has oftenest a weak (or weakened) form ; but it 
is sometimes vriddhied ; least often (when capable of guna], it 
has the ^wna-strengthening all without any apparent con- 
nection with either accent or meaning or gender. After final 
radical a is usually addded y (258) before the suffix. A few 
derivatives are made from the reduplicated root. 

Many words ending in u are not readily, or not at all, connectible with 
roots ; examples will be given only of those that have an obvious etymology. 

a. Examples' of ordinary adjectives are : wrti, 'wide', rju, 'straight', 
prtJiu, 'broad', mrdu, 'soft', sadhu, 'good', svadti, 'sweet', tdpu, 'hot', vdsu, 
'good'; jayu, 'conquering', darti, 'bursting'; pat/tJ, 'lying', reku, 'empty'; 
dhayu, 'thirsty', payti, 'protecting'. Final a appears to be lost before the 
suffix in -sthu (susthu, anusthu). 


b. Examples of nouns are : masc. ah$u, 'ray', ripu, 'deceiver', vayii, 
'wind-god', dsu, 'life', mdnu, 'man, Manu'; fern, isu (also masc.), 'arrow', 
sfndhu (also masc.), 'river'; tanu, 'body'. 

c. Derivatives from reduplicated roots are: cikitu, jigyu, sisnu. -tatnu 
(unless this is made with nu or tnu], ydyu or yayu and yfyu (with final a 
lost), pfpru (proper name), -didhayu; and babhru, -raru (araru), malimlu (?) 
have the aspect of being similar formations. 

d. A few derivatives are made from roots with prefixes, with various 
accentuation : for example, upayu, 'on-coming', pramayu, 'going to destruction', 
vikUndu, a certain disease, a&Tifpw, 'rein (director)', sdmvasu, 'dwelling to- 

e. From tense-stems, apparently, are made tanyu, 'thundering', and 
(with aoristic s) ddksu and dhdksu (all RV.J. 

f. Participial adjectives in u from desiderative "roots" (stems with loss 
of their final a) are sufficiently numerous in the ancient language (RV. has 
more than a dozen of them, AV. not quite so many) to show that the form- 
ation was already a regular one, extensible at will ; and later such adjectives 
may be made from every desiderative. Examples (older) are: ditsu, dipsu, 
cikitsti, titiksu, piplsu, mumuksu, fifliksu; and, with prefix, abhidipsu. 

These adjectives, both earlier and later, may take an object in the 
accusative (27 la). 

g. A few similar adjectives are made in the older language from caus- 
atives : thus, dharayu ('persistent'), bhajayu, bhdvayu, mahhayu, mandayu, 
cramayii; and mrgayu from the caus.-denom. mrgdya. 

h. Much more numerous, however, are such formations from the more 
proper denominatives, especially in the oldest language (RV. has toward 
eighty of them; AV. only a quarter as many, including six or eight 
which are not found in RV.; and they are still rarer in the Brahmanas). In 
a majority of cases, personal verbal forms from the same denominative stem 
are in use : thus, for example, to aghayu, aratlyu, rjuyu, caranyu, memasyw, 
sanisyii, urusyu, saparyu; in others, only the present participle in ydnt, or 
the abstract noun in yd (1149), or nothing at all. A few are made upon 
denominative stems from pronouns : thus, tvayu (beside tvaydnt and tvayd}, 
yuvayu or yuvayfi, asmayu, svayu, and the more anomalous ahamyti, and 
kirhyu. Especially where no other denominative forms accompany the ad- 
jective, this has often the aspect of being made directly from the noun with 
the suffix yu, either with a meaning of 'seeking or desiring', or with a more 
general adjective sense: thus, yavayti, 'seeking grain', varahayu, 'boar-hunting', 
stanasyti, 'desiring the breast'; urnayti, 'woolen', yuvanyti, 'youthful', 
bhimayti, 'terrible'. And so the "secondary suffix yu" wins a degree of standing 
and application as one forming derivative adjectives (as in aharhyu and kimyii, 
above, and doubtless some others, even of the RV. words). In three RV. 
cases, the final as of a noun-stem is even changed to o before it: namely, 
ahhoyti, duvoyti (and duvoyd beside duvasyd], dskrdhoyu. 

None of the words in yu show in the Veda resolution into iu. 

1181] STEMS IN w, u, uka^ aka. 395 

1179. "37 u. The long u is a vastly rarer suffix than 
the other long simple vowels, already described (1149, 1155 . 
It makes a small number of feminines corresponding to 
masculines in u, a very few independent feminines, and 
two or three very rare masculines : as to all which, see 
above, 355 c. 

1180. 3Sfi uka. With this suffix are made derivatives 
having the meaning and construction (271 g) of a present 
participle. The root is strengthened, and has the accent. 

The derivatives in uka are hardly known in the Veda; but they become 
frequent in the Brahmanas, of whose language they are a marked character- 
istic ; and they are found occasionally in the later language. In all prob- 
ability, they are originally and properly obtained by adding the secondary 
suffix fca (1222) to a derivative in u; but they have gained fully the char- 
acter of primary formations, and in only an instance or two is there found 
in actual use an w-word from which they should be made. 

The root is only so far strengthened that the radical syllable is a heavy 
(79) one; and it has the accent, whether the derivative is made from a 
simple root or from one with prefix. 

a. Examples, from the Brahmana language, are : vdduka, ndfuka, 
upakrdmuka, prapdduka, upasthdyuka (258), vydyuka, veduka, bhdvuka, 
kstidhuka, haruka, vdrsuka, samdrdhuka, ddn?uka, alambuka, yiksuka (GB.: 
RV. has fiksu], pramayuka (SB. has pramayu}. 

b. Exceptions as regards root-form are: nirmdrguka (with urdd/u'-strength- 
ening, as is usual with this root : 627) , -kasuka. AV. accents sdrhkasuka 
((^B. has sarhkdsuka) and vikasuka : RV. has sdnukd (which is its only 
example of the formation, if it be one ; AV. has also ghdtuka from ytian, 
and dpramdyuka) ; vasukd (TS. et al.) is probably of another character. 
Afandyuka (PB. et al.) is the only example noticed from a conjugation-stein. 

c. A formation in uka (a suffix of like origin, perhaps, with uka} may 
be mentioned here : namely, from reduplicated roots, jdgaruka, 'wakeful', 
dandafwfca, 'biting', ydyajuka, 'sacrificing much', vdvaduka (later), 'talkative'; 
salaluka is questionable. 

1181. fi aka. Here, as in the preceding case, we 
seem to have a suffix made by secondary addition of 3fi ka 
to a derivative in f a; but it has, for the same reason as 
the other, a right to be mentioned here. Its free use in 
the manner of a primary suffix is of still later date than 
that of uka; it has very few examples in the older language. 


a. In RV. is found (besides pavakd, which has a different accent, and 
which, as the metre shows, is usually to be pronounced pavaka] only sdyaka, 
'missile'; AV. adds piyaka and vddhaka, and VS. abhikr<S$aka. But in the 
later language such derivatives are common, usually with raising of the 
root-syllable by strengthening to heavy quantity : thus, nayaka, ddyaka (258), 
pdcaka, grdhaka, dravaka, bodhaka ; but also janaka, khanaka. They are 
declared by the grammarians to have the accent on the radical syllable. 
They often occur in copulative composition with gerundives of the same 
root: thus, bhakayabhaksaka, 'eatable and eater', vacyavacaka, 'designated 
and designation', and so on. 

That the derivatives in aka sometimes take an accusative object was 
pointed out above (27 Ic). 

The corresponding feminine is made sometimes in aka or in aki, but 
more usually in ika: thus, ndyikd (with ndyakd), pacika, bodhika, dravika; 
compare secondary aka, below, 1222 d. 

b. Derivatives in aka are said to be made from a few roots : thus, 
jalpdka, bhiksdka; but they are not found in the Veda (unless in "pavaka": 
see above), and appear to be very rare at every period. With dku is made 
in RV. mrdaydku, from the causative stem : pfdaku and the proper name 
fksvaku are of obscure connection. 

Derivatives in ika and Ika will be treated below, in connection with 
those in ka (1186). 

1182. cT tr (or rT|" tar}. The derivatives made by this 
suffix, as regards both their mode of formation and their 
uses, have been the subject of remark more than once 
above (see 369 ff., 942 if.). Agent-nouns are formed with it 
at every period of the language; these in the oldest lang- 
uage are very frequently used participially, governing an ob- 
ject in the accusative (271 d) ; later they enter into combin- 
ation with an auxiliary verb, and, assuming a future mean- 
ing, make a periphrastic future tense (942). 

Their corresponding feminine is in tri. 

a. The root has uniformly the ^tma-strengthening. A union- 
vowel i (very rarely, one of another character) is often taken : 
as regards its presence or absence in the periphrastic future 
forms, see above (943). 

"Without grwna-change is only ustr, 'plough-ox' (no proper agent-noun : 
apparently uks-tf : compare the nouns of relationship further on). The root 
grah has, as usual, I thus, -grahitf ; and the same appears in -taritf 
and -maritf. An u- vowel is taken instead by tdrutr and tarutf, dhdnutr, and 
sdnutr; long in varutr ; strengthened to o in man6tr and manotf. From a re- 
duplicated root comes vdvdtr. 

1183] STEMS IN aka, tr, in. 397 

b. The accent, in the older language, is sometimes on 
the suffix and sometimes on the root ; or, from roots combined 
with prefixes, sometimes on the suffix and sometimes on the 

In general, the accent on the root or prefix accompanies the participial 
use of the word; but there are exceptions to this: in a very few instances 
(four), a word with accented suffix has an accusative object; very much 
more often, accent on the root appears along with ordinary noun value. 
The accent, as well as the form, of mano'tr is an isolated irregularity. 

Examples are: jeta dhdnani, 'winning treasures'; yuydm mdrtarh frtitarah, 
'ye listen to a mortal'; but, on the other hand, yarhtd vdsuni vidhate, 
'bestowing good things on the pious'; and jeta jdndndm, 'conqueror of 

c. The formation of these nouns in tr from conjugation-stems, regular 
and frequent in the later language, and not very rare in the Brahmanas, is 
met with but once or twice in the Veda (bodhayitf and codayitri, RV.). In 
nestr, a certain priest (RV. and later), is apparently seen the aoristic s. 

d. The words of relationship which, in whatever way, have gained the 
aspect of derivatives in tr, are pitf, mat/, bhrdtr, ydtr, duhitf, ndptr, jdmatr. 
Of these, only mdtf and ydtr are in accordance with the ordinary rules of 
the formation in tr. 

e. Instead of tr is found tur in one or two RV. examples: j/amttfr, 

f. Apparently formed by a suffix r (or ar) are usr, savyasthr, ndnandr, 
dew, the last two being words of relationship. For other words ending in 
r, see 369. 

1183. ?T in. This is another suffix which has assumed 
a primary aspect and use, while yet evidently identical in 
real character with the frequent secondary suffix of the same 
form denoting possession (below, 1230). 

How far it had gained a primary value in the early language, is not 
easy to determine. Most of the words in in occurring in RV. and AV. are 
explainable as possessives ; in many the other value is possible, and in a few 
it is distinctly suggested : thus, kevaladfa, bhadravadin, nitodtn, dfdrdisin, 
dnamin, vivyadMn; with aoristic 8, saksfn; and, with reduplication, niyayin. 
As the examples indicate, composition, both with prefixes and with other 
elements, is frequent; and, in all cases alike, the accent is on the suffix. 

Later, the primary employment is unquestionable, and examples of it, 
chiefly in composition, are frequent. The radical syllable is usually strength- 
ened, a medial a being sometimes lengthened and sometimes remaining 
unchanged. Thus, satyavddin, 'truth-speaking', abhibhasin, 'addressing', 
mano/mrm, 'soul-winning'. la bhdvin has established itself a prevailingly 
future meaning: 'about to be'. 


The use of an accusative object with words in in was noticed above 
(271 b). 

1184. ^TJH lyas and J$ istha. These, which, from forming 
intensive adjectives corresponding to the adjective of root-form, 
have come to be used, within somewhat narrow limits, as suf- 
fixes of adjective comparison, have been already sufficiently treat- 
ed above, under the head of comparison (466 470). 

It may be further noticed that jyestha has in the older language (only 
two or three times in RV.) the accent also on the final, jyestha ; and that 
pdrsistha is made from a secondary from of root, with aoristic s added. 

When the comparative suffix has the abbreviated form yas (470) , its y 
is never to be read in the Veda as i. 

No other suffixes make derivatives having participial value 
otherwise than in rare and sporadic cases ; those that remain, 
therefore, will be taken up mainly in the order of their fre- 
quency and importance. 

1185. ^ tra. With, this suffix are formed a very few 
adjectives, and a considerable number of nouns, mostly 
neuter, and often having a specialized meaning, as signi- 
fying the means or instrument of the action expressed by 
the root. The latter has usually the ^^-strengthening, but 
sometimes remains unchanged. The accent is various, but 
more often on the radical syllable. 

Here, as in certain other cases above, we have probably a suffix origin- 
ally secondary, made by adding a to the primary tr or tar (1182); but its 
use is like that of a primary suffix. 

,a. Examples of neuter nouns are : gdtra, 'limb', pdttra, 'wing', pdtra, 
'cup', ytiktra, 'bond', vdstra, 'garment', prcftra, 'ear'; astra, 'missile', stotrd, 
'song of praise', potrd, 'vessel'; of more general meaning, ddttra, 'gift', 
ksetra, 'field', m<atra, 'urine', hotrd, 'sacrifice'. The words accented on the 
final have often an abstract meaning: thus, fesatra, 'authority', rastrd, 
'kingdom', fasfra, 'doctrine', sattrd. 'sacrificial session' (also jnatrd, 'know- 

b. Masculines are: ddhstra, 'tusk', mantra, 'prayer', attra (or atrd: 232, 
'devourer', ustra, 'buffalo, camel', and a few of questionable etymology, as 
rm'fra, 'friend', putra, 'son', vrtrd, 'foe'. Mitrd and vrtrd are sometimes 
neuters even in the Veda, and mitra comes later to be regularly of that 

C. Feminines (in tra) are : astra, 'goad', matra, 'measure', ftdfra, 'sac- 
rifice' (beside hotrd), danstrd (later, for ddhstra}; nastrd, 'destroyer'. 

d. Not seldom, a "union- vowel" appears before the suffix; but this is 
not usually the equivalent of the union-vowel used with tr (above, 1182 a). 


STEMS IN iyas, istha, tra, ka, ya, ra. 


For the words in itra have the accent on i: thus, arttra, 'impelling, oar', 
khanttra, 'shovel', pavftra, 'sieve'; janttra, 'birth-place', santtra, 'gift', etc.: 
the combination lira has almost won the character of an independent suffix. 
The preceding vowel is also sometimes a (sometimes apparently of the pres- 
ent-stem) : thus, ydjatra, 'venerable', krntdtra, 'shred', gayatrd (L -trl), 'song', 
pdtatra, 'wing'; but also dmatra, 'violent', -krtatrd, 'cutting-place', vddhatra, 
'deadly weapon', and varatrd, 'strap'. Tarutra, 'overcoming', corresponds 
to tarutr. 

The words still used as adjectives in tra are mostly such as have union- 
vowels before the suffix. A single example from a reduplicated root is 
johutra, 'crying out'. 

e. A word or two in tri and tru may be added here, as perhaps of 
kindred formation with those in tra : thus, dttri, 'devouring', arcdtri, 'beaming'; 
fdtru (fdttru: 232), 'enemy'. 

1186. 3fi ka. The suffix ^ ha is of very common use in 
secondary derivation (below, 1222); whether it is directly 
added to roots is almost questionable : at any rate, extreme- 
ly few primary derivatives are made with it. 

The words which have most distinctly the aspect of being made from 
roots are fuska, 'dry', floka (j/fru, 'hear'), 'noise, report', etc., and -sphaka, 
'teeming'; and stuka, 'flake', and stokd, 'drop', seem to belong together to 
a root stu. Other words in ka are of obscure connections, for the most part. 

But ka enters, in its value as secondary, into the composition of certain 
suffixes reckoned as primary: see aka and uka (above, 1180, 1181). 

A few words in which ika and Ika seem added to a root, though they 
are really of a kindred formation with the preceding, may be most conve- 
niently noticed here : thus, vrfdka (jAmifc), 'scorpion'; dnlka (?), 'face', dffifca, 
'aspect', mrdlkd, 'grace', vrdhlkd, 'increased, a$arlka and ufpcwifca, 'gripes', 
-rjlka, 'beaming'; and, from reduplicated root, parpharika, 'scattering' (?). 
Compare secondary suffix ka (below, 1222). 

1187. ET yet" It is altogether probable that a part of the 
derivatives made with this suffix are not less entitled to be rank- 
ed as primary than many of those which are above so reckoned. 
Such, however, are with so much doubt and difficulty to be 
separated from the great mass of secondary derivatives made with 
the same suffix that it is preferred to treat them all together under 
the head of secondary formation (below, 1210 13). 

1188. ^ ra. With this suffix are made a considerable 
number of adjectives, almost always with weak root-form, 
and usually with accent on the suffix. Also, a few words 
used as nouns, of various gender. 


In some cases, the suffix is found with a preceding vowel, 
having the aspect of a union-vowel. 

a. Examples of adjectives in ra of obvious derivation are : ksiprd, 'quick', 
chidrd, 'split', turd, 'strong', bhadrd, 'pleasing', paferd, 'mighty', pwfera, 'bright', 
hihsrd, 'injurious J ; with accent on the root, grdhra, 'greedy', dhira, 'wise' 
(secondary?), vfpra, 'inspired'. 

b. From roots with prefixes conie only an example or two : thus, niczra, 
'attentive', nimrgra, 'joining on'. 

c. Nouns in ra are, for example : masc. vird, 'man', vajra, 'thunderbolt', 
pura, 'hero'; neut. ksird, 'milk', riprd, 'defilement'; fern, dhdra, 'stream', 
stira, 'intoxicating drink'. 

The forms of this suffix with preceding vowel may best be considered 
here, although some of them have nearly gained the value of independent 
endings. Thus : 

d. With ara are made a few rare words : the adjectives dravard, 'run- 
ning', patard, 'flying'., (with prefix) w/ocard, 'suiting'; and the neuters 
gambhdra, 'depth', tdsara, 'shuttle', sdnara, 'gain': bharvard and vasard are 
doubtless of secondary formation ; and the same thing may be plausibly con- 
jectured of others. 

e. With ira are made a few words, some of which are in common use : 
thus, cy'ird, 'quick', isird, 'lively', madird, 'pleasing', dhvasird, 'stirring up', 
badhird, 'deaf; perhaps sthdvira, 'firm'; and sthird, 'hard', and sphird, 'fat', 
with displacement of final radical a; also sarird, 'wave' (usually salila), and 
one or two other words of obscure derivation. 

With ira are made gabhird or gambhird, 'profound', and pdumi, 'mighty'; 
and perhaps pdnra, 'body'. 

f. With ura are made a few words, of which the secondary character 
is still more probable: thus, ahhurd (an/m-ra?), 'narrow', dsura (dsw-ra?), 
'living', vithurd, 'tottering', yddura, 'embracing'. 

With wra, apparently, is made sthurd, 'stout' (compare sthdvira). 

1189. ^T la. This suffix is only another form of the 
preceding, exchanging with it in certain words, in others 
prevalently or solely used from their first appearance. 

Conspicuous examples of the interchange are cukld, sthuld, 
-micla, cithila, salild. 

Examples of the more independent use are: paid, 'protecting', dnila 
(or am'to), 'wind', trpdla, 'joyous'; later capala and tarala "(said to be accented 
on the final), and harsula (the same). Many words ending in la are of obscure 

1190. ^ va. Very few words of clear derivation are 
made with this suffix too few to be worth classifying. 
They are of various meaning and accent, and generally show 
a weak root-form. 

1195] STEMS IN ra. la, va, ri. ru, vi, snu, sna. 401 

Examples are : rfcrd, 'praising', pakvd, 'ripe', fjfcud, 'artful', ranvd, 
'joyful', urd/iud, 'lofty': vdkva, 'twisting'; urud, 'stall', sruvd, 'spoon'; eva, 
'quick, course', dfua, 'horse'. 

The words in va exhibit only in sporadic cases resolution of the ending 
into ua. 

1191. f^" ri. With this suffix are formed, directly or 
with preceding u, a small number of derivatives. 

Thus, for example : fubhrf, 'beautiful', bhuri, 'abundant'; and, with un, 
jdsuri, 'exhausted', ddpuri, 4 pious : , sdhuri, 'mighty'; angfiri (or anguli], 'finger'. 

1192. "^ ru. This suffix makes a few adjectives and 
neuter nouns, either directly or with a preceding vowel. 

Thus: dharu. -sucking', bhiru, 'timid', cdru, 'pleasant'; with preceding 
a-vowel : patdru, 'flying', vandaru, 'praising', piyaru, 'scoffing', and (from 
causative stem, with I for r] patayalu, 'flying', sprhaydlit, 'desiring' (late); 
with preceding e, maderu, 'rejoicing', saneru, 'obtaining', and peru (of 
doubtful meaning). 

1193. fcf m. By this suffix are made: 

Two or three derivatives from reduplicated roots : jdgrvi, 'awake', dddhrvi, 
'sustaining', didivi, 'shining'; and a very few other words ; ghrsvi, 'lively', 
dhruvf, 'firm' (and perhaps jvvri, 'worn out', for jirvi: BR.). 

Here may be mentioned cikitvti (RV., once), apparently made with a 
suffix vit from a reduplicated root-form. 

1194. T snu. With this suffix, with or without a union- 
vowel, are made a few adjective derivatives from roots, but 
still more from causative stems. 

a. From simple roots: direct, jianti, 'victorious', danksnti, 'biting', 
bhusnu, 'thriving', ni-satsm'i, 'sitting down', sthasnu, 'fixed'; with union- 
vowel i, camnw, 'wandering', rocisnu, 'shining', gamisnti (TB.), 'going', 
-marisnu, 'mortal', pra-janisnti, 'generating'. 

b. From causative stems: for example, cyavayisnti (AV.), 'setting in: 
motion', parayisnu, 'rescuing', posayimu, 'causing to thrive', abhi-focayisnti. 
'attacking with heat', pra-janayisnu (K.), 'generating'. 

c. It seems not unlikely that the s of this suffix is originally that of a 
stem, to which nu is added. Such a character is still apparent in kravisnv 
'craving raw flesh (ferain*)'; and also in vadhasnu, 'murderous', and vrdh- 
asnu (?), 'thriving'. 

1195. ^ sna. Extremely few words have this ending. 

It is seen in tlk&na, 'sharp', and perhaps in flaksnd, and -ruksna; and 
in desnd (usually trisyllabic: ddisna), 'gift'. Unless in the last, it is not 
found preceded by i; but it has (like snu, above) a before it in vadhasnd, 
'deadly weapon', fcardsna, 'fore-arm'. 

Whitney, Grammar. 26 


1196. eT tnu. This suffix is used nearly in the same way 
with ^ snu (above, 1194). 

a. As used with simple roots, the t is generally capable of being con- 
sidered the adscititious t after a short root-final, to which nu is then added : 
thus, fcrfniJ, 'active', hatnti, 'deadly', -tatnu, 'stretching'; and, from reduplic- 
ated roots, jigatnu, 'hasting', and jighatnu, 'harming'; but also dartnti, 
'bursting'. Also, with union-vowel, drauifntJ, 'running'. 

b. "With causative stems : for example, dravayitnu, 'hasting', madayitnu, 
'intoxicating', stanayitnti, 'thunder', -amayitnu, 'sickening'. 

C. With preceding a, in piyatnu, 'scoffing', a-rujatnu, 'breaking into'; 
and kavatnu, 'miserly' (obscure derivation). 

1197. H sa. The words ending in suffixal H sa, with or 
without preceding union-vowel, are a heterogeneous group, and 
in considerable part of obscure derivation. A few examples are : 

a. With sa simply: jesd, 'winning' (aoristic ?), ruksd, 'shining'; titsa, 
n., 'fountain'; bhisd, f., 'fear' (rather from the secondary root bhis}. 

b. With preceding i-vowel: tavisd (f. tavisl], 'strong', mahisd (f. ma/im), 
'mighty', 6/iama(?), 'seeking booty'; manlsd, f., 'devotion'. 

C. With preceding w-vowel : arusd (f. artm), 'red', apt/sa, 'ravenous', 
tarasa, 'overcomer', ptirusa and mdnusa (-us-a?), 'man'; plyusa, 'biestings'. 

1198. 5RH asi. A few words in the oldest language are 
made with a suffix having this form (perhaps made by the 
addition of i to as}. 

Thus, atasf, 'vagabond', dharnasf, 'firm', sanasi, 'winning'; and dhasf, 
m., 'drink', f., 'station'. 

1199. P7 ablia. A few names of animals, for the most 
part of obscure derivation, show this ending. 

Thus, vrsabhd and rsabhd, 'bull', farabhd, a certain fabulous animal, 
ferabha, a certain snake, gardabhd and rasabhd, 'ass'. AV. has the adjective 
sthulabhd, equivalent to sthuld. 

1200. a. 35fr^a#, ^V, 3rT^M^, Wf rt. All the words with 
these endings were mentioned above (383d). They have trace- 
able root-connection only in part : those in at are probably re- 
lated to the participles in ant. 

b. 5T^ ad. The words in ad are also given above (ibid.). 

c. (s{ ajj SsT if, 3s{ uf. The words with these endings were 
given at the same place (to be added is bhisdj, 'healer', of which 
the etymology is disputed) : trsndj appears to be a secondary de- 
rivative, from trma, 'thirst'. 

1201. A number of other primary suffixes are either set 
up by the grammarians and supported with examples of question- 
able value, or doubtfully deducible from isolated words traceable 
to known roots, or from words of obscure connection. 

1204] STEMS IN tnu. sa, asi. abha. at, ETC. 403 

A few such may be mentioned here : anda in karanda and vdranda and 
certain unquotable words (prakritized a-forms from the present participle); 
era and ora in unquotable words, and elima (above, 966 d: perhaps a further 
derivative with secondary ima from era); mara (ma or man with secondary 
ra added) in ghasmara etc.; sara ( in matsara, fcara in ptiskara and other 
obscure words, pa in piispa and a number of other obscure words; and so on. 

B. Secondary Derivatives. 

1202. Words of secondary derivation are made by the 
addition of further suffixes to stems already ending in evi- 
dent suffixes. 

But also, as pointed out above (1137 8), to pronominal 
roots, and to verbal prefixes and a few other indeclinable words. 

1203. Changes of the stem. The stem to which 
the suffix is added is liable to certain changes of form. 

a. Before a suffix beginning with a vowel or with y (which 
in this respect is treated as if it were i], final a and /-vowels 
are regularly lost altogether, while a final w-vowel has the guna- 
strengthening and becomes av; r and o and au (all of rare occur- 
rence) are treated in accordance with usual euphonic rule. 

An w-vowel also sometimes remains unstrengthened. 

b. A final n is variously treated, being sometimes retained 
and sometimes lost, even along with a preceding a; and some- 
times an a is lost, while the n remains : thus, vrsanvant, vrsana, 
vrsa, vrsatva, vrsnya, from vrsan. Of a stem ending in ant, the 
weak form, in at, is regularly taken : thus, vaivasvata (vivasvant). 

Other alterations of a final are sporadic only. 

1204. The most frequent change in secondary derivation 
is the vrddhi-sti'eiigthening of an initial syllable of the stem 
to which a suffix is added. 

The strengthened syllable may be of any character : radic- 
al, of a prefix, or of the first member of a compound : thus, 
acmna (aqvin), saumyd (soma), parthiva (prthivi), amitrd (amitra), 
samrajya (samraj), saukrtya (sukrtd), maifravarund (mitravdruna) , 
auccaihcravasd (uccaihcravas). As to the accompanying accent, 
see the next paragraph. 

a. If a stem begins with a consonant followed by y or v, the semi- 
vowel is usually vriddhied, as if it were i or u, and the resulting ai or au 
has y or v further added before the succeeding vowel. 

This is most frequent where the y or v belongs to a prefix as nz, 
uij S u altered before a following initial vowel: thus, naiyayika from 



nyaya (as if niydyaj, vaiya?vd from vyfyva (as if vfyacva], sauva^vya from 
svdfva (as if suvafva); but it occurs also in other cases, as sauvard from 
svdra, fauva from pt>an, against svayambhuva (svayambhu), and so on. 

AV. has irregularly kaverakd from kuvera (as if from kvera, without the 
euphonic y inserted). 

b. This strengthening takes place especially, and very often, before the 
suffixes a and ya ; also regularly before i, ayana (with ayani and kayani], 
eya (with ineya], and later lya; before the compound oka and ika, and later 
aki ; and, in single sporadic examples, before no, ena, ma, ra, and frua(?): 
see these various suffixes below. 

C. In a few exceptional cases, both members of a compound word have 
the initial -ufddTw-strengthening : thus, for example, saumapausnd (VS.: 
somapusdn), kaurupdncala (QB., GB.: kurupancald), caturvaidya (caturveda), 
dihalaukika (ihaloka). Again, the second member of a compound instead of 
the first is occasionally thus strengthened : for example, $atd?arada (RV., AV.), 
pancafdradiya (TB. etc.), 'somardudra (TS.), trisahasri, da$asahasra, purva- 
varaika (not quotable), caturadhyaya, gurulaghava. 

The 0tma-strengthening (except of a final tt-vowel: above, 1203 a) is 
nowhere an accompaniment of secondary derivation : for an apparent exception 
or two, see 1209h,i. 

1205. Accent. The derivatives with initial vrddhi- strength- 
ening always have their accent on either the first or the last 
syllable. And usually, it is laid, as between these two situa- 
tions, in such a way as to be furthert removed from the accent 
of the primitive; yet, not rarely, it is merely drawn down upon 
the suffix from the final of the latter; much less often, it re- 
mains upon an initial syllable without change. Only in the case 
of one or two suffixes is the distinction between initial and final 
accent connected with any difference in the meaning and use of 
the derivatives (see below, suffix eya: 1216). 

No other general rules as to accent can be given. Usually 
the suffix takes the tone, or else this remains where it was in 
the primitive ; quite rarely, it is thrown back to the initial syl- 
lable (as in derivation with initial vrddhi) ; and in a single case 
(ta: 1237), it is drawn down to the syllable preceding the suffix. 

1206. Meaning. The great mass of secondary suffixes 
are adjective-making : they form from nouns adjectives indicat- 
ing appurtenance or relation, of the most indefinite and varied 
character. But, as a matter of course, this indefiniteness often 
undergoes specialization : so, particularly, into designation of pro- 
cedure or descent, so that distinctive patronymic and metronym- 
ic and gentile words are the result ; or, again, into the de- 
signation of possession. Moreover, while the masculines and 
feminines of such adjectives are employed as appellatives, the 
neuter is also widely used as an abstract, denoting the quality 

1208! STEMS IN a. 405 

expressed attributively by the adjective : and neuter abstracts are 
with the same suffixes made from adjectives. There are also 
special suffixes 'very few) by which abstracts are made directly, 
from adjective or noun. 

A few suffixes make no change in the part of speech of 
the primitive, but either change its degree (diminution and com- 
parison), or make other modifications, or leave its meaning not 
sensibly altered. 

1207. The suffixes will be taken up below in the follow- 
ing order. First, the general adjective-making suffixes, begin- 
ning with those of most frequent use (a, ya and its connections, 
, #a); then, those of specific possessive value (in. vant and mant, 
and their connections) ; then, the abstract-making ones (ta and 
tva, and their connections); then, the suffixes of comparison etc.; 
and finally, those by which derivatives are made only or almost 
only from particles. 

1208. 5f a. With this suffix are made a very large class 
of derivatives, from nouns or from adjectives having a noun- 
value. Such derivatives are primarily and especially adject- 
ives, denoting 'having a relation or connection (of the most 
various kind) with' that] denoted by the more primitive word. 
But they are also freely used substantively; the masculine 
and feminine as appellatives, the neuter, especially and fre- 
quently, as abstract. Often they have a patronymic or gent- 
ile value. 

The regular and greatly prevailing formation is that which 
is accompanied with vrdtf^i-strengthening of the first syllable of 
the primitive word, simple or compound. Examples of this 
formation are : 

a. From primitives ending in consonants : with the usual shift of accent, 
ayasd, 'of metal' (dyas], manasd, 'relating to the mind' (mdnas), saumanasd, 
'friendliness' (sumdnas], brahmand, 'priest' (brdhmari), haimavatd, 'from the 
Himalaya' (himdvant), angirasa, 'of the Angiras family' (dngiras}; hastina, 
'elephantine' (hastfn], mdruta, 'pertaining to the Maruts' (martit); with 
accent thrown forward from the final upon the suffix, farada, 'autumnal', 
vairajd, 'relating to the viraf, pausnd, 'belonging to Pushan'; gairiksitd, 
'son of Girikshit': with accent unchanged, mdnusa, 'descendant of Manus'. 

The suffix is added (as above instanced) to the middle stem-form of 
stems in vant; it is added to the weakest in mdghona nnd vartraghna; the 
ending in remains unchanged; an usually does the same, but sometimes loses 


its a, as in pdusnd, trdivrmd, dd^ardjnd ; and sometimes its n, as in brdhmd, 
duksd, bdrhatsdma. 

b. From primitives in r :' jattra, 'victorious' (jetf or jetr, 'conqueror'), 
tvastrd, 'relating to Tvashtar', savitrd, 'descendant of the sun' (savitf). 

c. From primitives in u: usually with puna-strengthening of the w, as 
vasavd, 'relating to the Vasus', artavd, 'concerning the seasons' (f*w), ddnavd, 
'child of Danu' (cZanw), saindhavd, 'from the Indus' (sfndhu}; but some- 
times without, as madhva, 'full of sweets' (mddhu\ pdr^vd, 'side' (pdr$u, 
'rib'), paidvd, 'belonging to Pedu', tdnva, 'of the body' (tanU}. 

d. From primitives in i and *, which vowels are supplanted by the 
added suffix: pdrthiva, 'earthly' (prthivi), [sdrasvatd, 'of the Sarasvati', 
aindragnd, 'belonging to Indra and Agni' (indrdgni); pdnkta, 'five-fold' 
(panfcif), nairrtd, 'belonging to NfrrtC: but dvayd, 'barrenness' (if from a-vl}. 

e. From primitives in a, which in like manner disappears: yamund, 
'of the Yamuna', saraghd, 'honey' etc. (sardgha, 'bee'), kanind, 'natural child' 
(kanina, 'girl'). 

f. A large number (nearly as many as all the rest together) from prim- 
itives in a, of which the final is replaced by the suffix : for example, with 
the usual shift of accent, amitrd, 'inimical' (amitra, 'enemy'), vdrund, 'of 
Varuna', vaifvadevd, 'belonging to all the gods' (vifvddeva), ndirhastd, 'hand- 
lessness' (ntrhasla), vaiya^vd, 'descendant of Vyacva'; gdrdabha, 'asinine' 
(garddbhd), dafva, 'divine' (devd), mddhyandina, 'meridional' (madhydndina], 
pautra, 'grandchild' (putrd, 'son'), satibhaga, 'good fortune' (subhdgd), vddhry- 
afva, 'of Vadhryac,va's race'; with unchanged accent (comparatively few), 
vasantd, 'vernal' (vasantd, 'spring'), maitrd, 'Mitra's', atithigvd, 'of Atithigva's 
race', dafooddsa, 'Divodasa's'. 

The derivatives of this form are sometimes regarded as made by internal 
change, without added suffix. Considering, however, that other final vowels 
are supplanted by this suffix, that a disappears as stem-final also before 
various other suffixes of secondary derivation, and that no examples of deri- 
vation without suffix are quotable from primitives of any other final than a, 
it seems far too violent to assume here a deviation from the whole course of 
Indo-European word-making. 

1209. The derivatives made by adding 5f a without vrddhi- 
change of the initial syllable are not numerous, and are in 
considerable part^ doubtless, of inorganic make, results of 
the transfer to an a-declension of words of other finals. 

a. A number of examples of stems in a made by transfer were noticed 
above (399). The cases of such transition occur most frequently in com- 
position: thus, further, apa- (for ap or dp, 'water'), -rca, -nara, -patha, 
-gava, -diva; from stems in an, -adhva, -astha, -aha, -vrsa, but also -a/ma 
and -vrsna and vrsana; from stems in i, -angula and -rdtra; from the 
weakest forms of anc-stems (407) wcca, nicd, pardcd. 

b. Also occurring especially in composition, yet likewise as simple words 

1210] STEMS IN a 1 ya. 407 

often enough to have an independent aspect, are derivatives in a from norms 
in as (rarely is, us}: thus, for example, tamasd, rajasd, payasd, brahmavar- 
casd, sarvavedasd, devdinasd, parusd, tryayusd. 

c. Similar derivatives from adjectives in in are reckoned by the gram- 
marians as made with the suffix ina: thus, malina, 'polluted', frngina, 
'horned' (not quotable). The only Vedic instance noted is paramesthtna (AV.). 

d. From emc-stems (407) are made a few nouns in ka: thus, dnuka, 
dpaka, upaka, prdfika, parakd, etc. 

e. From stems in r, nestrd, potrd, prafastrd, from titles of priests. 

f. Other scattering cases are : savidyutd, avyusd, virudha, kdkuda, 

g. The Vedic gerundives in tva (tua) have been already (966 a) pointed 
out as made by addition of a to abstract noun-stems in tu. 

h. Trayd and dvayd come with grwna-strengthening from numeral stems; 
ndva, 'new', perhaps in like manner from nti, 'now'; and dntara from antdr (?). 

i. Bhesaid, 'medicine', is from bhisdj, 'healer', with ^wna-change; and 
perhaps devd, 'heavenly, divine, god', in like manner from div (compare 
avayd, above, 1208 d). 

One or two cases have been noticed above, in which the addition of a 
to another suffix has made a seemingly independent suffix. 

1210. T yet" With this suffix are made a very large 
class of words, hoth in the old language and later. 

The derivatives in ya exhibit a great and perplexing variety of form, 
connection, and application; and the relations of the suffix to others con- 
taining a t/a-element iya, lya, eya, ayya, eyya, enya are also in part 
obscure and difficult. In the great majority of instances in the oldest lan- 
guage, the ya when it follows a consonant is dissyllabic in metrical value, 
or is to be read as ia. Thus, in RV., 266 words (excluding compounds) 
have ia, and only 75 have ya always ; 46 are to be read now with ia and 
now with ya, but many of these have ya only in isolated cases. As might 
be expected, the value ia is more frequent after a heavy syllable : thus, in 
RV., there are 188 examples of ia and 27 of ya after such a syllable, and 
78 of ia and 96 of ya after a light syllable (the circumfiexed ya that is 
to say, ia being, as will be pointed out below, more liable to the reso- 
lution than ya or yd). It must be left for further researches to decide whether 
in the ya are not included more than one suffix, "with different accent, and 
different quantity of the i-element; or with an a added to a final i of the 
primitive. It is also matter for question whether there is a primary as well 
as a secondary suffix ya ; the suffix at least comes to be used as if primary, 
in the formation of gerundives : but it is quite impossible to separate the 
derivatives into two such classes, and it has seemed preferable therefore to 
treat them all together here. 

The derivatives made with ya may be first divided into those 
which do and those which do not show an accompanying vrddhi- 
increment of the initial syllable. 


1211. Derivatives in 7J ya with initial vrddhi-stiengthen- 
ing follow quite closely, in form and meaning, the analogy 
of those in f a (above, 1208). They are, however, decidedly 
less common than the latter (in Veda, about three fifths as many). 

Examples are : with the usual shift of accent, ddivya, 'divine' (devd), 
pdlitya, 'grayness' fpalitd), grdtvya, 'cervical' (grivd), drtvijya, 'priestly of- 
fice' (rMj), gdrhapatya, 'householder's' (grhdpati), janardjya, 'kingship' 
(janaTaj), sarhgrdmajitya, 'victory in battle' (sarhgrdmajti), sduvafvya, 'wealth 
in horses' (svdpva), dvpadrastrya, 'witness' (upadrastr); ddityd, 'Aditya' (dditi), 
sdumyd, 'relating to so-met 1 , dtithyd, 'hospitality' (dtithi), prdjdpatyd, 'belonging 
to Prajapati', vdimanasyd, 'mindlessness' (vimanas), sdhadevya, 'descendant 
of Sahadeva'; with accent thrown forward from the final upon the ending, 
Idukyd, 'of the world' (lokd), kdvyd, 'of the Kavi-race', drtvyd, 'descendant 
of Ritu', vayavyd, 'belonging to the wind' (vayu), rdivatyd, 'wealth' (revdnt): 
with unchanged accent (very few), ddhipatya, 'lordship' (ddhipati), frdfsthya, 
'excellence' (frestha), vdfyya, 'belonging to the third class' (vff, 'people'), 
patiihsya, 'manliness' (ptims). 

The AV. has once ndirbddhya, with circumflexed final 4 if not an error, 
it is doubtless made through ndirbddha; vdisnavydii (VS. i. 12) appears to be 
dual fern, of vdisnavi. 

1212. Derivatives in T ya without initial vrddhi-stiQiigth- 
ening are usually adjectives, much less often (neuter, or, 
in TT ya, feminine) abstract nouns. They are made from 
every variety of primitive, and are very numerous (in Veda, 
three or four times as many as the preceding class). 

The general mass of these words may be best divided accord- 
ing to their accent, into : a. Words retaining the accent of the 
primitive ; b. Words with retracted accent ; c. Words with acute 
yd (id); d. Words with circumflexed ya (ia). Finally may be 
considered the words, gerundives and action-nouns, which have 
the aspect of primary derivatives. 

a. Examples of derivatives in ya retaining the accent of their primitives 
are: d$vyct, 'equine' (dfva), dngya, 'of the limbs' (dnga), mtikhya, 'foremost' 
(mukha, 'mouth'), dvya, 'ovine' (dm), gdvya, 'bovine' (g6), vfyya, 'of the 
people' (vfy), dtiryct, 'of the door' (dur), ndrya, 'manly' (nf), vfsnya, 'virile' 
(vrsan), svardjya, 'autocracy' (svardj), suvirya, 'wealth in retainers' (suvTra), 
vifvddevya, 'of all the gods' (vi$vddeva), mayurctfepya, 'peacock-tailed', 

In the last word, and in a few others, the ya appears to be used (like 
ka, 1222 C, 2; ya, 1212 d, 5) as a suffix simply helping to make a possessive 
compound : so suhdstya (beside the equivalent suhdsta), mddhuhastya, ddfa- 
masya, mifrddhdnya. 

1212] STEMS IN ya. 409 

b. Examples with retraction of the accent to the first syllable (as in 
derivation with vrddhi-increment] are: kdnthya, , 'guttural' (kanthd), skdndhya, 
'humeral' (skandhd), vrdtya, 'of a ceremony' (vratd), meghya, 'in the clouds' 
(meghd), pftrya, 'of the Fathers' (pitr), prdtijanya, 'adverse' (pratijand). 
Hiranyaya, 'of gold' (hfranya), is anomalous both in drawing the accent 
forward, and in retaining the final a of the primitive. 

c. Examples with acute accent on the suffix are: divyd, 'heavenly 1 
(div), satyd, 'true' (sdnt), vyaghryd, 'tigrine' (vydghrd), kavyd, 'wise' (kavf); 
gramyd, 'of the village' (grdma), somyd, 'relating to the soma , bhumyd, 
'earthly' (bh&mi), sakhyd, friendship' (sdkhi), jaspatyd, 'headship of a family' 

d. Of derivatives ending in circumflexed ya (which in the Veda are 
considerably more numerous than all the three preceding classes together), 
examples are as follows : 

1. From consonant-stems: vi$yh, 'of the clan' (RV.: vfp), hrdya, 'of the 
heart' (hid), vidyutya, 'of the lightning' (vidytit), rdjanya, 'of the royal class' 
(rdjan), dosanyh, 'of the arm' (dosdn), firsanya, 'of the head' (flrsdnj, 
karmanya, 'active' (kdrman), dhanvanya, 'of the plain' (dhdnvan), namasya, 
'reverend' (ndmas), tvacasytt, 'cuticular' (tvdcas), barhisya, 'of barhts 1 , ayusyb, 
'giving life' (dyus), bhasadytt, 'of the buttocks' (bhasdd), prdcyh, 'eastern' 
(prdnc), etc. Of exceptional formation is aryamya 'intimate' (aryamdn). 

2. From w-stems : hanavyci, 'of the jaws' (hdnu), vayavyh, 'belonging 
to Vayu', pa?avya, 'relating to cattle' (pw&), isavya, 'relating to arrows' (isu); 
faravyh, f., 'arrow' (fdru, do.); and there may be added nauya, 'navigable' 
(especially in fern., naut/3, 'navigable stream': nat/, 'boat'). 

Under this head belong, as was pointed out above (chap. XIV., 964). 
the so-called gerundives in tavya (later tdvya], as made by the addition of 
yh to the infinitive noun in tu. They are wholly wanting in the oldest lan- 
guage, and hardly found in later Vedic, although still later tavya wins the 
value of a primary suffix, and makes numerous derivatives. The RV. has 
prdfavyh, 'to be partaken of (pra -\- Yaf), without any corresponding noun 
prdfu; and also urjavyb, 'rich in nourishment' (urj), without any intermedi- 
ate urju. 

3. From i and z-stems hardly any examples are to be quoted. VS. has 
dundubhya from dundubhi. 

4. From a-stems : svargyh, 'heavenly' (svargd), devatyh, 'relating to a 
deity' (devdtd), prapathyft, 'guiding' (prapatha), budhnyfr, 'fundamental' 
(budhnd), jayhanyh, 'hindmost' (jaghdna), varunyti, 'Varuna's', viryh, 'might' 
(vir&), udaryh, 'abdominal' (uddra), utsyh, 'of the fountain' (6tsa); and from 
a-stems, urvaryct, 'of cultivated land' (urvdrd), svahya, 'relating to the ex- 
clamation svdha 1 . 

The circumflexed yh is more generally resolved (into fa) than the other 
forms of the suffix : thus, in RV. it is never to be read as ya after a heavy 
syllable ending with a consonant ; and even after a light one it becomes fa 
in more than three quarters of the examples. 

5. There are a few cases in which yh appears to be used to help make 


a compound with governing preposition (next chapter, 1310): thus, apikaksyft, 
'about the arm-pit', upapaksya, 'upon the sides', udapya, 'up stream'; arid 
perhaps upatrnyh, 'lying in the grass' (occurs only in voc.). But, with other 
accent, dnvdntrya, 'through the entrails', tipamasya, 'on each month'; abhi- 
nabhyd, 'up to the clouds', ddhigartya, 'on the chariot floor'. 

1213. The derivatives in TJ ya as to which it may be 
questioned whether they are not, a least in part, primary 
derivatives from the beginning, are especially the gerund- 
ives, together with action-nouns coincident with these in 
form ; in the later language, the gerundive-formation (above, 
963) comes to be practically a primary one. 

a. In RV. occur about 40 instances of gerundives in ya, of tolerably 
accordant form : the root usually unstrengthened (but cetya, bhdvya, -hdvya, 
mdrjya, y6dhya; also -mddya, -vdcya, bhavyd); the accent on the radical